Every question we get asked, answered in one place
The company, getting started and how the work runs, then every question from every service page and every article, kept current with the pages they come from.
Loading FAQ hero image
About Xcelerate Brands
Who we are, what we do, and what we do not.
We are a commerce growth and operations partner. We run the marketplaces and storefronts a brand sells on, the advertising that feeds them, the creative and compliance work around them, and two commercial partnerships for brands that want a different arrangement from a retainer. The work is organised into five service families, each with its own page.
Amazon, Walmart, eBay and TikTok Shop, and Shopify for a brand's own storefront. Each has its own service page describing what running it involves.
An operator. We run accounts, advertising and listings inside your own seller accounts rather than writing a strategy document for somebody else to implement. Where a brand wants advice alone, that is a conversation we can have, but the work we are known for is doing it.
Since 2015. The About page carries the company facts and the team.
The registered office is in Sheridan, Wyoming, with teams in Europe and South Asia. The offices are listed on the contact page, and the map on the About page shows where the people are.
Brands selling on marketplaces and through their own storefronts, from established consumer brands to retailers with a catalogue to move. The case studies show the categories we have grown and what changed on each account.
Yes. Most of the demand is on the US marketplaces, and we also run accounts in the UK and the EU, including Amazon Renewed in Europe, and on the marketplaces' other regions where the brand sells there.
Not as a reseller of our clients' goods. The retail growth partnership is the one arrangement where we sell under a client's brand, and in it the inventory stays the client's until it sells. The partnership page explains the model.
A family is an area of work, such as Marketplace Management or Compliance & Account Health. A service is one thing inside it with its own page, such as Amazon or Account Reinstatement. The Services page lists every family and what is inside each.
Yes, through xB Academy. The courses page lists what is taught, and the community is where people who have taken a course keep asking questions.
Yes. BPO Services covers an outsourced team working under your own name, with your processes and your reporting. It sits under Strategic Partnerships because it is an arrangement rather than a project.
We do not promise rankings, reinstatements, approvals or revenue figures, because none of them is ours to promise. We do not buy reviews, manipulate rankings or work around a marketplace's policies. And we do not take on work where we do not see a realistic path to improvement.
Getting started
What happens between the first message and the first month.
Send the form on the contact page. Tell us the category, where you sell and what is not working, and the first reply will already say something useful about it.
A review of your category and your account before any talk of a retainer, sent back through the form. We look at who ranks above you, where their position comes from and whether it can realistically change, and we tell you what we would change first.
The category you sell in, the marketplaces you are on, and access to the account if you want the audit to be specific. Seller Central, Walmart Seller Center and the advertising consoles all allow read-only user permissions, which is all the audit needs.
By whether there is a realistic path to improvement. If the sellers above you hold their place for reasons additional spend will not change, we say so. Taking a fee for work that cannot move the numbers serves nobody.
It depends on the account and the channels, and we say how long each step takes when we propose the work. Access, a read of the catalogue and the advertising, and an agreed plan come before anything is changed.
Always. The seller accounts, the advertising accounts, the storefront and the brand registrations stay in your name, and we work inside them as users you can remove. Nothing about the arrangement makes leaving difficult.
Yes, and it is a common starting point. We read what the previous agency built before changing it, keep what is working, and say plainly what is not.
Yes, where the category supports it. A launch is a different plan from a takeover, and the audit is where we say which one applies and what it would need.
Yes. Project services such as Design, Product Listing Optimization, Keyword Research, Unit Economics, Cost Recovery, Trademark Registration and Brand Registry are bought as pieces of work. Managed services run month to month.
It depends on the channels, the catalogue and the scope, and it is set out in the proposal after the audit. We do not publish a price list, because two accounts in the same category can need very different work.
The proposal states the term for the work it describes. Managed work is planned in months because results on a marketplace take months. Project work ends when the project does.
A named account lead, and the specialists the account needs behind them. The team page introduces the people.
How the work runs
Reporting, operations, and what we measure.
Contribution after fees, advertising, returns and storage, alongside share of voice in the category. Growth that costs more than it earns is not reported as a result.
A regular report on the cadence agreed in the proposal, and a message whenever something on the account needs a decision from you. The reporting is built so a question about any number can be answered from the underlying data.
Inside yours. Listings, advertising, inventory and compliance are handled in your own seller and advertising accounts, so everything that is built stays with you.
Stock, pricing and demand decisions cross between channels, so they are made together rather than per marketplace. A promotion on one channel is planned knowing what it does to stock on the others.
Yes. Campaign structure, targeting, bids, budgets and the search term work are done by the team, against contribution margin rather than against the platform's own report of itself.
The Compliance & Account Health family covers it. We establish what actually triggered the action, assemble the case the process asks for, submit it through the right route and follow it. The decision is always the marketplace's to make.
No, and nobody honest can. A marketplace decides rankings, approvals and reinstatements. What we control is whether the work is done properly and reported honestly, and the case studies show what that has produced on real accounts.
Access is through named user permissions in your own accounts, limited to what the work needs and removable by you at any time. We do not hold your account credentials.
The marketplaces' own consoles and advertising platforms first, with reporting built on their data. Where a third-party tool earns its place for research or monitoring, it is used, and it is never a substitute for reading the account.
Yes. Everything is done in your accounts, so the history is visible there, and the reporting names what was changed and why.
On the notice the proposal states. Your accounts, listings, campaigns and registrations are yours and stay as they are; we hand over what we built and remove our access.
Every service has its own page with what it covers, how it runs and its own questions, and those questions are repeated further down this page under the service's name.
About Marketplace Management
Listings and catalogue, advertising, inventory and replenishment planning, account health and compliance, pricing and promotions, and the reporting that says what each channel actually contributed. On every marketplace we run, the whole operation rather than a slice of it.
No, and most brands should not. Each channel has its own economics, and one run well earns more than four run thinly. The first conversation is about which channels your catalogue and margins suit.
Yes. Most of the work starts on accounts already trading, which is usually where the largest short-term gains are: an established account with a structural problem is a faster win than a new channel.
One team across all of them, so stock, pricing and promotional decisions are taken knowing what they do elsewhere. Brands that split channels between suppliers usually discover the interaction later, in the inventory.
Amazon account management is the ongoing operation of a brand's Seller Central or Vendor Central presence: listing content and keywords, advertising, account health and compliance, inventory and forecasting, and the reporting that ties it to profit. It is continuous rather than a project, because Amazon's auction, catalogue rules and policies all change without notice.
Ten areas, weighted differently for each account: compliance operations, Seller and Vendor Central administration, listing optimisation, Amazon SEO, A+ and Brand Store creative, sponsored advertising, DSP, account health and reinstatement, inventory and FBA reimbursements, and reporting on contribution margin. A brand losing money to stockouts needs different attention from one whose listings never rank.
It depends on catalogue size, how many marketplaces are in scope and whether advertising is included, so we quote after the audit rather than from a rate card. The test we apply is whether the fee is smaller than the money currently being lost to stockouts, unclaimed reimbursements, untended advertising and listings that do not convert. If the audit does not find that, we say so.
Advertising changes show in two to three weeks, because that is how long a campaign needs to gather enough clicks per keyword to be read honestly. Listing and keyword changes take four to eight weeks, since ranking responds to sales history rather than to the edit. Account health work can resolve in a week or take months, depending entirely on Amazon.
Yes. The Seller Central or Vendor Central account stays in your name, we work inside it under user permissions you grant and can revoke, and every asset we create is yours. You can end the arrangement and keep the listings, the campaign structure, the creative and the data.
An audit, then the fixes that stop money leaving. In practice that is suppressed or miscategorised listings, stranded and aged inventory, advertising campaigns running without negative keywords, unclaimed FBA reimbursements, and missing or non-compliant images. Strategy comes second, because none of it survives an account that is bleeding.
In-house wins when Amazon is the whole business and you can hire three or four specialists, because nobody will ever care about your catalogue as much as your own people. An agency wins when Amazon is one large channel among several, when the work needs six skills at a fraction of a person each, or when the account is in trouble now and hiring takes three months. The failure case for both is one overloaded generalist.
Seller Central is third party: you sell to the customer, set your own prices, hold the inventory, and keep the margin and the risk. Vendor Central is first party: you sell to Amazon at wholesale, Amazon sets the retail price, and you trade margin for volume and for not owning the inventory. They have different economics, different advertising tools and different failure modes, and a brand can run both at once.
Yes. Vendor Central is a different discipline from Seller Central: purchase orders, chargebacks, cost negotiations, retail analytics, and content submitted through a different pipeline. Brands running both need the two coordinated, because they compete with each other on the same detail page if nobody is watching.
Amazon SEO is ranking inside Amazon's own search, which is a purchase engine rather than an information engine. It rewards conversion and sales velocity for a keyword, not backlinks or domain authority. A listing ranks because people who searched that term bought it, so the work is making a listing eligible for the right terms and then making it convert once it appears.
Backend search terms are hidden keyword fields that make a listing eligible to appear for terms the visible copy does not contain. They still matter, and they are the most commonly wasted field on Amazon: duplicated words, competitor brand names that are not permitted, and commas that do nothing. They will not rank a listing on their own. They decide what it can rank for at all.
Platinum keywords are a legacy Seller Central field left over from the old Platinum Merchant programme. For almost every seller they do nothing and can be left empty. They are worth mentioning only because the field still appears, and a seller reasonably assumes an empty field is a missed opportunity. It is not.
Advertising buys sales history for a keyword, and sales history is the main input to organic ranking on that keyword. So PPC is how a new listing earns an organic position it could not otherwise reach. The search term report is also the cheapest keyword research available, because it records what real buyers typed before they bought your product.
There is no universal number, and an agency quoting one is describing their other clients rather than your business. ACoS is only meaningful against contribution margin: 30% is excellent on a product with 60% margin and ruinous on one with 25%. The figure worth managing is total advertising cost of sale against profit, not advertising cost against advertising revenue.
Often, and it depends far more on the first appeal than on the tenth. Amazon wants a plan of action identifying the root cause, what has been corrected and what will prevent recurrence, with evidence. Generic appeals and repeated resubmissions of the same text make reinstatement harder. We cannot promise reinstatement, and an agency that does is selling something Amazon has not agreed to.
Ungating is Amazon's approval process for restricted categories and brands, usually requiring invoices from an authorised supplier and sometimes documentation or test reports. You need it if the category refuses your listing attempt. It is administrative rather than strategic work, and it fails most often because the invoices do not meet Amazon's requirements rather than because the seller was refused.
Amazon regularly loses, damages or miscounts inventory in its own warehouses and owes the seller for it. Some is refunded automatically and a meaningful share is not, which is why it has to be claimed. Most accounts that have never run a reconciliation are owed something, and claims expire, so the older the account the more has usually gone.
From sales velocity, lead time, seasonality and planned advertising, reconciled against what the supply chain can actually deliver. The point is not accuracy for its own sake, it is avoiding the stockout: ranking decays while a product is unavailable, and the advertising that bought that ranking has already been paid for.
For most brands selling small, standard-sized, steadily moving products, yes, because Prime eligibility and Amazon's delivery promise lift conversion enough to cover the fees. It stops being worth it on large or heavy items, on slow movers that accrue storage fees, and on products whose margins cannot absorb fulfilment and referral fees together. It is an arithmetic question per product, not a decision for the whole catalogue.
No. Amazon allows individuals to register as sole proprietors and many sellers start that way. An LLC or equivalent is usually taken for liability and tax reasons rather than because Amazon requires it, and Brand Registry needs a registered trademark rather than a particular company structure. Take advice from an accountant in your jurisdiction rather than from an agency.
There is no minimum ASIN count. What matters is whether the account is large enough that the fee is small against what is recoverable, and whether there is inventory to sell. We would rather tell a brand it is too early than take a retainer that cannot pay for itself, and we say that at the audit rather than at renewal.
Frequently. The usual arrangement is that the brand keeps product, pricing and supply, and we run the marketplace execution. What has to be agreed at the start is who decides on price and promotions, because that is the one place a brand team and a marketplace team will disagree and both will be partly right.
Monthly reporting on contribution margin by ASIN rather than gross revenue, advertising performance against agreed targets, inventory position and forecast risk, account health status, and a written note on what was tried and what did not work. Weekly figures are available and are usually noise. The monthly view is the one worth acting on.
Walmart, eBay, TikTok Shop and Shopify, plus Google Ads and Meta advertising alongside them. Brands most often come to us for Amazon and add a second channel once the first is stable, which is the right order: a second marketplace multiplies the operational load before it multiplies revenue.
A product nobody wants, a price the category will not bear, a supply chain that cannot keep stock in the country, or margins too thin to fund advertising. Marketplace work makes a viable product findable and profitable. It cannot make an unviable one viable, and an agency that takes the retainer without saying so is charging you to find that out slowly.
A conversation about what the account is doing now and what is going wrong, then an audit of the account itself. The audit is where the specifics come from and it is worth having even if you do not go on to work with us, because it produces an ordered list of what is losing money, which is useful to anybody who owns the account.
Walmart marketplace management is the ongoing operation of a brand's Walmart Seller Center presence: item setup and catalogue compliance, listing content and keywords, Walmart Connect advertising, pricing position, fulfilment through WFS or your own network, inventory planning, and reporting on whether the channel earns its cost.
For brands with competitive pricing and a catalogue Walmart shoppers buy, usually yes, because the channel is far less contested than Amazon and cost per click through Walmart Connect is often lower. It is not worth it for premium-priced niche products, or where margins cannot absorb another channel's operational cost. It is a question worth answering before onboarding rather than after.
Three differences matter most. Walmart weighs price competitiveness heavily in its ranking and runs parity checks against other retailers, so an overpriced item can be suppressed however good its content is. Its item setup standards are stricter, so catalogues imported from Amazon arrive with errors. And it publishes a content quality score, which makes listing quality unusually measurable.
You can import them, and it is the most common reason a Walmart launch underperforms. Titles and attributes that pass on Amazon are frequently rejected or de-prioritised here, the required category fields are different, and Walmart's search responds to different phrasing. The catalogue needs rewriting for the channel, not copying into it.
Walmart Connect is Walmart's advertising platform, covering Sponsored Products, Sponsored Brands and display placements on and off the site. It works much as Amazon Ads does, in a smaller auction. Lower competition usually means a lower cost per click, which makes it a reasonable place to buy early visibility while a new listing builds sales history.
Walmart Fulfillment Services is Walmart's own fulfilment network, the equivalent of FBA. It carries the delivery promise that improves search placement and conversion, so it is usually worth it for standard-sized, steadily moving products. Seller fulfilled can compete where you can meet the delivery speed yourself, since fulfilment speed feeds ranking either way.
Most often a price above Walmart's competitiveness threshold, a failed parity check against another retailer, missing required category attributes, or an image that does not meet specification. Suppression on Walmart is more frequently a pricing or data problem than a policy one, which is good news because both are fixable quickly.
Walmart ranks on relevance to the query, content completeness and quality, price competitiveness, fulfilment speed, and sales performance. The unusual part is how much weight price and delivery carry compared with Amazon, which means listing work alone will not rank an item that is priced above the category.
Walmart publishes a score for each item measuring how complete and rich its content is: attributes filled, image count, description quality, rich media present. It is one of the few ranking inputs a marketplace states directly, which makes it a rare case where the work needed is unambiguous.
Account approval and setup typically takes days to a few weeks. Getting a catalogue live cleanly takes longer and depends almost entirely on data quality: a well structured catalogue with complete attributes can be live in two to three weeks, while one imported from an Amazon feed usually spends a month in error resolution.
Yes, including spec sheets, category attribute mapping and the bulk upload process, plus the rejection and error resolution that follows. This is the least visible and most decisive part of a Walmart launch, because items that never go live cleanly cannot be optimised later.
Yes, and they are difficult to separate here. Walmart's ranking rewards sales performance, so advertising buys the early velocity a new listing needs, and listing quality decides whether that spend converts. An agency running one without the other is optimising half a system.
It depends on catalogue size, whether item setup is included and whether advertising is in scope, so we quote after assessing the catalogue. The honest framing for a brand adding Walmart is that the first months are setup cost against uncertain revenue, and it should be entered as an investment in a channel rather than as an immediate return.
We do not sell them, and we would advise caution about anybody who does. In this market the phrase usually describes done-for-you dropshipping stores sold as passive income, which is a different business from managing a brand's own catalogue, carries real account risk, and is often outright fraudulent. What we do is manage marketplaces for brands that own their products.
That is the majority of this work. The advantage is one team seeing both catalogues, both ad accounts and the shared inventory, which is the thing two separate agencies cannot do. The first question is usually which products belong on which channel, because putting the whole catalogue everywhere raises operational cost without raising revenue proportionally.
A rollback is Walmart's temporary price reduction badge, and it does drive visibility and conversion. Used deliberately for a launch or a seasonal push it buys velocity. Used routinely it trains the category to wait for the discount and makes the undiscounted price look like the wrong one, which is a slow way to lose margin.
Yes. The Seller Center account stays in your name and billing, we work in it under permissions you grant and can revoke, and the listings, content, campaigns and data are yours. You can end the arrangement and keep all of it.
Both platforms monitor prices elsewhere, so they are not independent decisions. Pricing lower on Walmart to win competitiveness can affect Amazon's own price checks, and pricing higher can get items suppressed on Walmart. A brand selling on both needs one pricing policy across channels rather than two teams optimising in opposite directions.
Monthly reporting on contribution margin by item, advertising performance against agreed targets, content quality and catalogue health, channel share against Amazon, and a written note on what was tried. Including, where it applies, an honest read on whether the channel is worth continuing.
Usually, and the diagnosis is normally one of four things: items never went live cleanly, content was copied from Amazon, pricing sits above the competitiveness threshold, or fulfilment speed is losing the placement. All four are fixable, and none of them are fixed by advertising harder.
Start with the products that are priced competitively and already sell, not the whole catalogue. A focused launch of thirty well-built items outperforms two thousand imported ones, and it keeps the item setup work small enough to do properly. Breadth can follow once the channel is proven.
A price the channel will not accept, a product with no demand on Walmart specifically, or a supply chain that cannot support a second channel's stock. Walmart is price led in a way Amazon is not, and no amount of content or advertising work makes an uncompetitively priced item rank.
A conversation about the catalogue and its pricing, then an assessment of whether Walmart fits it. If it does, item setup and content come before advertising. If it does not, we would rather tell you at that point than take a retainer to find out slowly.
eBay account management is the ongoing operation of a seller account: listing structure and content, item specifics and catalogue data, Promoted Listings, pricing and listing format, postage settings, and the seller performance metrics that decide placement. On eBay the last of those matters more than on any other marketplace, because performance is scored over a rolling period.
For the right catalogue, yes. It is less contested than Amazon in many categories, fees are often lower, and it supports used, refurbished and open-box inventory that Amazon restricts. It is a poor fit for brands wanting tight control of presentation and pricing across a large new-product catalogue, where Amazon's structure serves better.
eBay's search engine, Cassini, weighs relevance to the query, listing quality and completeness, price competitiveness, sales history, postage speed and seller performance. The distinguishing feature is that seller account standing feeds directly into placement, so a good listing on a poor account is held back by something that has nothing to do with the listing.
You need a minimum number of transactions and sales value over the previous twelve months, a defect rate at or below eBay's threshold, a low late dispatch rate, cases closed without seller resolution below the limit, and tracking uploaded within the stated handling time. The requirements vary by marketplace region, and the metric that most often fails is tracking uploaded on time rather than anything to do with the product.
Better search placement, the Top Rated Plus badge on listings that qualify, and a discount on final value fees. The badge affects conversion because it appears in search results next to competitors without it, and the fee discount is meaningful at volume. Together they make the status worth protecting rather than chasing back after it is lost.
Selling limits rise as an account builds a record: consistent sales, on-time dispatch, low defects and prompt case resolution. Requests can be made through Seller Hub once the account has performed for a period, and they are granted more readily on accounts with clean metrics. There is no way to buy an increase, and services offering one are worth avoiding.
Through a multi-variation listing, which holds sizes, colours or other variants under a single item. It is almost always better than separate listings, because sales history, reviews and search authority accumulate in one place instead of being split across twenty weak listings. The structure has to be set up before selling begins, since converting later loses history.
Item specifics are the structured attributes on a listing: brand, size, colour, material, model and so on. They matter because they decide which filtered searches your listing appears in at all. A shopper narrowing by brand and size will never see a listing that left those fields empty, however well the title is written.
Standard charges a percentage of the sale price, and only when a buyer clicks the promoted listing and then purchases within the attribution window. Advanced is a cost per click auction with keyword targeting, charged whether or not the click converts. Standard suits most sellers; Advanced suits high volume accounts that can manage bids actively.
Amazon has far more traffic and stricter control: one catalogue, one detail page, and competition on the Buy Box. eBay gives sellers more control over presentation and format, supports used and refurbished inventory, and often costs less in fees. Brands with new products and a large catalogue usually lead with Amazon; sellers with varied, used or hard-to-find inventory frequently do better on eBay. Many brands should be on both.
eBay's measure of transactions that went wrong in ways attributable to the seller, principally cases closed without seller resolution and cancellations initiated by the seller for out-of-stock items. Exceeding the threshold costs Top Rated status and can trigger selling restrictions, so it is monitored continuously rather than reviewed monthly.
Fixed price for anything with predictable demand and repeatable stock, which is most brand inventory. Auctions suit genuinely scarce, collectible or hard-to-value items, where letting the market set the price finds a number you would not have guessed. Running a whole catalogue as auctions is a way to sell everything slightly too cheaply.
Yes, including bulk listing tools, variation structure, item specifics mapping and inventory feed connections. For an existing catalogue the more common work is restructuring rather than creating: consolidating split listings and completing the attribute data that was skipped at launch.
Delivery speed and the estimated delivery date shown on the listing both feed placement, and handling time is one of the inputs to Top Rated status. Reducing stated handling time and uploading tracking promptly is often the cheapest ranking improvement available on an eBay account, because it costs nothing but discipline.
Frequently, and it is the arrangement we prefer. The two share inventory and pricing decisions, so having one team see both prevents the most common failure, which is a stockout on the stronger channel caused by an unplanned surge on the weaker one.
It depends on catalogue size, whether listing creation is included and whether advertising is in scope, so we quote after auditing the account. For accounts with damaged metrics, the honest framing is that the first phase is recovery rather than growth, and pricing should reflect that.
Usually, and it takes time rather than cleverness. Metrics are scored over a rolling period, so the only route is a clean run of transactions: dispatch on time, upload tracking, resolve cases before they close against you. Anybody promising a fast fix is describing something eBay does not offer.
Yes. The account stays in your name, we work under permissions you grant through Seller Hub and can revoke, and the listings, feedback history and data are yours. Feedback and seller standing are tied to the account and cannot be transferred, which is another reason nobody should ever sell through an agency's account.
Monthly reporting on contribution margin by listing and category, Promoted Listings performance, seller metric trends with any risk flagged early, and a written note on what was tried. Where eBay is a clearance channel rather than a growth one, the reporting should say that plainly rather than dress it up.
Yes, and eBay is often the better channel for it. Condition description, grading consistency and photography matter far more here than on new inventory, because the buyer is assessing risk as much as product. Amazon Renewed is the equivalent programme on Amazon and we run that too.
Inventory the market does not want, prices above what the category clears at, or an account with restrictions that require time rather than intervention to lift. eBay is also slower to reward improvement than Amazon, because so much of its ranking rests on rolling history, and that patience has to be part of the plan rather than a surprise.
An audit of the account: seller metrics and standing first, then the catalogue structure and listing quality. Metrics come first because they constrain everything else, and there is no point optimising listings on an account that is one late dispatch away from losing its status.
TikTok Shop management is running a brand's shop on TikTok end to end: registration and compliance, catalogue and product content, affiliate and creator partnerships, live selling, paid media, promotions, and fulfilment. It differs from other marketplace management because demand is created by content rather than captured from search, so creator and content work is the growth engine rather than a supporting activity.
Amazon captures existing demand: somebody searches, you compete for that search. TikTok Shop creates demand: content appears in a feed and prompts a purchase nobody set out to make. That means keywords and bids matter far less, creators and video matter far more, and a listing that would rank well on Amazon can sit unsold here because nothing is driving attention to it.
It depends on the category more than on the brand. Products that demonstrate well in a few seconds, sit at an impulse price point and have margin to absorb commission and vouchers do well. Considered, high-price or hard-to-demonstrate products struggle, and there is no amount of management that changes that. It is worth assessing honestly before setup rather than after six months.
Sellers offer commission to creators, who then feature products in their content and earn a share of resulting sales. Plans can be open to any creator, targeted at specific ones, or negotiated individually. Commission rate, product appeal and how easy you make sampling all decide whether creators pick you up, and it is a continuous recruitment exercise rather than a launch task.
Through the affiliate marketplace at volume and through direct outreach to creators already covering the category, backed by sample seeding. The hit rate is deliberately low, so it is run as a pipeline with continuous outreach rather than as a campaign. Brands that treat it as occasional outreach conclude creators are not interested, when what was missing was volume and follow-up.
Enough to compete for attention in your category while leaving margin after TikTok's fees, vouchers and returns. Rates vary widely by category, and the mistake at both ends is common: too low and no creator picks it up, too high and the volume that arrives is unprofitable. It should be set from your own unit economics rather than from a benchmark.
Yes, and before recruiting creators rather than instead of it. Brand-owned content proves the product works in the format and gives creators a reference for what converts. A shop with no content of its own is a harder proposition to a creator deciding whether to accept a sample than it is to a shopper.
Live selling is a real-time stream where a host demonstrates products and viewers buy without leaving. It works well in categories where demonstration and urgency matter, and it moves a substantial share of TikTok Shop volume in those categories. It is operationally demanding: it needs a host, a schedule, stock readiness and offers built for the format.
Video Shopping Ads and related formats place product-linked video into the feed, with campaign types that optimise towards purchases. The important discipline is that paid spend should amplify content already performing organically. Putting budget behind a video the feed has ignored is the fastest way to conclude the channel is expensive.
Register and verify the shop, upload and get products approved, set up fulfilment and returns, then build the demand side: brand content first, then affiliate plans and creator recruitment, then paid amplification of whatever is working. The setup is administrative and takes days to weeks; the demand side is the actual job and is continuous.
TikTok Shop is built around physical goods, and support for digital products is limited and varies by market. Categories, eligibility and rules change often enough that the current position should be checked against TikTok's own seller documentation rather than taken from any agency page, including this one.
Registration and verification usually take days. Getting a catalogue live and approved takes longer and depends on category and data quality. The realistic timeline to meaningful sales is longer still, because it depends on creator recruitment and content rather than on setup, and that is measured in months rather than weeks.
Yes: the logistics, tracking and follow-up. Seeding is where creator programmes stall, because sending samples is easy and tracking who received what, who posted, and what it produced is the part that turns it into a channel rather than a cost.
TikTok takes a commission on sales, with rates varying by category and market, plus whatever affiliate commission you offer creators and the cost of any vouchers or promotions. The combined figure is materially higher than a marketplace referral fee alone, which is why margin has to be checked before entering rather than after.
Some crossover is normal and mostly not worth worrying about, because the audiences are reached in different states of mind: one is searching, the other was not. The real risk is operational, not commercial. A video that performs can clear a month of inventory in a day and leave the channel that was already working out of stock.
Yes, and it is the arrangement that avoids the inventory problem above. Shared stock and pricing decisions need one team seeing all of it, particularly on a channel whose demand is spiky by design.
Treating it as a third marketplace listing, under-resourcing creator recruitment, offering commission that does not compete, launching without brand content, and running out of stock during the one surge that worked. Almost none of the common failures are about the product itself.
Yes. The shop stays registered to your business, we work under permissions you grant and can revoke, and the catalogue, creator relationships and content are yours. Creator relationships in particular are worth owning, because they are the asset that took longest to build.
Monthly reporting on GMV by creator, video and campaign, margin after commission, vouchers and returns, creator recruitment pipeline, and content performance. Plus a straight view on category fit, which on this channel is the question that matters most in the first six months.
It depends on whether content production and creator recruitment are in scope, which is where most of the work sits. We quote after assessing category fit, and if the assessment says the catalogue does not suit a content-led channel we will say that instead of quoting.
A product that cannot be shown working in a few seconds, a price point that does not suit impulse purchase, or margins that cannot absorb commission plus vouchers plus returns. Management makes a viable product visible on this channel. It cannot make an unsuitable one suitable.
An honest assessment of category fit first, using your products, price points and margins. If it fits, setup and brand content come before creator recruitment, because creators are choosing between shops and the shop needs to look worth choosing.
eCommerce SEO is the work of getting an online store and its product pages to rank in Google: site structure and crawlability, page speed, structured data, keyword targeting per page, content that answers buying questions, and internal linking between them. It differs from marketplace optimisation because Google ranks what it can crawl, understand and trust, whereas Amazon and eBay rank largely what already sells.
The discipline is the same and the obstacles are Shopify's own. Its URL structure creates duplicates through collections, tags and variants; its apps inject scripts into every page; and its theme decides much of your Core Web Vitals before you write anything. Shopify SEO is therefore weighted more heavily towards technical cleanup than SEO on a platform where the developer controls the templates.
Technical fixes can show within two to four weeks, because they change what Google is able to index rather than how it judges quality. Content and authority take three to six months, and competitive commercial terms longer than that. Any agency promising first page in a month is describing a keyword nobody searches.
It depends on catalogue size, how much technical debt is in the theme and whether content production is included, so we quote after the audit. The useful comparison is against paid traffic: if a store is spending on ads to reach terms it could rank for, the SEO work has a payback period you can actually calculate.
Yes, where the work requires it and with your agreement. Much of Shopify SEO is in the theme: heading structure, template markup, schema, image handling and script loading. We work on a duplicate theme and publish only after review, so nothing changes on a live store without somebody deciding it should.
We will tell you what each one costs in page weight and what it appears to be doing, and recommend removals. The decision stays with you, because an app that is slowing the site may also be running something the business depends on. App bloat is one of the most common causes of poor mobile performance on Shopify.
Core Web Vitals are Google's measures of loading, interactivity and visual stability: Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift. They are a ranking input, and a modest one compared with relevance. They matter far more for conversion: mobile shoppers abandon slow pages regardless of where the page ranked.
Most commonly because several pages are competing for the same term, because the collection has no content beyond a grid of products, or because pagination and filters have created dozens of near-identical URLs. Google has to pick one and often picks none. Deciding which URL owns which term is usually the single highest-value change.
Yes, and it is the most commonly skipped work in eCommerce SEO. Descriptions taken from a supplier feed appear on every retailer selling that product, so there is nothing to distinguish your page. You do not need a thousand words: you need the specifics, the use case and the answers buyers actually ask, written once per product that matters.
You need content that answers what buyers ask before they buy, and a blog is one way to hold it. What does not work is a blog with no plan: articles that rank for nothing commercial, are not linked to the pages that sell, and exist because somebody said to post weekly. Content earns its place when it supports a collection you want to rank.
Structured data is markup that tells Google what a page contains: that this is a product, this is its price, this is its rating and availability. It matters because it decides whether your result in search shows those things. Two results side by side, one with a price and stars and one without, do not get the same click-through rate.
Yes, and it is the moment SEO risk is highest. A migration without a mapped redirect plan loses rankings that took years to build, and the loss shows up weeks later when it is hardest to diagnose. The work is URL mapping, redirect testing, structured data parity and a monitored launch rather than a switch flipped on a Friday.
Yes, through Google Ads and Meta, and it is often worth running alongside SEO rather than after it. Paid search tells you within weeks which terms convert, which is the fastest way to decide what the organic work should target. Ranking for a term that does not convert is a slow way to learn the same thing.
By commercial intent first and volume second. A term with 500 searches from people ready to buy is worth more than one with 50,000 from people researching. We also weigh difficulty against the store's current authority, because targeting terms the domain cannot reach this year produces a report full of movement and no revenue.
Answer engine optimisation is being the source an AI assistant quotes when somebody asks it a question. Assistants lift passages that state things plainly: definitions, specifications, comparisons and direct answers. A product page written as marketing copy has nothing quotable in it, and increasingly that is a channel rather than a curiosity.
No, and neither can anybody else, because nobody controls Google's index. What we commit to is the work, the reporting and an honest read of whether it is producing revenue. An agency guaranteeing a position is either targeting terms nobody searches or planning to explain later why the guarantee did not apply.
Monthly: organic sessions and revenue by landing page, conversion rate by template, indexation and technical health, and what changed since last month with a written note on what was tried and what did not work. Rankings sit underneath as diagnosis, because a ranking that does not move revenue is not a result.
Yes. The theme changes, the content, the schema, the redirect map and the keyword strategy are all yours and stay in your store. There is no proprietary layer that stops working when the retainer does, and nothing is hosted anywhere but your own Shopify account.
Frequently, and it is usually the fastest arrangement. We specify the technical work and review it, your developer implements it in the theme they know. What needs agreeing at the start is who owns the release process, because SEO fixes that sit in a branch for six weeks are not fixes.
Often more than for a large one, because a small catalogue can realistically make every page good. The question is not catalogue size but whether people search for what you sell in language you can compete in. If the demand is there, a focused store with thirty excellent pages will beat a neglected one with three thousand.
Then the two need to be planned together rather than run as rivals. Marketplace and direct sales use different levers, different economics and different content, and a brand doing both should know which channel it wants a given customer to arrive through. We run both, which makes that conversation easier than it is between two agencies.
It can, if it is done badly. Pages stuffed with copy to satisfy a keyword tool convert worse, and traffic from terms with no buying intent lowers your conversion rate by arriving at all. Good eCommerce SEO targets commercial intent and puts content where it helps a decision rather than where it inflates a word count.
An audit, then foundations. Crawl and indexation problems, duplicate URLs, broken redirects, missing or invalid schema, and the speed issues in the theme and app stack. Content and link work start in month two, once the site can actually be indexed properly, because content published onto broken foundations is content Google may never read.
A conversation about the store and what it is trying to reach, then a technical and content audit. The audit stands on its own: it produces an ordered list of what is limiting organic performance, which is useful whether or not you go on to work with us.
About Growth & Advertising
On contribution after fees, ad spend, returns and storage, rather than on the ROAS a platform reports about itself. Every platform measures the part of the journey it can see, which is why a campaign can look efficient in the dashboard and unprofitable in the accounts.
Yes, and deliberately together. They compete for the same queries and the same page, and a brand that buys traffic it already ranks for is paying for something it owns.
Sponsored ads and programmatic display sit in this family because they are advertising, and they are run on the channel page for the marketplace they belong to. The economics are the same question: what did the spend contribute after everything it costs.
Paid changes show within a cycle or two of the buying window. Organic and conversion work is slower and compounds, which is why they are planned together rather than sequentially.
A Google Ads agency runs the account that decides what a brand pays to be found for: campaign structure, keyword and audience targeting, the product feed behind Shopping, bidding and budget allocation, conversion tracking, ad creative, and the weekly work of reading search terms and cutting waste. The strategic part is deciding what to optimise towards; the automation handles the arithmetic.
Agency fees are usually a percentage of spend or a flat monthly retainer, and ours is quoted after an audit rather than from a rate card, because a £10,000 account with one campaign and a £10,000 account with forty are not the same job. The fee is separate from the ad budget, which is paid to Google directly from your own account.
There is no universal figure, and an agency quoting one is describing its other clients. A 3x return is excellent on a product with 70% gross margin and loses money on one with 25%. The useful target is set from your contribution margin, which is why we ask for margin data before quoting a target rather than after missing one.
Waste reduction shows in the first two to four weeks, because cutting spend on queries that never convert is immediate. Structural and feed changes need four to eight weeks for the bidding algorithms to relearn. Anything that changes the bid strategy resets the learning period, which is why good management does not change strategies weekly.
Yes. The account stays in your name and your billing, we work in it through manager account access you grant and can revoke in a click, and the campaigns, audiences, conversion history and feed are yours. Any agency that insists on running your spend through their own account is holding your data hostage.
Performance Max is a campaign type that spends across Search, Shopping, Display, YouTube, Gmail and Discover from a single budget, with Google deciding the mix. It works well for eCommerce with a good feed and clear audience signals. It works badly when it is given no structure, because it will find the cheapest conversions available, and the cheapest conversions are usually people who were going to buy anyway.
Standard Shopping shows products in search results and gives you visibility and control over product groups and bids. Performance Max includes Shopping inventory but adds every other Google surface and takes most of the control in exchange for automation. Many accounts run both, with Shopping holding the products that need protecting and PMax handling breadth.
Most commonly missing or invalid GTINs, price or availability mismatches between the feed and the landing page, missing shipping or returns policies, image requirements, or restricted content in a title or description. Disapprovals are usually a data problem rather than a policy dispute, and clearing them is the fastest available win on most Shopping accounts.
Yes, and on eCommerce accounts it is frequently the highest-value work available. The feed decides which searches your products are eligible for at all, and most feeds are a raw export from a store platform with titles written for a product page rather than for search. Bid changes cannot fix a title that does not contain the words people type.
Usually yes, and less than most accounts do. Brand terms are cheap and convert well, which makes them flattering in a report, and some of that traffic would have arrived free. Bid on brand where competitors or resellers are bidding on you, or where the organic result does not own the page, and measure it separately so it does not inflate the account's apparent performance.
Google Ads buys position immediately and stops the day you stop paying. SEO earns position slowly and keeps it. They work best together: paid search tells you within weeks which terms actually convert, which is the fastest way to decide what the organic work should target, and organic reduces how much you have to pay for the same click over time.
Almost always work with the existing account. Conversion history is an asset the bidding algorithms use, and starting from a blank account throws it away along with any audiences you have built. We rebuild inside the account instead, which is slower to look tidy and much faster to perform.
We check it before trusting it. Conversion actions, values, deduplication between Google and analytics, enhanced conversions, and consent handling, reconciled against the store's own revenue. It is common to find an account reporting twice the orders the shop recorded, and every optimisation made on top of that was made towards the wrong number.
It varies far too much by position, query type and industry for a benchmark to be useful. What matters is the direction and the comparison against your own account: a rising click-through rate on the same keywords at the same position means the ad copy is working. An absolute number quoted without a query type behind it is not information.
Yes, and it is usually worth having both read together. Google captures demand that already exists; Meta creates it. Measured in isolation each one will claim conversions the other contributed to, so blended reporting across both, and across marketplaces where a brand also sells, is the only view that reflects what actually happened.
Enough for the account to gather data faster than the market changes, which in practice means enough clicks per campaign per week to read a result rather than a coincidence. Below that, the honest advice is usually to spend on fewer keywords rather than to spread thinly, and we will say so rather than take a management fee on a budget that cannot learn.
Amazon, Walmart and TikTok advertising are separate platforms with their own auctions and their own agencies inside ours. Where a brand sells on Google and a marketplace at once, the useful question is which channel should own which product, because the same catalogue competing with itself across channels raises the cost of every sale.
An audit, then the waste. Negative keywords, disapproved products, campaigns competing with each other for the same query, and product groups that cannot be profitable at any conversion rate. Structural rebuilding comes second, because a rebuild on top of broken tracking or a broken feed is a rebuild that has to be redone.
Monthly reporting with a written note on what was tried, including what did not work, and a call whenever something needs a decision rather than on a fixed schedule that fills time. Weekly figures are available on request and are usually noise: the auction moves enough day to day that reading it weekly invites changes that undo each other.
No, because nobody controls the auction, competitors' budgets or Google's algorithm changes. What we commit to is the work, honest measurement against contribution margin, and telling you when the problem is the product, the price or the page rather than the advertising. Guaranteed ROAS is a sales technique, not a service.
That usually means the account is efficient and the ceiling is elsewhere: the catalogue, the price, the landing page, or a target so tight the automation refuses to bid on anything new. Growth almost always costs efficiency at the margin, so the real question is what return you are willing to accept on the next unit of spend, which is a business decision rather than an advertising one.
Yes, and it is the usual starting point. The audit covers structure, tracking accuracy, wasted spend, feed health and the things being counted as conversions that are not. It produces an ordered list of what is losing money, which is useful whether or not you go on to work with us.
A price the market will not pay, a product page that does not convert, out of stock inventory, or margins too thin to fund a click. Advertising buys attention. It cannot make an unconvincing offer convincing, and an agency that takes the retainer without saying so is charging you to discover it slowly.
A conversation about the account, the catalogue and what the business needs the advertising to earn, then an audit of the account itself. We ask for margin data early, because a target set without it is a target that will be hit and still lose money.
A Meta ads agency runs the Facebook and Instagram account that decides who a brand pays to reach and what it shows them: campaign structure, the product catalogue behind dynamic ads, the conversion signal the platform optimises against, budget allocation between prospecting and retargeting, creative strategy and production, and the weekly work of reading results and cutting waste. The strategic part is deciding what to optimise towards. The automation handles the arithmetic.
Agency fees are usually a percentage of spend or a flat monthly retainer. Ours is quoted after an audit rather than from a rate card, because an account spending £10,000 a month on one product and an account spending the same across four hundred are not the same job. The fee is separate from the ad budget, which is paid to Meta from your own account.
For most consumer brands, yes, and for a reason that has nothing to do with the platform being fashionable: it is still the largest place to create demand for a product nobody is searching for yet. What has changed is what the work consists of. Targeting is largely automated, so the advantage has moved to creative volume, signal quality and the economics you are willing to accept. Brands that judge it on last-click alone usually conclude it stopped working, then find their search and marketplace performance falls with it.
Nothing, in practice. Meta is the company and Facebook is one of its placements, so the same account buys Facebook, Instagram, Messenger and the Audience Network. The words are used interchangeably because the rename happened faster than the vocabulary did, and an agency that runs Meta ads is running Facebook and Instagram ads.
Yes, from the same account and usually within the same campaigns, with placements chosen per creative. Instagram is not a separate channel with a separate plan; it is a set of placements that reward a different edit of the same idea. The commonest waste on Instagram is one square asset stretched across Reels and Stories it was never designed for.
Usually a large part of it. Most of what was lost is signal rather than demand: browser tracking misses a material share of events, and the platform can only optimise towards what it can see. Conversions API sending server-side events, correct parameters, deduplication against the pixel and honest purchase values recover much of it. What does not come back is the certainty of last-click attribution, which is why measurement moves to blended reporting and incrementality tests.
It should be sized to the traffic that exists rather than set to a ratio. Retargeting can only spend against an audience the prospecting created, so an account that over-funds it reports a flattering ROAS while paying repeatedly for people who were going to buy anyway. When retargeting frequency climbs and revenue does not, the money belongs at the top.
We handle the strategy, the iteration and the production through our media team, and we work with a brand's own creators and studio where they have one. What matters more than who makes it is that enough distinct angles are made: an account testing four variations of one idea is not testing, and it is the commonest reason a good account stops improving.
Waste usually comes out in the first month and shows up immediately. Structural changes need a fortnight or so of stable spend before the numbers mean anything, because every significant edit costs the campaign its learning. A fair read on whether the account is on a different trajectory is about ninety days, and anybody promising a verdict sooner is reading noise.
There is no number worth quoting, and an agency that gives you one before seeing your margins is guessing. The target follows from contribution margin, repeat rate and what the business needs the channel to do: a brand at 70% margin with strong repeat purchase can grow profitably at a ROAS that would bankrupt a brand at 25% with none. We work the target out from your economics and then report against it.
Yes. Some engagements are a full takeover, some are the Meta half of an account where the in-house team runs search, and some are an audit and a plan somebody else executes. What does not work is two teams editing the same account without one of them owning the structure, so we agree that first.
Always. We work inside your Business Manager with the access the work requires, and the account, the pixel, the catalogue and the audiences stay yours. If the engagement ends, nothing has to be rebuilt and no data leaves with us.
Campaign and budget management, creative planning and iteration, catalogue and product set maintenance, audience and exclusion work, event and tracking checks, weekly reads of what changed, and a monthly report on contribution margin rather than on platform ROAS. The unglamorous half is the half that compounds: exclusions kept current, catalogue errors cleared, learning phases not reset by needless edits.
Enough for the conversion event to happen often enough for the algorithm to learn from it, which for most eCommerce brands means a daily budget that can produce a handful of purchases a day rather than a handful a week. Below that, results swing so much that no test can be read and money is spent buying noise. The exact figure follows from the product's price and conversion rate, which is why an audit precedes any number.
More than most brands are producing, and the number matters less than the spread. Ten assets that are one idea in ten aspect ratios give the auction nothing to choose between; five genuinely different angles, each cut for the placements that matter, give it a real decision. For an account of any size we would expect a steady flow of new angles every month rather than a quarterly batch.
Yes, though the work is different enough to say so. Lead generation is judged on lead quality and what happens after the form, so the account has to be tied to the CRM and optimised towards qualified leads rather than submissions. Optimising towards raw form fills is the fastest way to buy a great cost per lead and no revenue.
Yes, and the early weeks look different. With no pixel history there is nothing to build lookalikes from, so the account starts broad, leans harder on creative, and optimises towards an earlier event until purchases are frequent enough to optimise on directly. The first month is buying information as much as revenue, and a plan that does not say so out loud is selling optimism.
From the evidence the business already has. Reviews and support tickets say which objection stops the purchase; search terms say what language buyers use; the account's own history says which angles have never been tried. Tests are queued against those rather than against a list of platform features, and one variable moves at a time so the answer means something.
Advantage+ is Meta's automated shopping campaign: it takes a broad audience, a creative pool and a budget, and makes the targeting and placement decisions itself. It works well with a clean conversion signal, a real catalogue and enough creative, and it wastes money quietly when the existing-customer budget cap is left wide open and the account has nothing to compare it against. We run it as part of an account rather than as the whole of it.
By sizing it to the audience that exists, excluding recent buyers, watching frequency, and reading the account blended rather than campaign by campaign. When a retargeting campaign reports a very high ROAS and total revenue has not moved, it is being paid for sales that were going to happen. A holdout test settles the argument in a fortnight.
Read access is enough for most of it and makes the work materially better: the catalogue, the real order values, the return rate and the repeat rate all live there, and none of them can be inferred accurately from the ad platform. Where a server-side event setup is involved we work with whoever owns the store's code rather than around them.
An audit in the first week, the waste stopped in the second, structure and tracking corrected through the third, and the creative pipeline started immediately because it has the longest lead time. Spend is not increased until the signal is trustworthy: scaling an account that is optimising towards the wrong event just buys more of the wrong thing.
There is usually a fortnight where performance is noisier, because significant structural edits reset a campaign's learning and that is the cost of changing anything. We sequence the work so the account is never rebuilt all at once, keep the best-performing campaigns running while their replacements learn, and say in advance which week is likely to look worst.
We only work with eCommerce brands, and we run the marketplaces as well as the ad accounts. That matters on Meta specifically because a Facebook or Instagram campaign frequently converts on Amazon rather than on the brand's own site, and an agency that cannot see the marketplace half will report that the campaign failed. Judging channels in isolation is the commonest way a growing brand cuts the spend that was working.
Yes. Most of our clients sell into several markets, and the account structure follows the economics rather than the map: markets with different margins, delivery costs or return rates are separated so one cannot hide inside another's average.
A monthly report built around contribution margin, blended acquisition cost and what changed, with the platform's own figures beside them as context. Plus a live view you can open whenever you like. Nothing is reported that nobody can act on, and the report says what did not work as plainly as what did.
It depends on the category, the price and the channel, and a benchmark from another brand's catalogue tells you very little. The useful number is your own, tracked per step and per device, and whether it moves.
Usually not. A redesign changes everything at once, which means nothing can be attributed and any gain is luck. Most of the money is in specific friction on specific steps.
Enough for a result to be distinguishable from noise, which depends on the current rate and the size of the effect. Where a store does not have it, we fix what is certainly wrong rather than run tests that cannot conclude.
The principle does, but the levers are different: on a marketplace the page is the listing and the controls are images, content and reviews. That work lives with the platform's own service.
Campaign structure, keyword and product targeting, bids and placements, budgets, search term harvesting and negation, creative for Sponsored Brands, and DSP where it applies. Plus the reporting that ties all of it to contribution rather than to the platform's own ROAS figure.
PPC usually means the sponsored ad types alone. This covers those plus DSP and the relationship between paid and organic: what a listing already earns for free, and where spend is buying something it would have had anyway.
Only at a certain scale. It has minimums and it buys audiences rather than keywords, which suits accounts with a catalogue and a retargeting pool worth reaching. We will say when it is not worth it.
The one your margin supports, which is a different number per product and is not the number the console suggests. We work it out from contribution after fees, returns and storage before setting any target.
No. Advertising buys traffic. The listing decides what happens to it. If the conversion rate is the problem we say so and fix that first, because spending into a poor listing is the most expensive way to discover it.
Buying display and video inventory automatically, targeted at defined audiences rather than at keywords. For a marketplace brand it usually means Amazon DSP, which adds Amazon's purchase data to the targeting and can place ads both on and off Amazon.
Sponsored Display sits inside the advertising console and is bought per campaign against products and audiences. DSP is a separate buying platform with wider inventory, deeper audience controls and minimum commitments.
Amazon DSP has one, and it varies by market and by whether the buying is managed or self-service. We will tell you the current position when we know which market and which route applies to you.
On incrementality where the data supports it, and on contribution after spend where it does not. Attributed return alone flatters retargeting, because some of those people were coming back anyway.
About Creative & Media
eCommerce design is the creative that sells a product inside a marketplace or store: main and gallery images, infographics, A+ or enhanced brand content, Brand Stores, packaging and advertising assets. It is judged on click-through rate and conversion rather than on appearance, because its whole purpose is to win a decision made in seconds against competitors on the same screen.
Not as a standalone service. We build and develop Shopify stores as part of Shopify work, and this page is about marketplace and eCommerce creative rather than web design. If you need a corporate website designed and built from scratch, an agency that specialises in that will serve you better than we would, and we would say so.
Because it is the only asset that competes before anybody reads anything. On a search results page a shopper sees your main image at roughly thumbnail size beside twenty alternatives, and that comparison decides the click. Every other asset on the listing only gets seen if the main image has already won.
Amazon requires the main image to show the product alone on a pure white background, filling most of the frame, with no text, logos, watermarks, borders or props that are not part of the product. Enforcement varies, and creative that ignores the rules risks suppression at exactly the moment it is working, which is the worst time to lose a listing.
As many as it takes to answer the questions that stop a purchase, which is usually six to nine. The number matters less than the sequence: size and scale, what is included, how it is used, what it is made of, how it compares, and who it is for. Filling slots with variations of the same photograph adds nothing.
A+ content, formerly Enhanced Brand Content, replaces the plain product description with designed modules of images and text. It is worth it where a product needs explanation, comparison or brand context to justify its price. It is worth less on a simple commodity item, and it is seen only by shoppers who scroll, so it should never carry information the gallery needs to convey.
Premium A+ adds larger modules, video, interactive comparison tables and hover features, and takes up more of the page. It is available to brands meeting Amazon's eligibility criteria, which change. The design discipline is the same either way: modules that answer objections in order, rather than a longer brochure.
The Brand Story module is a carousel that appears above A+ content on every listing a brand owns. It is one of the few places on Amazon to say something about the brand rather than the product, and it cross-links to your other listings, which makes it useful for range discovery as well as for positioning.
If you have more than a handful of products, yes. It is the only page on Amazon whose layout you control, it is where Sponsored Brands traffic lands, and it typically converts better than a product page because a visitor arriving there is exploring a range rather than comparing one item. Most Brand Stores are built once at launch and never revisited, which wastes the asset.
Yes, through Amazon's Manage Your Experiments where the brand is eligible, and through panel testing tools such as PickFu where it is not or where speed matters more than precision. Main image assumptions are wrong often enough that testing is not a refinement, it is how the decision should be made.
A full gallery and A+ set for one product typically takes two to three weeks including revisions, longer where new photography or video is needed. A catalogue is quicker per product after the first, because the brand system and the template decisions have already been made.
Both. We direct and produce studio, lifestyle and video assets, and we design with existing assets where a brand already has a library worth using. Reshooting a catalogue that already has good raw material is usually the wrong first spend.
Yes. Walmart has its own rich media formats and content quality scoring, eBay has its own listing structure, and TikTok Shop is video first. The brand system carries across, the specifications do not, and assets built once for Amazon and pasted elsewhere usually fail a requirement quietly.
Yes, and it is the better starting point when they exist. Where they do not, or where they were written for print and packaging rather than for a marketplace, we would rather agree the marketplace-facing rules first than produce assets that will be inconsistent with whatever comes next.
Each design service shows its price on its own page: listing images and A+ content per ASIN, main images per image, a Brand Store and each page beyond it, and brand guidelines in two levels. New photography or video is quoted against a defined scope. The useful comparison is against advertising: creative is a fixed cost that raises the conversion rate of every click you will ever buy, whereas advertising is a recurring cost that buys the clicks.
It changes click-through rate and conversion rate, which is where its effect appears. It cannot change demand for the product, its price competitiveness or whether it is in stock. Creative work is at its most valuable on a product with proven demand that is losing the comparison, and least valuable on one nobody is searching for.
You do. Source files, layered artwork and photography are handed over, and you keep them if the arrangement ends. Nothing is held back as leverage, which is worth asking any design supplier before starting rather than afterwards.
When performance slips or the category shifts, rather than on a schedule. Competitors update their images, seasonal context changes, and a main image that won a year ago can be losing quietly now. Monitoring the category and refreshing on decay is more useful than an annual redesign nobody measured.
Yes: Sponsored Brands assets, DSP and display, and paid social for Meta and TikTok. Building it from the same brand system as the listings matters more than it sounds, because a shopper who clicks an ad and does not recognise the product page has been given a reason to hesitate.
The products, any existing brand assets and photography, the ASINs or listings to work on, and an honest account of what you think is not working. Access to the listings is useful but not needed on day one; the category audit can be done from outside the account.
Yes, and it is a common starting point. Rejections are usually specification problems rather than judgement calls: background, framing, text on a main image, or a file that does not meet size requirements. They are quick to fix once identified and expensive to leave, since a suppressed image can take a listing with it.
Yes. Creative is taken on as a standalone engagement for brands managing their own accounts or working with another agency. It is useful for us if we can see the listing performance data, but it is not a condition.
A look at the category and your current listings against the competitors a shopper actually compares you with. That audit is where the specifics come from, and it usually makes the priority obvious: main image first, then the gallery, then rich content.
As many as the product needs to answer its questions, which is rarely the same number twice. Five covers a simple product; seven suits one with more to explain; ten is the full gallery for a product that needs a comparison and several uses shown. The tier is chosen by what the product has to explain, not by filling every slot.
Every image is delivered at 2000 by 2000 pixels, large enough for the zoom marketplaces offer on product images and sharp on any screen.
No. These prices are for design work, built from your product photography or from renders. Product photography is quoted separately.
Yes, where each platform's rules allow it. Tell us which channels the images are for when you order, and each set is checked against those platforms' image requirements before delivery.
The main image has to show the product itself on a pure white background, filling most of the frame, with nothing added that is not part of the product, such as text, badges or graphics. The detail varies by category, so each image is checked against the current requirements for its category before delivery.
Yes. Each image is priced on its own, so variants can be made for a split test. On Amazon, brands enrolled in Brand Registry can test main images on eligible listings with Manage Your Experiments.
Up to ten in one order, each priced per image.
No. The work starts from your product photography. Product photography is quoted separately.
Premium A+ uses larger, full width modules and adds interactive ones such as hotspots, carousels and Q&A, and it is available only to brands that meet Amazon's eligibility for it. Standard A+ is available to brand owners enrolled in Brand Registry.
A carousel that sits above the A+ content and introduces the brand, with links to other products in the range. One Brand Story can be applied across many ASINs, which is why it is priced for up to twenty.
Its job is conversion: it answers the questions still open when a shopper scrolls. The alt text and written text can be read by search engines, which is why every design carries both, but A+ is bought for the sale it helps close rather than as a ranking tactic.
Yes. Larger orders are discounted per design, as shown in the pricing on this page.
Brand owners enrolled in Amazon Brand Registry can create a Store, and Amazon does not charge for it. The price here is for its design and build.
Enough for a shopper to reach any product quickly: usually a home page and a page for each category. Each page beyond the home page is priced on its own, up to fifty.
Yes. A Store page is one of the destinations a Sponsored Brands campaign can use, which is why landing pages are designed with the campaigns in mind.
Essential sets the core: logo design, colour palette, typography and starter guidelines. Professional adds the research and strategy behind the brand, advanced logo guidelines, extended brand elements and the system applied across the brand's channels.
They keep every listing, A+ page, Store and ad consistent, so a shopper who sees an ad recognises the page they land on. They are also what lets a new designer or agency produce work that fits from the first day.
Yes. Where a logo exists and works, the guidelines are built around it, and the logo work is limited to its variants and the rules for using it.
About Commerce Services
Both, and each service says which it usually is. Listing optimisation, keyword research and unit economics are normally projects. eCommerce SEO is ongoing, because search visibility is built and maintained over months.
Yes. These are capabilities rather than channels, and they are applied to marketplaces and storefronts we do not manage as often as to ones we do.
Compliance, reinstatement and cost recovery now sit in Compliance & Account Health, because they are the same kind of work: a case assembled and filed with a marketplace or a registry. BPO Services sits in Strategic Partnerships, because an outsourced team is an ongoing arrangement rather than a project.
You do. Files, keyword sets, models and the listing content stay with you, including if we stop working together.
Marketplace search engines rank listings on conversion and sales velocity inside their own platform, so the terms that matter are the ones shoppers use at the point of buying. Google ranks pages on relevance and authority across the open web, which brings in research and comparison queries a marketplace never sees. The same product needs both, worked separately.
Properly once, then maintained. Demand shifts with seasons, competitors and new products, so the map is reviewed rather than rebuilt. A full redo is warranted when the catalogue changes materially or when you enter a new market.
Both. Third party volumes are estimates and are treated as one input, while your own account data is the only source that knows what converted rather than what was searched. Where the two disagree, your data wins.
A project. The research produces a map, and the map is then used by whoever does the listing, content and advertising work. Where we run those channels for a brand, keeping the map current is part of that engagement rather than a separate one.
A clustered term list with the intent behind each group, estimated demand, a realistic difficulty against your domain, one owning page named per cluster, and a written note of where no suitable page exists yet.
Technical fixes can move things within weeks because they change what is already indexed. Content and authority take months. Anybody promising ranking in a fixed number of weeks is describing a coincidence rather than a plan.
Yes. The discipline is the same and the obstacles differ by platform: Shopify's collection and variant URLs, WooCommerce's plugin markup, and whatever a custom build does with faceted navigation.
No. Amazon and the other marketplaces rank listings on conversion and sales velocity inside their own search engine. This is about a store's visibility in Google, which is a different algorithm and a different set of levers. We run both, separately.
Usually yes, and for different jobs. Paid buys demand now. Organic compounds and lowers what you have to buy later. Most accounts should run both and know how much overlap they are paying for.
Indexing of new terms usually takes days. Ranking movement depends on sales accumulating against those terms, so it is weeks rather than days, and a listing with very little traffic takes longer because there is less evidence for the platform to read.
Amazon, Walmart, eBay and TikTok Shop most often, plus Shopify product pages where the same catalogue also has to work in Google. The fields and the rules differ by platform, so copy is written per platform rather than copied across.
Only where that is the right answer. Most catalogues have a small number of listings carrying most of the revenue and a long tail where the cheapest gain is backend fields and images. The audit decides where the work goes.
No. Marketplace listings rank inside the marketplace's own search engine on conversion and sales velocity. eCommerce SEO is a store's visibility in Google, a different algorithm with different levers. We run both, separately.
Yes. Image work sits with the creative team and copy with the listing team, and both work to one brief, because a main image that wins the click and a bullet that answers the next question are one decision split in two.
What one unit of a product earns after every cost of selling it on a given channel: the platform's fees, fulfilment, storage, returns, the advertising that produced the sale and the cost of the unit itself. The result is a contribution margin, per SKU and per channel.
No. The platform's reports cover the platform's own fees and leave out landed cost, returns handling, and most of what advertising actually costs per sale. The model puts all of it in one place.
Managed clients get it as part of the work, because advertising and pricing decisions are made against it. It is also sold on its own to brands that run their own channels and want the number.
Any number. The model is built to be refreshed from reports rather than filled in by hand, so a catalogue of a few hundred products is no harder to keep current than one of ten.
About Compliance & Account Health
No. The marketplace decides who sells on it and the trademark office decides what is registered. What we control is whether the case is complete, accurate and submitted the way the process expects, which is the part most failed cases never had.
Amazon most often, because that is where most of the demand is, and Walmart, eBay and TikTok Shop, which have their own policies and their own enforcement.
Account management is the month-to-month running of a channel. This is the work when something has stopped: an account, a listing or a brand that needs a case made for it. Managed clients get it as part of the service; anybody can buy it on its own.
No, and nobody can. Reinstatement is the marketplace's decision, made against its own policies and its own record of the account. What can be controlled is the quality of the case, the evidence behind it and whether it addresses the real cause. That is the part that is work rather than chance.
The platform sets the pace, and it varies by action type, by how complete the first submission is, and by how busy the review teams are. We do not quote a turnaround because we do not control one. What we do control is avoiding the delay an incomplete case creates.
A response that answers the notice rather than the cause, sent without the documentation the policy asks for. Both are fixable before submission, which is the reason the diagnosis comes first and takes as long as it takes.
Yes, and the work has the same shape. Establish which requirement applies, assemble the documentation against it, submit and follow up. Approval is the platform's to give, and we will tell you when the paperwork you have is not going to be enough.
Amazon most often, because its policy surface is the largest, plus Walmart, eBay and TikTok Shop. Each has its own process and its own view of what documentation counts as evidence.
No. This is marketplace policy work. Where a matter needs a lawyer, for example a rights dispute that has moved beyond the platform's own process, we say so rather than continuing past the point where we are useful.
No, and nobody honest can. The marketplace decides. What we control is whether the case is complete, accurate and submitted the way the process expects.
The platform sets the timeline and it varies by case type, from days to months. We can say how long each step on our side takes and how quickly we respond when the platform comes back.
That is the most common starting point. A rejected plan usually answered the notice rather than the cause, and the next submission has to be different rather than longer.
Yes. Each has its own policies and its own process, and the work is the same shape: cause, case, submission, follow-up.
No. The marketplace decides. What we control is whether the exact reason was found and the case that answers it was complete and filed where it will be read.
A retraction from the complainant is the fastest route, and we pursue it. Where none comes, a counter is filed with the evidence that the complaint does not hold.
Yes. Each category has its own document requirements and its own form of application, and most refusals come from the documents rather than the product.
Usually the reviews stay with the ASIN. Rank and the Buy Box can slip while the listing is down, which is why we watch both after reinstatement rather than closing the case at 'live'.
Discrepancies between what a platform's own reports show and what its own policies commit to. Lost or damaged inventory, incorrect fees, returns never restocked, and refunds issued without a return. What is claimable, and how long you have to claim it, are both set by the platform and both change.
Read access to reporting is enough to run the audit. Submitting claims needs the ability to open cases, which can be a limited user rather than full account access. We work at whatever level you are comfortable granting and say what it rules out.
Terms are agreed before any work starts, and they depend on the size of the account and the state of its reporting. Nothing is filed on your account before you have seen what the audit found and agreed how the work is scoped.
Yes, and it is the platform's limit rather than ours. Each marketplace enforces a window after the event inside which a claim can be raised, which is why discrepancies found late are sometimes not recoverable. It is the argument for auditing on a rhythm rather than once.
Wherever a platform publishes a policy committing it to reimburse sellers and gives enough reporting to check against, which in practice means Amazon most of the time. Where a channel does not report in enough detail to audit properly, we say so rather than charging to look.
Months rather than weeks, and the number depends on the office, the classes and whether the application is examined cleanly or draws an objection. Any specific figure quoted at the start is a guess, so we describe the stages and tell you where yours has reached.
No, but most brand protection tools on the major marketplaces require one, Amazon Brand Registry included. Selling is possible without it. Defending what you sell largely is not.
No. This is application preparation and process management. Where a matter requires a licensed attorney, including jurisdictions that demand local representation and disputes that move beyond examination, we tell you rather than carrying on.
The clearance search is where that surfaces, which is why it comes before anything else. Depending on how close it is, the options are a different mark, narrower classes, or filing in the knowledge that objection is a real risk. We set out which applies rather than filing and hoping.
It depends on where you sell and where local representation is required. We will tell you which of your markets can be handled directly and which need local counsel, before any work starts rather than partway through.
Amazon requires a mark at a stage it accepts, and which stages qualify has changed more than once. We check the requirement as it stands when you apply. If you have nothing filed at all, the registration is the first piece of work and it is slow, which is the argument for starting early.
Amazon decides, and the wait depends on its queue and on whether the application arrived complete. A complete application avoids the round trips, and that is the only part of the timeline anybody outside Amazon can influence.
Enhanced content, the brand advertising formats, the Store, brand analytics, and the tools for reporting counterfeits and hijacked listings. Most of it goes unused by brands who enrolled and stopped there.
It gives you the process for reporting them and the evidence route Amazon expects. It does not remove anybody automatically. Amazon assesses each report, so what you control is how well the report is made and how quickly it is raised.
No. Compliance is about your own account and the policy actions taken against it. Brand Registry is about protecting your brand from other sellers and gaining access to the tools reserved for brand owners. The two meet where a rights complaint is involved.
About Strategic Partnerships
A retainer buys a team to run a channel you already own. These are commercial models: a marketplace programme with its own approval and economics, an operating model where we build and run the route to market while the inventory stays yours until it sells, and an outsourced team working inside your operation.
No. In the retail growth partnership you keep ownership until the point of sale, which is the line between that model and wholesale. Amazon Renewed has its own supply requirements, which we work through with you.
The retail growth partnership is designed so you do not: no separate 3P or DTC organisation, no new team, and your wholesale relationships and logistics keep working as they do.
A conversation about the catalogue and the market. For Renewed, whether the category and supply meet the programme's requirements. For the partnership, which parts of the range have retail demand worth reaching.
Both. The programme is the same and the work around it is not: the EU adds VAT registration per market, WEEE, CE marking, customs and multi-language service, while the US adds almost none of that and a different sourcing problem instead. Tell us which market you are entering and the plan changes accordingly.
It depends far more on documentation than on Amazon. Applications stall on supplier invoices that do not meet the 90 day rule and on quality manuals written after the fact. Prepared properly first, approval is a matter of weeks rather than an open ended wait, and new entrants without help commonly spend months on it.
Yes, and a rejection is usually about evidence rather than eligibility. The application is rebuilt from what Amazon actually asked for, the gap that caused the refusal is closed, and it goes back in. What does not work is resubmitting the same package and hoping for a different reviewer.
The physical SIM. From the iPhone 14 onwards the US variants are eSIM only with no SIM tray, and Amazon's EU Renewed listing check flags the missing physical SIM and refuses the listing. International, EU, UK and Hong Kong variants keep the tray. It is the single most common way a new EU seller ends up with capital in stock that cannot be sold, and it is avoided by verifying the variant before purchase.
Established networks in Hong Kong, the UAE and inside the EU: electronics wholesalers, carrier trade-in aggregators, OEM authorised resellers and B2B liquidation. Which one suits you depends on the market you are selling into, the generation you want and how fast you need it, and we negotiate the pricing rather than passing on a list.
Not every one, but more than one, and which ones depends on where the stock sits and where it sells. Pan-EU fulfilment puts inventory in several countries and creates the obligation in each. We map it against what you are actually doing rather than registering everywhere by default.
The argument for Renewed is not the margin on a single unit, it is who you are competing with. Enrolment is restricted, so most Renewed categories have a handful of active sellers rather than hundreds, and the Renewed Guarantee means the listing converts close to a new one. The barrier that makes it slow to enter is the same barrier that protects you afterwards.
It is an operating model for selling a brand's products through retail channels without the brand building the function that does it. Xcelerate identifies the retail growth opportunity, builds the commercial plan and runs the selling across consumer and business demand. The brand keeps ownership of its inventory until the point of sale, and the products are sold under the brand's own name at retail pricing.
In wholesale, a buyer purchases inventory from you, takes ownership of it and resells it on their own terms. Here nobody buys your stock. You retain ownership until it sells, and what we provide is the retail growth strategy, channel execution, demand planning, forecasting and operating support that turns that inventory into retail revenue.
You do, until the point of sale. That is the defining feature of the model and the clearest difference from a wholesale arrangement.
Either yours or ours, depending on the channel and the operating model agreed. Both are available and which applies is settled before anything goes live, because it determines where the account, the data and the daily operations sit.
No. The model is designed so that a brand can take part without restructuring its existing wholesale relationships or logistics. It runs alongside what you already have rather than replacing it.
No. That is a large part of the point. There is no separate internal organisation to stand up, no new team to recruit and no new operating function to create in order to reach these channels.
Both consumers and business buyers. The two behave differently in volume, pricing structure and buying cycle, so they are planned and served separately rather than treated as one audience at different order sizes.
It is recurring rather than a one-off plan. We forecast expected retail demand and plan availability against it, and revise both as the channels produce real numbers, so stock and demand stay aligned as the partnership grows.
Yes. Products are sold under your brand at retail pricing, which is what keeps your pricing position and brand presentation intact as the reach expands.
A brand with products that could reach more of the retail market than their current route to market allows, and which would rather not build an internal channel organisation to get there. If you already run these channels well yourself, a management engagement is the more useful conversation.
A conversation about your catalogue, your current route to market and where you think the reach is missing. We will tell you which parts of the range we think have retail demand and which we do not, before anything is agreed.
eCommerce BPO is outsourced operational capacity for online retail: named people who run marketplace operations, advertising, catalogue work, customer support and reporting on your account and your systems. It differs from a call centre in continuity, and from freelancing in that the people are managed, trained and covered for absence.
A virtual assistant is usually one contractor you manage directly. BPO is a managed team: people matched to skills, trained on documented process, supervised, and covered when somebody is unavailable. The distinction matters most when you are away, ill or busy, which is exactly when a directly managed contractor stops being useful.
Marketplace specialists, advertising specialists, catalogue and data staff, customer support, inventory and order operations, reporting analysts, designers, and account managers who coordinate them. Most engagements are a mix of part-time skills rather than one full-time role, because that is usually what the work actually is.
Yes. Named staff rather than a rotating pool, with a nominated cover for absence. This is the point of the arrangement: an outsourced specialist becomes more valuable every month they learn your catalogue, and that value resets to zero each time the person changes.
It is priced on the roles and hours required, and it is quoted after scoping rather than from a rate card. The comparison worth making is against the fully loaded cost of hiring the same skills directly, including recruitment, management time and the months before a new hire is productive.
Usually, and cost is rarely the main reason brands do it. The stronger arguments are speed, since a trained team starts in weeks rather than months, and divisibility, since you can buy a quarter of an advertising specialist and a half of a catalogue specialist, which no employment contract offers.
We have offices in the United States and Pakistan, and teams are assigned with your working hours in mind, including a defined overlap with your own team. Overlap matters more than location: work that needs a question answered stalls without it.
Agreed at the start, with a stated overlap with your own working day. Marketplace support has response time requirements that feed seller metrics, so where support is in scope the coverage is planned around those rather than around convenience.
Through user permissions you grant inside your own accounts, scoped to what the role requires and revocable in a click. Nobody works through shared logins or through our own accounts, and access is removed when a person leaves the engagement.
Least-privilege access, no shared credentials, no data taken outside the systems it belongs to, and access removed promptly when roles change. Access review should be a condition of any outsourcing arrangement, and it is reasonable to ask for it in writing before starting rather than after.
Cover is nominated in advance and the documentation is maintained throughout, which is why the documentation step is not optional. A handover from written process takes days; a handover from somebody's memory takes weeks and loses things nobody notices for a month.
Typically two to four weeks: scoping, documentation, then assignment and training. Documentation is what varies. A brand with written procedures starts quickly, and one where the process lives in one person's head spends the first fortnight getting it out.
No, and most brands do not. Building the documentation is part of the onboarding and one of the more useful by-products, because you keep it. Several clients have said the documentation was worth more than the first months of the service, since it made their own team replaceable in the healthy sense.
Yes. Your Seller Central, your helpdesk, your PIM, your project tool. Working in your systems rather than exporting into ours is what keeps the arrangement reversible: if it ends, the work and its history stay where they always were.
A supervised first period where output is reviewed before it goes out, then agreed quality measures for the ongoing work: error rates on catalogue changes, response times on support, adherence to the documented process. Reviewed with you rather than reported at you.
That is the usual and the better path. One role, one channel, a defined scope, and expansion once the process is documented and the first person is productive. Brands that outsource six functions at once spend the first months managing the arrangement instead of the work.
We do both, and they are different arrangements. Agency management means we own the outcome and the strategy. BPO means you own both and we provide the hands. Some brands use us for agency work on one channel and BPO capacity on another, which is a sensible way to keep strategy in house where it matters most.
Either you, if you want hands and already have the management capacity, or a named account manager on our side if you would rather outsource the coordination as well. The second costs more and is usually the right answer when the scope covers three or more functions.
Some does not, and we would rather say so at scoping. Work requiring constant judgement calls on brand, pricing or supplier relationships belongs in house. Repetitive work with a documentable process outsources well. Most roles are a mix, and the useful exercise is separating the two rather than moving the whole job.
English is the default. Other languages depend on availability and are confirmed at scoping rather than promised in advance, because staffing a language badly is worse than not offering it.
With notice, and with everything you need to continue: the documentation, the account access that was always yours, and a handover. Nothing is held in systems you cannot reach, and the process documentation built during onboarding is yours whether or not the arrangement continues.
A scoping conversation about what the work actually is, in hours rather than in job titles. That usually reveals whether you need one person or three part-time skills, and it is worth having even if you go on to hire directly instead.
From the blog
The questions each article answers. The heading opens the article.
Sponsored Products, almost always. It is closest to the purchase, easiest to measure, and it produces the search term data every other campaign type depends on.
Usually only for defending your own brand terms. The format is built to sell a range, so with two or three products the headline banner has little to carry.
Reaching shoppers away from the search results: on competitor product pages, and back to people who viewed your listing and did not buy. It is prospecting and retargeting rather than harvesting demand.
No. Sponsored Products catches shoppers at the end of a search; Sponsored Display reaches them earlier, on competitor product pages, off Amazon and after they have viewed your listing without buying. Expect a higher ACoS, judge it over a longer window, and hold it to a target set for that job rather than to the account average.
Both, doing different jobs. Automatic campaigns find search terms you would not have thought of. Manual campaigns let you set a bid for the ones that convert. Running only automatic means never controlling what you pay; running only manual means never discovering anything.
Enough to keep discovering, which is usually a small and steady share rather than a large one. Automatic is a research budget: once its useful terms have been promoted to manual, most of what is left is the waste it also finds.
Yes, and it is the single most important thing you do to them. Without negatives an automatic campaign keeps paying for terms you have already promoted to manual, so you bid against yourself, and it keeps buying the irrelevant matches it turns up.
Close match, loose match, substitutes and complements. The first two match your product to searches; substitutes and complements show it on other products' detail pages, and usually deserve a lower bid than the search groups. Bidding on them separately takes about ten minutes and stops one expensive group hiding inside an average.
It is the same idea applied to a different search engine. You bid on what a shopper types, you pay when they click, and placement depends on both bid and relevance. What differs is intent: an Amazon search is usually much closer to a purchase.
Amazon knows what was bought, not just what was clicked, so attribution is far more direct. There is also no landing page to design: the product page is the destination, and it does most of the converting.
Usually yes, because paid placements sit above and among organic results, and competitors can buy visibility against your own listings. Defending the terms you already rank for is often the cheapest advertising you will do.
Sponsored Products stop serving for that listing. Advertising cannot rescue a listing that is not winning the Buy Box, so that is the first thing to check when a campaign suddenly stops spending.
Reduce rather than pause. Running out mid campaign wastes the rank you paid for, but pausing outright loses placement that is expensive to buy back. Lowering bids as cover shortens is usually the better trade.
There is no single best one, and any answer that does not ask what stage your product is at is guessing. A launch runs above target on purpose, a mature product runs at it, and one being wound down runs below it. The same account can be doing all three at once.
Give a change enough clicks to be readable, not enough days. A keyword with thirty clicks says something; a keyword with four says nothing regardless of how long it took to get them. Most accounts change strategy far too often and read noise as a result.
You should. A launch and a mature hero product in the same account want opposite things, and forcing them to share one ACoS target guarantees the launch is starved and the hero is overpaid for.
Three questions usually settle it. Does the product have organic rank on the terms that matter? Is there room to spend more before the returns collapse? Is a competitor bidding on it? No rank means launch. Rank with room to spend means scale. Rank without room, with competitors present, means defend. Rank without room and nothing to protect means harvest.
It depends on catalogue size and how much is automated, but twenty to thirty hours a week on a single account is the point at which most brands are doing mechanical work by hand. Report pulling, formatting and threshold checking are usually the largest share and require no decisions, so they are the first to remove.
When the account generates more decisions each week than there is time to make properly. The practical test is the search term report: if it is being read weekly or less, the account is being maintained rather than optimised, and that gap widens on its own.
No. Automation applies rules you wrote, inside limits you set, against targets you chose. Control moves from making every adjustment to deciding what the adjustments are for. What does lose control is a system nobody has reviewed in a month, so the review has to stay a real task in the week.
What the account is for, how campaigns are structured, and the reading of search terms. The first cannot be inferred, the second decides what every later decision can see, and the third is where new phrasing and competitor movement are actually noticed. Flagging search terms is mechanical; reading them is not.
Automation solves a capacity problem. If you are overriding the tool most weeks, several products want different things at once, and the questions have become why the account behaves as it does rather than what to do next, that is a judgement problem and more automation will not answer it.
Within limits you set, yes. Unbounded, no. An automation that can raise a bid without a ceiling will eventually find a term that converts once at a price you would never have agreed to, and keep buying it. A minimum click count, a bid ceiling and a maximum change per adjustment are the three limits.
Reporting and alerting, before any bidding. Collecting and formatting the numbers takes hours and involves no judgement, so automating it costs nothing and buys back the time you need for the parts that do.
Almost always because the traffic is arriving at a listing that cannot convert it, not because the bids are wrong. Advertising buys the visit; the main image, title, price, reviews and availability decide what happens next. Check the conversion rate on the terms that are spending before touching a bid.
Enough that a zero means something. If the listing converts at ten percent, ten to fifteen clicks with no sale is a reasonable line. Below that you are reading noise, and cutting a keyword on four clicks removes a term you never measured.
Only after ruling out the listing. Lowering bids on a page that does not convert buys fewer visits at the same poor rate, which improves the number without fixing anything. If conversion is the problem, a bid change hides it for a month.
It means the ad is doing its job and the page is not. A strong click through rate with a weak conversion rate is the clearest signal there is that the problem sits on the detail page rather than in the campaign.
A drop in impressions on one day is almost always a listing change rather than the bids: a main image that failed a check, an attribute that moved the product out of its category, a split variation, or a price that tripped a pricing alert. Find the day the impressions fell and read the listing's change history for that day. A decline over weeks is the auction, and that is the bids.
Because the listing is now being compared against something new on the results page: a competitor's deal, a lower price beside yours, a run of poor reviews, or a season that has passed. Read the search terms report for the terms that used to convert and open the results page for each as a customer; the answer is usually beside your ad, not inside the campaign.
Amazon has not announced a change to the auction. What sellers reported on the Seller Forums in September 2026 is more advertising slots on the results page, more spend landing on product page placements, which convert worse than top of search, and a drop several of them dated to the end of August. Those are reports, not a rule. The way to know for your own account is the placement report and the search term report compared thirty days before and after the change you noticed.
The page. A campaign cannot lose conversion on its own while the listing keeps it, so a fall in both at once means the detail page is being compared against something new or has lost relevance: a title without context, thin bullets or images, missing A+ content, reviews saying what the page does not, a price that has been undercut. Read the page as a buyer and the reviews as a critic before touching a bid.
Both, separately, because they fail for different reasons. Branded terms falling usually means competitors bidding on your brand or your own broad and automatic campaigns taking the click the organic listing would have earned; add the brand terms as negatives in the generic campaigns and run them in one of their own. Generic terms falling means relevance or the page. The search term report split by term type shows which it is.
For the ninety day change log: every bid rule, budget change, negative list edit and placement modifier, with dates. Then for the search term report and the placement report compared before and after the drop, and for the split between a cost per click that rose and a conversion rate that fell. An agency that runs the account should be able to show all of that in a day; one that cannot is running rules it has not read in a while.
Amazon says it does not bill for traffic it identifies as invalid and credits it after review, and a case that names the campaign, the dates and the click pattern is how to ask. A seller on the September 2026 thread reported invalid traffic confirmed by Amazon and still uncredited after five weeks, so treat the credit as a claim to chase rather than a refund to expect.
Amazon PPC is a paid advertising model where sellers pay only when shoppers click on their ads. It helps increase product visibility, drive targeted traffic, and boost sales through strategic amazon ppc management.
Amazon PPC is essential because it improves product ranking, increases visibility, and accelerates sales growth. With the right ppc advertising services or strategy, sellers can outperform competitors and scale faster.
You can reduce ACoS by optimizing keywords, using negative targeting, improving product listings, and adjusting bids based on performance. Effective amazon ppc management ensures better budget control and higher profitability.
You can manage PPC yourself initially, but working with an experienced amazon ppc agency provides advanced optimization, better scaling strategies, and improved ROI, especially in competitive markets.
For a handful of terms that carry real volume, yes, because a keyword with its own budget cannot be starved by a neighbour. Across a whole account it produces hundreds of campaigns nobody can review, which costs more than it returns.
Rarely, and less often than it is recommended. Amazon buying intent is spread through the day, the data thins out quickly once split by hour, and the saving is usually smaller than the attention it takes to maintain.
The best strategy is a hybrid approach. It combines internal control, an Amazon advertising agency for execution, and automation tools. This reduces manual work and keeps optimization consistent.
If you lack internal capacity, working with an amazon advertising agency is more efficient. However, businesses with strong internal resources can benefit from amazon hiring and building a dedicated team.
Effective amazon ppc management improves visibility, conversion rates, and campaign efficiency, which directly increases Amazon order volume and overall revenue.
Software is not mandatory but highly beneficial. It helps automate repetitive tasks, improve reporting, and optimize campaigns across any online advertising platform.
ROAS differs depending on the stage of your business growth. Most brands aim to balance profit and growth, not just high returns.
Spend is the wrong trigger on its own. A better one is whether the decisions have become genuinely hard: several products at different stages, seasonality that changes the maths, or competitors who respond to what you do. That can happen well below the spend levels agencies usually quote.
Most good arrangements are exactly that. The software does the mechanical work at a speed no person matches, and the agency decides what it should be doing. Paying an agency to do by hand what a tool does in a second is paying for the wrong thing.
While TikTok Shop has been gaining popularity as a shopping platform, it is important to consider various factors before making a purchase. Always check the credibility of sellers, read reviews, and be mindful of the platform's fees, creator costs, and your true ROI to ensure safe and informed shopping experiences.
To set effective ROAS goals for Amazon PPC, focus on listing relevance, keyword prioritization, and conversion optimization rather than relying solely on generic benchmarks. Use the "Defend / Grow / Scale" approach to tailor ROAS targets to different campaign types, optimizing for both profit and growth.
When evaluating TikTok Shop Ads, it’s crucial to track "true ROI" by factoring in all associated costs, including platform fees, ad spend, and creator costs. Keeping separate targets for in-platform ROAS and "true ROI" will help you accurately measure profitability and avoid misjudging the campaign’s contribution.
For Walmart Connect in 2026, define ROAS goals based on your specific unit economics and operational readiness. Consider your inventory, pricing, and listing quality before setting targets, and apply the "Defend / Grow / Scale" tiers to align with your expansion stage.
It runs the paid channels against a number the business agreed to, and takes responsibility for that number rather than for activity. In practice that means campaign management, reading what search terms are actually bought on, fixing what stops visits converting, and reporting decisions rather than statistics.
Performance marketing is judged on a measurable outcome, usually revenue against spend. Digital marketing is the wider category and includes work whose return is real but slower to attribute, such as brand and content. The difference is what the engagement is accountable for.
Spend, sales, return on ad spend and total advertising cost of sale, and beside them what was changed, why, what was learned and what happens next. A report made only of numbers describes what happened without saying whether anybody was steering.
When the decisions are settled and the work is mechanical, software is cheaper and faster. An agency earns its fee on judgement: several products at different stages, seasonality that changes the maths, or competitors who respond to what you do.
Only if the listing converts. Advertising buys visits, and everything that turns a visit into a sale sits on the detail page. A campaign is often blamed for a conversion problem it cannot solve.
Indirectly and heavily. Better reviews raise conversion, higher conversion raises relevance, and higher relevance wins placement at a lower bid. Reviews are an advertising input even though nothing in the console says so.
Once a week for most accounts. Bids need enough clicks behind them to mean something, and daily changes usually react to noise rather than to performance.
It depends on your margin and on what the campaign is for. A launch campaign that is buying reviews and rank can run above break even on purpose. A profitability campaign cannot.
Not on click count alone. Judge a keyword once it has had enough clicks to have plausibly produced a sale at your conversion rate, and negate the search terms rather than the keyword where the intent is simply wrong.
At least seven days, and fourteen on lower volume keywords. Amazon attributes a sale to the click that led to it, and shoppers do not always buy on the day they click, so a change made on Monday is not fully visible on Tuesday. Reacting sooner means reacting to an incomplete number.
Higher than your profitability target, on purpose. A launch is buying rank and reviews as well as sales, and judging it on immediate ACoS alone will end it before it works.
Usually four to eight weeks, and the end of it is a decision rather than a date: you move on when the product holds organic placement for its main terms without paid support.
Both, doing different jobs. Automatic finds the terms you did not think of, and manual buys the ones you already know matter. The automatic campaign feeds the manual one.
At least three, separated by the job each does rather than by product. A discovery campaign, automatic or broad match, finds new search terms on a modest budget. A harvest campaign holds the exact match keywords promoted out of discovery once they convert, and carries most of the budget. A defence campaign covers your own brand and product terms, so a competitor cannot buy the shopper who was already looking for you.
Most agencies charge either a monthly retainer or a percentage of ad spend, commonly somewhere between ten and twenty percent. The useful question is not the rate but what is included, because the work below takes a known number of hours and a price far under that is buying you a monthly report.
One to three for most accounts, done by hand. Catalogue size matters less than how much of the reporting and switching is manual, which is why the same account can take one specialist an hour and another most of a morning.
Spend, sales, ACoS and total ACoS, the changes made and why, what was learned from the search term report, and what happens next month. A report that only contains numbers is a statement, not management.
Revisit the structure and the targets. Products get added, seasons shift and competitors change what they bid on, so a structure that was right in January quietly stops being right; the symptom is campaigns that neither fail nor improve. A product that has finished launching should no longer run at launch ACoS, and one entering its season may deserve to run above target on purpose.
Negative exact blocks only that precise term. Negative phrase blocks anything containing it. Phrase is powerful and easy to overreach with, because one careless entry can block a family of terms that were converting.
Enough that a zero means something. If your listing converts at ten percent, ten to fifteen clicks with no sale is a reasonable line. Below that you are excluding a term you have not measured.
Yes, in two ways. A negative phrase that is too broad removes converting terms along with the waste, and negatives that are never reviewed keep blocking terms long after the reason has gone.
At ad group level for terms that are wrong for one product, and at campaign level only for terms that are wrong for everything in the campaign. Putting every negative at campaign level is quicker, and sooner or later blocks a term that was fine for another product in the same campaign.
Indirectly but genuinely. Sales and conversion on a term contribute to how Amazon ranks you for it, and advertising is a way of getting those sales before you rank. It is not a shortcut to rank on a term you cannot convert.
Usually not entirely. Ranking first organically and appearing in a sponsored slot on the same page takes up more of the result, and stopping removes the sales that keep the rank. Reduce the bid and watch, rather than pausing outright.
Listing. Advertising buys visits, and everything that turns a visit into a sale sits on the detail page. Sending paid traffic to a listing that does not convert is the most reliable way to spend money and learn nothing.
It lists the exact words real shoppers used before buying your product, with the conversion rate beside each, which makes it better keyword research than a tool's category averages. Terms that convert well and appear nowhere in your title or bullets are the cheapest listing improvement available. Working the other way round fills a listing with terms nobody buys on.
Backend keywords, properly called backend search terms, are a hidden field on an Amazon listing that makes it eligible to appear for searches the visible copy does not contain. Shoppers never see them. They control what a listing can rank for, not where it ranks.
Amazon enforces a byte limit on the field, and exceeding it can cause the whole field to be ignored rather than truncated. That failure is silent, so the length is worth checking rather than assuming, and the limit is best treated as a budget to spend deliberately.
No. Terms are separated by spaces and commas simply consume characters from a limited budget without adding any indexing value.
Platinum keywords are a legacy Seller Central field from the old Platinum Merchant programme, which was invitation-only and is effectively historical. For almost every seller the field does nothing and can be left empty.
Yes, for what they have always done: making a listing eligible for searches its visible copy does not cover. They have never been a ranking lever on their own, and expecting them to lift position rather than open eligibility is the most common reason sellers conclude they stopped working.
No. It is against Amazon policy, and beyond the policy it invites complaints from the brand owner. Those complaints go to Amazon and land on your account health, which is a costly way to buy a few impressions.
It means the order cannot proceed until something is approved. Usually that is on the buyer's side: a payment or address verification, an Amazon Business purchase awaiting sign-off from a manager, or an internal check. Occasionally it means Amazon needs an approval from the seller relating to a restricted or gated product.
No. Wait until the status resolves. Shipping against an unresolved order risks sending goods for an order that will be cancelled and not paid for, and it will not speed the approval up.
You can, but seller-initiated cancellations count against your order defect metrics, and an order waiting on a buyer-side approval is not your fault to absorb. Waiting for it to resolve or cancel on its own is almost always the better option.
The pending status itself does not. What can affect them is how you react to it: cancelling the order yourself, or shipping late once it finally clears. Both are avoidable by letting the status resolve and dispatching promptly when it does.
Brands can restrict who lists against their catalogue through Amazon's gating and Brand Registry tools, and Amazon gates some categories itself for safety or compliance reasons. The requirement is set by the brand or the category rather than by anything about your account.
Payment and verification checks usually clear within hours. Amazon Business approvals depend entirely on the buyer's own organisation and can take days, because the person who has to approve it may not know they are being waited on.
Amazon Business accounts can require internal sign-off for purchases above a spending threshold or in certain categories. The order waits until the nominated approver in the buyer's organisation acts. It is a setting on their account and nothing to do with your listing.
Ungating is Amazon's approval process for restricted categories and brands. It typically requires invoices from an authorised supplier, and sometimes documentation or test reports. Most failed attempts fail on invoice formatting or supplier recognition rather than on an outright refusal.
An FBA reimbursement is money Amazon owes a seller when inventory is lost, damaged, disposed of, refunded without being returned, or miscounted inside the fulfilment network. Amazon reimburses much of this automatically, and a meaningful share only after a claim.
By reconciling the inventory ledger, reimbursements report, returns report and shipment records against each other to find events that should have produced a reimbursement and did not. Most accounts that have never done this find something.
Monthly or quarterly, depending on unit volume. The claim windows run from the event, so an annual exercise will always find that some of what it identifies has already expired. A schedule catches things while they are still claimable.
Yes. The reports are all in Seller Central and the cases are filed through normal support channels. The work is the reconciliation rather than the filing, and it is a data exercise that scales badly by hand once an account is doing real volume.
Yes, where an item has been charged at the wrong size tier or weight. This is one of the more common and least noticed cases, because the individual amounts are small and only become significant multiplied across every unit sold since the product was measured.
Yes. Claim windows run from the event rather than from when you noticed it, so the longer an account has run without reconciliation the more has already expired. That is the main argument for doing it on a schedule rather than once.
It depends entirely on how they operate. Accurate, well evidenced claims filed within policy are routine. Services promising guaranteed recoveries or filing aggressively at volume put your account at risk, and the account carrying that risk is yours rather than theirs.
Walmart Fulfillment Services has its own reimbursement process for inventory lost or damaged in its network. The principle is the same and the reports, deadlines and claim routes are different, so a process built for Amazon does not transfer directly.
Not if it is lost or damaged before a customer orders it. Since 31 March 2025 Amazon reimburses that stock at its manufacturing cost, the seller's own figure or Amazon's estimate, excluding shipping, handling and customs duties. Items lost or damaged after a customer order are still reimbursed at the sales price on the order, minus applicable fees.
Yes, with conditions. You must be the seller of record: your name on packing slips, invoices and packaging, and you responsible for returns. What Amazon prohibits is buying from another retailer and having that retailer ship directly to your customer, which is the model most commonly marketed as Amazon dropshipping.
For small, standard-sized, steadily moving products, usually yes, because Prime eligibility lifts conversion enough to cover the fees. It stops being worth it on large or heavy items, slow movers accruing storage fees, and products whose margins cannot absorb fulfilment and referral fees together. It is a decision per product rather than per catalogue.
Private label means selling products under your own brand, usually manufactured to your specification. You own the listing and the detail page rather than sharing it, which removes Buy Box competition and gives the highest margin potential, in exchange for the highest capital requirement and product development risk.
Wholesale means buying genuine branded products from distributors or the brand and reselling them. Margins are lower than private label and setup is faster, and the constraints are supply relationships and gating approvals rather than product development.
It is cheap to start and difficult to sustain. Supply is inconsistent, retail receipts are not the supplier invoices Amazon accepts when a listing is challenged, and it does not scale without effectively becoming wholesale. It works as a way to learn the platform rather than as a business to build on.
It depends on the product. FBA generally produces higher margin on goods that move steadily, because of better unit costs and higher conversion. Dropshipping produces lower margin but requires almost no capital, so return on capital can be better even when return on revenue is worse.
You can run both models across a catalogue: FBA for core products where the fees are justified, and seller fulfilled or supplier fulfilled for items where they are not. The two are not exclusive, because one is a fulfilment method and the other is a sourcing model.
Most commonly for shipping from another retailer, which is a direct policy breach, and for the metric damage that follows unreliable third party dispatch: late shipments, cancellations for out of stock, and defect rates above Amazon's thresholds.
Generally no. Under the first-sale doctrine, genuine goods that have been legitimately purchased can be resold, and a trademark alone does not prevent that. What you do need is documentation Amazon will accept if the listing is challenged, which in practice means supplier invoices in your registered business name.
An authorised reseller has a commercial arrangement with the brand or its distributor. It is a business relationship rather than a legal requirement for resale, though it makes supply, gating approval and defending complaints considerably easier.
Yes, and the rules differ. eBay is the most permissive on used and refurbished stock, Walmart applies its own approvals and weighs price competitiveness heavily in placement, and your own Shopify store has no gating at all but supplies no traffic either. A reseller across several channels needs to decide which products belong where.
Selling through Amazon Business changes quantity expectations, pricing tiers and documentation rather than the underlying rules about authenticity and invoices. The gating and complaint process is the same. What changes is that buyers may need quantity pricing and their own organisation may require internal approval.
Buy from distributors or the brand rather than from retail, keep invoices in your registered business name covering the quantity sold, match products to the exact ASIN, and keep condition and packaging exactly as the listing describes. Most complaints are answered with documentation, so the documentation has to exist before the complaint arrives.
Invoices from a legitimate supplier, in the seller's registered business name, dated within Amazon's stated recency window, and covering the quantity concerned. Retail receipts, screenshots and orders from other marketplaces are not accepted, which rules out most of what beginners are commonly advised to do.
Many brands and categories are gated, requiring approval before you can list. Approval usually rests on supplier invoices, and sometimes on documentation or authorisation letters. Gating should be checked before buying stock, not after the listing is refused.
It can be, where you have supply advantages, exclusive or semi-exclusive distribution, or an operational edge in a category others avoid. It is difficult where the only differentiator is price, because you are competing on the same detail page as everybody else selling the identical item.
No. Collections created for merchandising convenience, seasonal groupings and automated tag collections often have no search demand behind them. They still get crawled and still compete. Index the collections that match how customers search and let the rest exist for navigation only.
By what somebody typing it wants. A category term wants a collection, a specific product term wants a product page, and a question wants an article. Write the decision down per term, because ownership that lives in somebody's head is ownership two pages will eventually both claim.
They help in two ways: they give Google a clear hierarchy to read, and with breadcrumb structured data they replace the raw URL in the search result with a readable path. Neither is dramatic on its own and both are cheap.
Most commonly because several URLs exist for the same collection through tags, filters and sorting, because the page has no content beyond a product grid, because pagination splits the range across many URLs, or because a blog post and a product page are competing for the same term. Usually more than one of those at once.
Yes, on the collections that matter commercially. Google is being asked to rank a page for a term that nothing on the page discusses. A few hundred words that genuinely help somebody choose within the category is enough, and it does not need to be an unread wall of copy under the grid.
Check the canonical tags on tag, filter and sorted variants, since these are the largest source of duplicates and are often wrong or absent in a theme. Decide which URL should own the term, make every internal link point at it, and let the canonicals agree with that decision.
They create a second address for every product, which is a duplicate unless the canonical points back at the product's own URL. Most themes handle it correctly and it is worth verifying rather than assuming, because when it is wrong it is wrong across the whole catalogue.
As many as there are genuine ways a customer would narrow the range, and no more. Collections created for merchandising convenience that nobody searches for still get crawled, still compete, and still dilute the ones that matter.
Usually yes for filter and sort combinations, which generate large numbers of near-identical pages and consume crawl budget without ranking. The exception is a filter that matches a real search, such as a size or colour people genuinely look for, which may deserve its own indexed page.
Shopify's own infrastructure is fast and is rarely the constraint. What the platform does limit is control: you cannot remove its core scripts or change how apps inject theirs, so optimisation is largely about what you add rather than about the platform underneath.
Below the fold, yes. Above the fold, no. Lazy loading the hero or main product image delays the Largest Contentful Paint, which is the opposite of the intended effect and a common mistake when lazy loading is applied site-wide by a plugin.
After any app install or theme change, and monthly otherwise. Speed regresses gradually through accumulation rather than suddenly, so a scheduled check catches it while it is still one app rather than five.
Considerably. Most eCommerce traffic is mobile, mobile devices parse JavaScript more slowly, and mobile connections are less reliable. Google's Core Web Vitals assessment is also based on mobile, so the mobile number is the one that counts.
Usually the accumulation of apps rather than the theme. Most apps inject JavaScript into every page rather than only the pages where their feature appears, so a store with fifteen apps loads a dozen scripts on pages that use none of them. Oversized images are the other common cause.
They are a ranking input and a modest one compared with relevance and authority. Marketplaces weigh speed differently again, because the page belongs to them rather than to you: <a href="/services/marketplace-management/amazon">Amazon</a> and <a href="/services/marketplace-management/walmart">Walmart</a> own the markup on their own pages. They matter more for conversion: mobile shoppers abandon slow pages regardless of where those pages ranked, so the commercial argument is stronger than the SEO one.
There is no number, because apps differ enormously in weight. The useful test is whether each one is still used by the business and how much it loads on pages where its feature does not appear. Stores routinely carry apps installed for a campaign years ago that still load on every page.
Sometimes, and it is rarely the cheapest fix. A rebuild on a fast theme with the same fifteen apps and the same unoptimised images ends up in the same place. Audit what is loading first, because that answer often makes the rebuild unnecessary.
Google's thresholds are the target: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1, measured on mobile. Most Shopify stores fail on LCP first, and images are usually the reason.
It is usually the single largest improvement available, because images are the biggest thing on most pages and the Largest Contentful Paint element is an image on nearly every product and collection page. Correct format, correct size for the display, and explicit dimensions so nothing shifts.
Product with offer details on product pages, BreadcrumbList throughout, Organization once for the business, and FAQPage where a page genuinely answers questions. Review and AggregateRating only where real reviews are visible on the page.
Search Console reports errors and warnings across the site, which is the fastest way to find problems at scale. Google's Rich Results Test checks a single URL and shows what would be eligible. Both are worth using: one finds the pattern, the other confirms the fix.
Yes, and it is common on Shopify. A theme adds Product schema, an app adds its own, and the two disagree on price or availability. The result is usually no rich result at all, which is why verifying what is already on the page matters before adding more.
Yes, and this is enforced. Markup describing a price, rating or availability that the page does not show is against the guidelines regardless of intent, including when it is a caching accident rather than a deliberate one.
Product schema is structured data stating that a page describes a product, along with its name, price, currency, availability, condition and rating. It lets a search result display those details rather than only a title and description.
Not directly. It improves how a result looks, which improves click-through rate, and click-through rate on a result you already hold is often more valuable than moving up a position. It also makes a page easier for assistants to quote accurately.
Usually a validation error, conflicting markup from a theme and an app both adding Product schema, or markup that does not match what is visible on the page. Search Console reports these and is the fastest place to find out. Rich results are also never guaranteed even when the markup is valid.
No. Review and rating markup must reflect genuine reviews visible on the page, and the guidelines are enforced. Marking up ratings you do not have risks losing rich results across the whole site, which is a large penalty for a small gain.
Most themes add some, and its completeness and correctness vary considerably. It is also common for an app to add a second, conflicting set. Verifying what is actually on the page is worth doing before adding anything new.
It helps an assistant read facts unambiguously, which makes a page easier to quote accurately. That matters more as product research moves into assistants, and it is a reason to treat structured data as content work rather than as a technical afterthought.
Rarely. Variants of one product usually share a description and differ on the attributes, which belong in structured data rather than in prose. Writing a description per colour produces near-duplicate pages competing with each other, which is the problem you were trying to solve.
Not for SEO reasons alone. A long tail of low-demand products is not a ranking problem as long as they are not duplicated and not consuming crawl budget through filter combinations. Remove them when they cost money to hold, not because they do not rank.
Write them separately. A marketplace listing rewards scannable bullets for a shopper comparing options, and a product page frequently rewards depth because Google prefers pages that cover a topic. Pasting one into the other produces copy that underperforms in both places.
It should contain the words your customers actually use, which are frequently not the manufacturer's words. A supplier writes thermal regulation membrane and buyers search for waterproof breathable. Write for the second and the first can appear in the specifications.
The customer service inbox, the reviews on competing products, and the questions section of a marketplace listing for the same item. All three are records of what real buyers were unsure about, which is better evidence than any keyword tool.
It is a problem when it is thin and duplicated, which is the usual combination. A short unique description that answers the real objection outperforms a long generic one, so the fix is specificity rather than length.
On the products you want to rank, yes. Supplier copy appears on every retailer selling that item, so Google has nothing to distinguish you on and falls back to authority, which a smaller store loses. On long-tail products nobody searches for, supplier copy is fine indefinitely.
Long enough to answer the questions that stop the purchase, which is usually 150 to 400 words. Length is not the ranking factor. A short description that answers the real objection beats a thousand words written to satisfy a word count.
The ones already getting impressions and few clicks in Search Console, because Google is willing to show them and buyers are not choosing them. Then the products with the most commercial demand behind them. Rewriting an entire catalogue is how a rewriting project stalls and produces nothing.
No. Specifications are facts and should match the supplier exactly, because inventing variation is how a store publishes wrong dimensions. Put them in structured data as well as on the page, which is where a search engine or an assistant can read them.
It can produce text quickly, and the risk is producing text that is fluent, generic and answers none of the questions your customers actually ask. Used with real inputs, the customer service inbox, reviews, and the actual objections, it is a useful drafting tool. Used to generate variation for its own sake it produces a different kind of thin page.
There is no penalty in the sense of a punishment. What happens is more mundane and more damaging: Google picks one version to show and it is usually not the smaller store's. The result looks like a penalty and is really a filter.
The method transfers and the sources do not. Google keyword tools describe how people search Google. A marketplace search term report describes what buyers typed inside that marketplace before buying your product, which is better evidence for that channel and useless for the other.
Indirectly, and less than people hope. It can bring traffic to a listing, and traffic that converts becomes sales history, which is a ranking input. But an article cannot pass authority into a marketplace the way a link passes authority in Google, so the mechanism is weaker and slower.
Marketplace work, currently, because the pool is smaller and the feedback loop is faster, which hides weak practitioners for less time. Google SEO has more practitioners and a longer feedback loop, which means poor work can look plausible for a quarter before the absence of results becomes undeniable.
Yes. eBay's Cassini weighs listing quality, price, sales history, postage speed and seller performance together, and seller account standing feeds placement in a way neither Google nor Amazon does. A good listing on a poor account is held back by something unrelated to the listing.
Barely, in the sense the other platforms mean. Demand there is created in the feed by content rather than captured from a search box, so creators and video do the work keywords and bids do elsewhere. Product content still matters, but for conversion rather than for discovery.
Marketplaces usually show results faster because the audience is already there and ready to buy, and a listing can be moved with advertising. Google takes longer and the position is more durable once earned, because it is not resettable by a competitor outbidding you.
eCommerce SEO is ranking a storefront in Google, which ranks pages it can crawl, understand and trust using signals well beyond the page. Amazon SEO is ranking a listing inside Amazon, which ranks what sells for a keyword. Conversion and sales velocity are dominant there and there is no equivalent of a backlink.
Not directly. They are separate systems with separate indexes. Amazon listings can themselves rank in Google, which is a reason to write them well, but work done inside Amazon does not transfer authority to your own domain.
Almost not at all for ranking inside the marketplace, which has no equivalent signal. External traffic to a listing can help indirectly by producing sales, and sales are a ranking input, but that is a different mechanism from the one links serve in Google.
No. Marketplace shoppers compare rather than read, and length is not a ranking factor there. Scannable bullets that answer objections work better. On your own store, depth frequently helps because Google rewards pages that cover a topic properly.
Yes, and the thing to check is whether they treat them as the same job. An agency applying marketplace tactics to a store produces keyword-stuffed titles and no technical work; one applying Google tactics to a listing produces long copy nobody reads. Knowing which discipline applies where is the actual skill.
Whichever channel is already producing revenue, usually. Marketplace work compounds faster because the audience is already there and ready to buy. Store SEO compounds more durably because you own the asset. Most brands should do both eventually and rarely have the capacity to start both at once.
Most often because the same item is available cheaper elsewhere on the open web, which frequently means on the brand's own website during a promotion. The seller centre reports the item as unpublished for price; the work is finding which cheaper listing triggered it.
Yes. The comparison is against the same product available cheaper anywhere publicly, and it makes no exception for the seller's own store. A site-wide discount code moves the effective price on every channel that is watching.
Amazon pricing is largely a contest between sellers of the same item for the buy box. Walmart has that contest plus an external comparison that can remove the listing from the site entirely, which is a compliance problem rather than a competitive one.
Only if it is given a ceiling based on the item's lowest public price anywhere, not on margin. A repricer watching only other Walmart offers will raise a price into the range that gets the listing unpublished and report it as a win.
Yes, and more visibly than on Amazon. Two comparable listings with a meaningful price gap tend to place differently, which is consistent with a retailer whose brand is built on price. Pricing and ranking are one decision on this marketplace.
Not automatically. Sometimes the honest answer is that the item is not viable on this channel at a price you can live with, and carrying it anyway is how sellers accumulate products that make nothing. Match where the margin survives it.
It is a view in the seller centre listing items currently not visible to shoppers and why. Almost nobody reads it daily, and it is the fastest signal that a pricing decision made on another channel has reached this one.
Continuously by a repricer for the competitive comparison, and whenever a promotion is planned anywhere else in the business for the external one. The second is a process problem rather than a software one: whoever runs store promotions has to know the marketplaces are watching.
The price rules apply either way. Fulfilment affects the buy box contest and delivery promise rather than the external comparison, so a WFS item can be unpublished for price exactly as a seller-fulfilled one can.
Yes, if the item is publicly available cheaper through them or through anybody they supply. For wholesale and distribution models the ceiling is set by other people, which makes it something to monitor rather than something to decide.
The number of days between an order being placed and it shipping. It sets the delivery date shown to the shopper, so a shorter lag time makes the listing more attractive and every order harder to meet. The right value is the one you hit consistently, not the one you hit on a good week.
Usually the delivery promise rather than the listing. A competitor with a shorter lag time or Walmart fulfilment shows a sooner arrival date, and shoppers choose that more often than they choose a small price saving.
Either ring-fence a quantity per channel so each has stock that is only its own, or synchronise from one source of truth close to real time. Ring-fencing wastes stock and never oversells; syncing uses stock efficiently and oversells occasionally.
Yes. Cancellations count against seller performance at the account level, so a problem that started with one item's stock setting can affect buy box eligibility across everything you sell.
For fast-moving items where delivery speed decides the sale, usually yes: it removes the lag time question and improves the promise. For slow, heavy, fragile or seasonal items, storage economics often beat delivery speed and seller-fulfilled wins.
Something that holds one source of truth and pushes quantities to every channel, and that alerts on low stock, missed lag times and cancellation spikes. The specific tool matters less than whether one system is authoritative; two systems that both think they are is worse than a spreadsheet.
Enough to cover the supplier's real lead time at the item's current rate of sale, plus room for the variability you actually see rather than the variability you would like. For private label with long lead times that is considerably more than it feels like it should be.
For a handful of items on one channel, yes. The moment the same stock is listed on more than one marketplace it stops being a list and becomes an arithmetic problem that runs continuously, which is what software is for.
As close to nothing as the operation allows. Rather than aiming at a threshold, treat every cancellation as having a cause worth finding, because the causes repeat: an optimistic lag time, a channel sync gap, or a supplier who is less reliable than the feed suggests.
Daily for fast movers, weekly for everything else, and a full reconciliation whenever the numbers disagree. A discrepancy corrected silently is a symptom nobody investigated.
Walmart leans harder on structured attributes, publishes a listing quality score that tells you which part of a listing is weak, and weighs price more visibly. Amazon leaves more of the same work to keyword placement in text fields.
A per-item score in the seller centre broken into content and discoverability, offer, ratings and reviews, and post-purchase quality. It is a checklist of what the platform can measure, not the ranking itself, and most sellers never open it.
They populate the filters on a search results page. A shopper filtering to a size or material you left blank has removed you from consideration entirely, which is exclusion rather than a lower position, and no title work compensates for it.
As a way to launch, yes. As a final answer, no. Walmart publishes a title format per category and follows shopper behaviour closely, so a title stuffed with terms reads as a worse product and performs like one.
Indirectly and strongly. They affect whether people click and buy, which is a large ranking input, and they are also a compliance matter: a main image that breaks the category requirements can get the listing suppressed.
Content changes are indexed within days. The performance signals that follow, clicks and conversions, take longer to accumulate, so judge a change over weeks rather than days and change one thing at a time on items that matter.
There is no direct equivalent to Amazon's search terms field. The nearest thing is completing the attributes and writing the description properly, which is where terms that do not fit the title belong.
The ones that already sell, and enough of them to see a pattern rather than a coincidence. Optimising a catalogue evenly means spending most of the effort on items nobody was going to buy.
Treat it as a diagnostic rather than a dial. A listing scoring well has not left anything obvious undone, which is a good position to rank from, but the score itself is not the mechanism.
Yes, and it is usually more efficient, provided they know the differences rather than assuming the platforms are one job. The attribute discipline in particular has no Amazon equivalent and gets skipped by teams who learned on Amazon.
Inventory lost or damaged in fulfilment, returns refunded and never received back, fees charged on incorrect weights or dimensions, and shipments received short of what was sent.
Some cases are handled automatically and many are not, which is why reconciling the reports matters. The amounts are individually small enough that nobody notices them going unclaimed.
Each claim type has its own window, published in the seller help documentation and subject to change, so use the current published figure rather than one from an article. The practical consequence is constant: a discrepancy found late is one you own.
Weight and dimensions, primarily. Because those are recorded once and applied to every unit thereafter, a single measurement error becomes a recurring overcharge until the record is corrected.
Measure and weigh the packaged item yourself and compare with what the fee is being calculated on. It is worth doing for anything near the boundary of a size band, where a fraction of an inch changes the rate.
Proof the item existed, reached them and did not come back: the shipment record, the receiving record, the order or return identifier, and the settlement line. Claims fail on missing references more often than on merit.
Monthly is enough for most sellers, weekly for anyone with significant fulfilment volume, because reconciling recent shipments is far quicker than reconstructing old ones.
If it recovers money that would otherwise go unclaimed, a share of something beats all of nothing. It is worth knowing what proportion you are giving away on discrepancies a report would have shown you, and deciding rather than defaulting.
Fulfilment loss claims apply to inventory held by Walmart. Seller-fulfilled sellers have the equivalent exposure with their carriers instead, which is a separate process with its own windows and evidence.
Yes. One item that is always damaged is a packaging problem, and repeated losses at one location is a handling pattern. Claiming the same thing every month without changing anything is treating the symptom.
Costs per click are generally lower because fewer sellers advertise, but there is also less traffic. The same budget buys more clicks against a smaller pool, so the risk shifts from not affording enough to buying more than is worth having.
The vocabulary and the structure transfer almost intact. What does not transfer is the optimisation cadence: less traffic means terms take longer to accumulate meaningful data, and an Amazon weekly routine applied here prunes terms on noise.
Both. Automatic discovers terms you would not have chosen, manual controls what you pay for the ones that work, and the search term report is what moves terms from the first to the second.
Usually relevance rather than bid. An advertised item still has to be considered relevant to the search, and relevance comes from the listing content, so an incomplete listing cannot be advertised into terms it does not support.
Long enough to accumulate a number of clicks per term that would justify a decision, which at lower traffic is typically weeks rather than days. A keyword with four clicks and no sales is not evidence of anything.
Do not carry an Amazon target across unchanged. A return that would be poor on Amazon can be acceptable at a lower cost per click with the same margin, and the reverse holds in competitive categories.
No. Most catalogues have a small number of items worth advertising, and every additional item costs the same attention as the first. Start with the ones whose listings are complete and that already sell.
For wholesale and any item several sellers offer, advertising an item whose buy box belongs to somebody else is buying traffic for a competitor's offer. Settle the buy box first.
An amount you would not mind losing entirely, because the first period is buying search term data rather than a return. What matters more than the figure is leaving it running long enough for that data to exist.
Yes, and it is usually better, provided they adjust the cadence and the benchmarks rather than running the Amazon playbook. The failure is not a lack of skill, it is applying a routine tuned for volume to an account that does not have it.
The default sort order. It tries to show the listing most likely to result in a satisfied purchase rather than the closest word match, which is why two identical listings can rank differently based on how their sellers have performed.
Most often seller performance rather than the listings. Defects, late dispatch and cases closed without resolution all feed Best Match, so an account can decline in visibility while every listing stays exactly as it was.
They are eBay's structured fields: brand, model, size, colour, condition and whatever else the category defines. They populate the filters on a results page, so a missing one removes the listing from any search that filters on it, which is exclusion rather than a lower position.
Use the eighty characters on words a buyer would actually type, front-loaded: brand, model, defining attribute, item type, size. Punctuation and capitalised filler spend characters that could carry a search term.
It is one of the offer terms Best Match weighs, alongside dispatch time and the returns policy. Cutting them to protect margin can cost more traffic than it saves, so the arithmetic is worth doing per item rather than as a blanket policy.
No. Your eBay listing is yours rather than a shared product page, condition is a first-class concept, and seller performance feeds the search directly rather than only the buy box. The habits transfer badly.
Engagement with a listing is part of how it performs, and eBay buyers browse and watch far more than Amazon buyers do. A listing can build momentum over days rather than converting on first view, which is worth knowing before judging one too early.
All the required ones and as many recommended ones as genuinely apply. eBay marks which are missing per category, and the recommended ones are recommended because buyers filter on them.
As a way to launch, yes. As a final answer, no: the title limit is different, item specifics have no Amazon equivalent, and condition matters here in a way it usually does not there.
Relevance changes are picked up quickly. The performance signals that follow take longer to accumulate, so change one thing at a time on listings that already get traffic and judge it over weeks.
A minimum number of transactions and account age, a low transaction defect rate, a low late dispatch rate, few cases closed without seller resolution, and tracking uploaded on time to recognised carriers. The exact thresholds are published per eBay site and change, so work from your seller dashboard.
Chiefly a seller-cancelled order because the item was unavailable, and a case closed without the seller resolving it. It is not a negative review, which is measured separately and does not itself create a defect.
Almost always cancellations from stock that was not really there, or late dispatch. Both are operational rather than about how customers felt, and both count regardless of whether anyone complained.
Until the bad period falls out of the rolling evaluation window, which is measured in months rather than weeks. There is no way to shorten it other than continuing to do everything correctly while it passes.
It affects Best Match placement, carries the badge, and on qualifying listings a final value fee discount where the site offers one. Everything it requires is also what a well-run account does anyway, so it is more a by-product than a project.
The one you meet consistently including bad weeks, not the one you meet when everything runs properly. A one-day window is attractive to buyers and unforgiving, and the late dispatch rate is measured against whatever you promised.
Tracking has to be uploaded within the handling time and show timely dispatch. A parcel that genuinely went out on time with tracking added afterwards can still be counted late, which is a paperwork failure scored as an operational one.
It is the hardest model for it. The handling time depends on a supplier you do not control and the stock levels are reported by somebody else, so every supplier failure becomes your defect. It is possible with a genuinely reliable supplier and rare otherwise.
Not directly. Feedback and defects are different measures. A negative review matters for conversion and can accompany a case, but the status is decided by defects, late dispatch, unresolved cases and tracking.
Before it, not during it. Lengthen handling times honestly, reduce listings you cannot reliably fulfil, and make sure stock levels are accurate across every channel. A defect taken in December is still in the window in spring.
Only in defined circumstances: policy breaches such as profanity or personal information, causes eBay recognises as outside your control, feedback withdrawn by the buyer, and retaliatory or extortionate feedback. A buyer who is simply unhappy or factually wrong is not grounds for removal.
Briefly and factually, addressed to future readers rather than to the buyer. State what happened, state what you did, stop. Arguing tells every future buyer that a problem with you becomes a dispute.
A request asking the buyer to change what they left. The number available is capped and requests expire, so use them after the problem is genuinely fixed rather than as a first move.
No. It breaches eBay policy, and it is the same transaction extortion looks like from the other direction. Fix the problem first without mentioning feedback, then ask.
Not directly. Status is decided by transaction defects, late dispatch, unresolved cases and tracking. Feedback affects how buyers perceive you and how listings convert, which matters commercially but through a different mechanism.
Individually, little, particularly against a long history. Repeatedly for the same reason, a great deal, because it is describing a fault in the operation rather than an unlucky transaction.
Yes, almost always. Most negatives come from buyers who could not get a response, and a quick resolution often produces a revision. The public reply is for the ones that cannot be resolved.
Reply once, calmly, with the facts, and leave it. It is not usually removable, and a second reply reads as an argument. Future buyers weigh the tone of a reply at least as heavily as its content.
The ones who look at negative feedback do, which is exactly the audience deciding whether to risk you. That is why the reply is worth writing carefully even when nobody will change their mind about the original order.
Describe items accurately including flaws, photograph what will actually arrive, answer messages quickly, and resolve rather than escalate. A refund costs the item; a case closed against you costs visibility for months.
A cap on how many items an account may have listed and how much it may sell in value within a calendar month. Some categories carry their own limits in addition to the account level ones.
Because an unlimited new account is how fraud scales. The limit is the cost honest sellers pay for a marketplace that buyers trust, and it lifts as the account produces evidence of being what it says it is.
Mostly by selling well and being paid without incident: completed sales, on-time dispatch with tracking, few defects, cases resolved. eBay reviews accounts periodically and raises limits on its own where the history supports it.
Yes when there is a real reason: an account performing well, consistently hitting its ceiling, with a catalogue and the ability to fulfil it. It does not help on an account with defects or a short history, because the request is assessed on the same evidence.
Periodically rather than on a published schedule, which is why the useful strategy is producing the evidence continuously rather than timing a request.
Yes. A spike in defects, a run of cancellations or a category restriction can reduce room on an account that had earned it. The usual cause is selling stock that was not really available.
There are account level limits and, in some categories, additional limits of their own. A strong record in one category does not automatically carry into a new one.
What you sell, where you source it, how you fulfil orders and what you expect to sell monthly. A request that reads like a business explaining itself is treated as one.
The value cap counts sales rather than listings, so an account with plenty of room to list can still reach its ceiling partway through the month and stop until it turns over.
It is evidence the room would be used, which helps. It is also two thirds of your selling time lost if it happens on the twelfth, so it is worth solving rather than tolerating as a strategy.
Because the same unit is usually listed on several channels at once and sold twice inside the window between synchronisations. It is arithmetic rather than carelessness, and the fix is either ring-fencing stock per channel or syncing closer to real time.
Yes. A seller-cancelled order for unavailable stock is a transaction defect, and defects reduce Best Match placement, threaten Top Rated status and can reduce selling limits.
Ring-fencing wastes inventory and never oversells; synchronising uses stock efficiently and oversells occasionally. On eBay, where defects are expensive, ring-fencing is the safer default until volume justifies proper real-time software.
It works until an item sells quickly, which is exactly when failing is worst. It is the most common arrangement and the one that produces the cancellations sellers cannot explain.
Slow movers, discontinued items, hard-to-find parts, and anything where condition varies. The audience browses and searches for specifics, and used, refurbished and open-box are ordinary rather than exceptional.
Generally yes, because sales history accumulates on one listing rather than being spread across several, and that history is what Best Match reads.
End the listing rather than leaving it live. An ended listing costs the sales it would have made; a cancelled order costs a defect that stays in the evaluation window for months.
Something that holds one authoritative stock figure and pushes it to every channel, with alerts for low stock and cancellations. Which tool matters far less than whether exactly one system is authoritative.
Daily for anything that sells quickly and weekly for the rest, with a full reconciliation whenever the numbers disagree. A discrepancy corrected quietly is a symptom nobody investigated.
Only as reliably as the supplier's feed, which is the whole difficulty. You are publishing their stock position as your own, so their errors arrive as your defects, and eBay charges for those more than most channels do.
A dashboard of the measures Amazon uses to decide whether an account is selling responsibly: order defect rate, late shipment and on-time delivery, pre-fulfilment cancellations, valid tracking and policy violations. The thresholds are published in Seller Central and change, so use the figures on your own dashboard.
Usually a metric that had been drifting for weeks, and usually one that started as an inventory problem: cancellations from stock that was not there, or late shipments from a handling time nobody revisited. Policy and intellectual property matters are the other main route.
The root cause, specifically and without blaming anybody; what you have already done, in the past tense and concrete; and what prevents it recurring, as a process with an owner and a frequency. Short and specific reads better than long and apologetic.
It varies with the cause and with how well the first submission is written, which is the part you control. A well diagnosed appeal answered once beats several vague ones, because each submission is read and a repeat that adds nothing wastes the reading.
Invoices from a supplier the brand recognises, with details matching your seller account exactly. Receipts and order confirmations are not invoices. A retraction from the rights owner is worth more than any appeal, so establish with them whether the complaint is genuine before writing to Amazon.
Yes, and each one is read. Send a second only when it contains something the first did not: new evidence, a better diagnosis, or a process you have since put in place.
Yes. Performance feeds buy box eligibility and how much of your catalogue Amazon is willing to show, so an account with drifting metrics is quietly selling less long before anything is suspended.
Weekly, before anything else. It is the score that governs whether the rest of the work matters, and every problem it reports is cheaper to fix in the week it appears.
For a complex or repeated suspension, experience with what these submissions are read for is worth paying for. For a first suspension with an obvious cause, a carefully written plan of action from somebody who understands the operation is usually enough.
Stop producing defects, then wait. The rate is measured over a rolling window, so it improves as the bad period ages out rather than because of anything you can do today. That is also why protecting it before a difficult season matters more than reacting during one.
A monthly subscription for a professional account, a referral fee that is a percentage of the sale price and varies by category, and then fulfilment, storage, returns and advertising, which together usually exceed the referral fee considerably.
Yes. An individual account avoids the monthly subscription and charges per item sold instead, but referral fees apply either way, and every route to actually selling carries fulfilment or postage costs.
Amazon's commission on a sale, a percentage of the total charged to the buyer including shipping rather than of the item price alone. It varies by category and the current figures are published in Seller Central.
By size tier and weight, with size usually deciding more than weight. A light item in a slightly oversized box can cost more to ship than a heavier compact one, which makes packaging a pricing decision.
Usually returns, storage on slow stock, or advertising. Those three are the ones most often left out, and any one of them can turn a healthy looking margin into nothing.
As a percentage of revenue rather than as a project cost, because that is how it behaves. What is reasonable depends on the category and the stage, and carrying a target across from another channel unchanged is how accounts end up measuring against a benchmark that never applied to them.
Not by default. It replaces the fulfilment fee with picking, packing, postage and staff time, and it costs the delivery promise that Prime buyers choose on. It suits slow, heavy, fragile or seasonal items rather than everything.
Inbound shipping and prep, removals and disposals, promotion fees on top of the discount, photography and listing content, software, and people. The last is usually the largest and the least often counted.
Subtract the referral fee, fulfilment, landed cost, storage at this product's turnover, returns at this category's rate and the advertising it needs, from the sale price. Do it before sourcing, because after the first quarter's statements the answer is no longer useful.
No, and the differences are large enough to change which channel a product belongs on. A product that is marginal on Amazon can be comfortable on Walmart, and one that sells through content rather than search may belong on TikTok Shop instead.
Register the business, verify the account, add tax and bank details, then list products with categories, attributes, images and stock. That part is administrative. The part that decides whether you sell anything is content, which is a separate job entirely.
Because nobody browses a TikTok Shop catalogue. Purchases begin with a video somebody was not looking for, so a complete listing earns you the ability to convert traffic without producing any. Traffic comes from your content, from creators, and from live sessions.
Creators list your product, make content about it, and take a commission on sales. It is the channel's main distribution lever, and commission set too low is the usual reason a product attracts no creator interest: they choose from a marketplace of offers.
Enough to compete in your category, which is a question about what else creators are being offered rather than about your target margin. Treat it as the price of distribution, in the way a wholesaler treats a retailer's markup.
Things that demonstrate in a few seconds, at impulse price points, with supply you can sustain. Products people want to compare specifications on sell badly here, because comparison happens on a search marketplace, not while scrolling.
No, but live selling works when it is prepared: promoted beforehand, stock reserved, somebody managing comments, and a presenter who can demonstrate for an hour. Trying it once unprepared and concluding it does not work is a fair conclusion from bad evidence.
Advertising amplifies content, it does not replace it. Putting budget behind a video that is not converting organically buys a larger version of the same result. Make content, see what performs on its own, then fund that.
The content transfers with light editing and it is a reasonable way to launch. Do not expect it to sell on its own, because the listing is where traffic lands here rather than where it is won.
Some categories are restricted and some are prohibited, and the rules are enforced rather than advisory. Check before sourcing, because finding out after listing means stock you cannot sell on the channel you bought it for.
Long enough to make content consistently and find creators who convert, which is months rather than weeks. The channel rewards repetition, so the sellers who do well are usually the ones who kept posting after the first few videos did nothing.
The model is not prohibited. What is enforced is dispatch within the promised window, accurate descriptions, genuine goods and a returns process that works, and dropshipping arrangements fail those more often than holding stock does.
Because impulse buyers are impatient, a viral video surfaces batch problems in a week rather than a month, creators need the product in their hands to make content about it, and thin margins leave nothing to pay creator commission with.
You can, and delivery time is the thing that decides whether it works. Weeks of shipping against an impulse purchase produces cancellations and complaints, both of which are measured and both of which reduce how much the channel shows you.
Yes, and more than one. You cannot brief content about a product nobody has touched, and the creators doing the selling need the item in their hands to make anything worth watching.
Enough to fund creator commission on top of the usual costs, because commission is the price of distribution on this channel. If the arithmetic leaves nothing for it, there is no distribution and you are relying on your own posting alone.
It counts against seller performance, which affects how much the channel is willing to show you, and it is not confined to the orders that were late. A bad month is expensive for longer than the month.
With a domestic supplier you have a relationship with, samples in hand and one product chosen deliberately, it can work. The scattergun version, many products nobody has held from suppliers nobody has spoken to, does not.
One, or very few. Every product on this channel needs content made for it, so listing fifty is fifty products with no content rather than a wider net.
The process is the marketplace's, and the rate is usually higher than on a search marketplace, because a decision made in seconds is reconsidered in seconds. Budget for it rather than being surprised by it.
It is the model the channel suits best: you hold the stock, control the quality and keep the margin that funds the content. Most sellers who do well here are selling something they own.
Generally no. The shop is built for physical goods, and everything it enforces assumes shipping, tracking, delivery windows and returns. Availability varies by market and changes, so check the seller centre for your own market rather than trusting an article.
Sometimes permitted, often not, depending on the category rules rather than on how clever the idea is. It also carries the cost and delay of shipping something of almost no value.
No. That is misdescription, it is enforced, and it produces complaints from buyers expecting a parcel. It is one of the specific behaviours the policies exist to catch.
Not through a token listing. Off-platform diversion breaks the rules on every marketplace, and the penalty falls on the account rather than the listing. Using TikTok content to build an audience that buys on your own store is a different thing and is fine.
Sell the transaction on your own store and use TikTok for the audience. You lose the in-app checkout and you keep the customer relationship, which for a business whose product is knowledge is the more valuable of the two.
Yes. A physical product with digital material included is a physical product as far as the marketplace is concerned, and the digital part becomes a reason to choose yours over an identical item.
Because availability differs by market and changes over time, and because some of what circulates is a policy violation described as a workaround. The seller centre for your own market on the day you ask is the only reliable answer.
Amazon sells digital goods through defined programmes with their own rules, so the answer there is yes with conditions rather than a flat no. It is a different question with a different answer.
Only if you have a physical range that fits, or are willing to create one. Otherwise use the platform as an audience channel and keep the transaction where digital goods are ordinary.
It depends on the product, not the account. FBA usually wins on small, light, fast moving items where the fulfilment fee is low and the buy box matters. FBM usually wins on large, heavy, slow moving or high value goods, where Amazon's size bands and storage costs are punishing. Most established accounts run both.
Yes, indirectly. Delivery speed and seller performance are inputs into the buy box, so an FBM offer competing against an FBA one at the same price is starting behind unless its delivery promise is comparable. Seller Fulfilled Prime closes some of that gap, at the cost of meeting Amazon's delivery performance requirements yourself.
Yes. The listing stays the same; the offer changes. What you cannot do instantly is move the stock, so a switch either waits for the FBA inventory to sell through or involves a removal order, which costs money and takes time. Plan the switch a season ahead rather than in the week you notice the fees.
It lets an FBM offer carry the Prime badge, provided you meet Amazon's delivery speed, on-time and cancellation requirements from your own warehouse, with Amazon's carriers or approved alternatives. It suits sellers who already run a fast, reliable fulfilment operation and want Prime placement without moving stock into Amazon's network.
Substantially. FBA returns go back into the network and are graded for resale, and the customer service is handled for you. FBM returns come to you, and so does the conversation. The cost difference is easy to underestimate in a category where a tenth of orders come back.
Not directly, and the effect is real anyway. Ranking follows sales velocity and conversion, and FBA listings usually convert better because of Prime and the delivery promise. The lift comes through conversion rather than through any preference in the algorithm for the fulfilment method itself.
It keeps costing storage, and after enough time it attracts an aged inventory surcharge on top. The options are to sell it down with price and advertising, to have it removed to your own warehouse, or to have it disposed of. All three cost something, which is why slow moving lines are the ones to think hardest about before committing them to FBA.
Yes, and it is a sensible hedge on your best sellers. The FBA offer serves normal demand; the FBM offer keeps the listing live if FBA runs out. Losing a ranking to a stockout costs more than the thinner margin on a handful of merchant fulfilled orders.
They cap how much you can send into the network, which turns FBA into a constrained resource in the run-up to a peak. Sellers with limited capacity usually spend it on the products with the best margin per cubic foot and fulfil the rest themselves, which is a different calculation from the per unit one.
Whichever protects the cash. FBM keeps the money out of stock for longer and lets you learn the demand curve before committing to inbound shipments; FBA gets a competitive listing live faster. For a first product with unproven demand, the cash argument is usually the stronger one.
Enough to cover the lead time plus a buffer, expressed in days rather than units. For a product with a 60 day lead time and steady demand, that is roughly 75 to 90 days of cover. For erratic demand or unreliable lead times it is more, and for slow moving lines it is deliberately less, because storage and aged inventory charges make deep cover expensive.
There is no single number, because it is set by your lead time. The right question is whether cover is comfortably above the time it takes to replenish. A product with 30 days of cover and a 90 day lead time is already out of stock; it just has not happened yet.
Because some of it is reserved or stranded. Reserved units are committed to customer orders, moving between fulfilment centres or being processed. Stranded units belong to a listing that is not currently sellable. Both count in the total and neither will fill an order today, so plan against available inventory rather than the total.
Amazon caps how much inventory you can send in, based on your sales performance and the network's capacity. The limit moves, so it is worth checking before planning a large inbound shipment, and worth allocating deliberately when it is tight: the products that should get the space are the ones earning most per unit of space.
Once the catalogue is beyond a few dozen products, yes, because the arithmetic is repetitive and the cost of missing one line is a stockout. What software cannot do is decide the lead time buffer or which product deserves constrained capacity. It automates the counting, not the judgement.
Use the closest comparable product you have, discount it for being unproven, and order a first quantity you can afford to be wrong about. The first replenishment is the one that carries real information, so the plan should be to reorder early rather than to order deep.
Decide, rather than wait. The options are to sell it down with price and advertising, remove it to your own warehouse, or dispose of it. Every month of hesitation costs storage and moves the stock closer to a surcharge, and the discount that would have cleared it gets deeper the longer it sits.
Ranking follows sales velocity, so a listing that cannot sell loses position, and it loses review velocity at the same time. Recovery takes longer than the stockout did, which is why protecting cover on the top lines is worth more than the margin on the units it costs.
It should come down as cover approaches the lead time, and stop before the listing runs out. Clicks on an unavailable product are spend with no possible return, and the ranking damage continues while the campaign is still paying for traffic.
Weekly for the products that make most of the revenue, monthly for the long tail. The weekly review is short: what is below its reorder point, what has changed in the rate of sale, and what is aging. It is the rhythm rather than the depth that prevents stockouts.
Brand-controlled modules of images and text that appear below the buy box on a listing, replacing the plain text description. It is available to sellers enrolled in Brand Registry, and it exists to help a shopper who has scrolled past the images and bullets decide.
Not the way the description does. A+ modules are not indexed in the same way, which is why the keywords that lived in the description need a home in the bullets, attributes or backend fields when A+ replaces it. Its value is conversion, which affects ranking indirectly and more strongly.
Premium adds larger images, video, interactive comparison tables and hover content, and is eligibility-gated. The extra modules earn their cost on considered or technical purchases where a video or a bigger comparison changes a decision, and add little on a cheap, obvious item.
Usually, on listings with enough traffic to measure it. The gain comes from answering the objection that was stopping the purchase, so a set of modules written from your own reviews outperforms a template every time. Launch it alone, with nothing else changing, or the number cannot be attributed.
Enough to answer the real objections and no more. Four or five well-chosen modules beat a full page of them: a shopper this far down is scanning, and the modules after the decision point are not read.
Start with the products carrying most of the revenue and the ones with the most returns, because both have the most to gain. A long tail item with a handful of sessions a week will not produce a readable result either way.
The brand story module, yes, and it should be. The comparison chart should be shared within a family. The objection-answering modules should not, because the objection differs by product and a generic module answers nobody.
Yes. It is what a shopper using a screen reader gets, it is the only machine readable text in a module, and writing it properly costs a sentence per image. Describing what the image shows is the standard, not naming the file.
Submission goes through a review, usually measured in days rather than hours, and rejections are commonest for claims that cannot be substantiated, references to price or shipping, and contact details. Writing to the guidelines first is faster than appealing afterwards.
Where Manage Your Experiments is available for the brand, yes, and it is the cleanest way to answer whether a set of modules is doing anything. Without it, the alternative is a before and after with nothing else changed, which is weaker evidence but still evidence.
Yes. Amazon's generative AI tools in the A+ Content Manager are available to Brand Registered sellers in the US, on modules with the AI Ready badge, starting from a brand-owned ASIN and free for select English language modules. They draft copy and lifestyle imagery; the claims, specifications and images still have to be checked against the product before publishing.
Two jobs on one page. Relevance, which is making the listing eligible to appear for the searches you want, through the title, bullets, attributes and backend terms. And conversion, which is making the shopper who lands actually buy, mostly through images, the first bullet and price. Amazon rewards the second by showing you more of the first.
Conversion changes show within days on a listing with real traffic. Ranking changes follow sales velocity, so they take a couple of weeks to read and longer to settle. Anything judged inside a week is being judged on noise.
As many as fit naturally in something a person would read on a phone, which is usually the brand, the product, and one or two deciding attributes. The rest of your terms are already in the bullets, the attributes and the backend fields, and Amazon reads all of those.
They are indexed, so they contribute to relevance, and their bigger effect is on conversion, which feeds ranking indirectly and more powerfully. Writing them as answers to the objections that stop a purchase does both jobs at once.
On a considered purchase, usually yes, because it is where the comparison happens. It is worth less on a cheap impulse item where nobody scrolls that far. Remember that it replaces the plain description on listings that have it, so anything doing a job there needs to be carried over.
Review the ones carrying most of the revenue quarterly, and whenever something changes: a new competitor, a pattern in the reviews, a shift in what the search terms report shows. The rest of the catalogue can wait for a scheduled pass.
Usually because conversion dropped, not because the edit was rejected. A changed main image or a reordered title can cost a percentage point of conversion, and rank follows conversion. The other common cause is an attribute change that moved the listing into a different browse node.
Read them, do not copy them. Their listing is optimised for their reviews, their price and their objections. What is worth taking is the structure and the questions they answer, not the words.
They help when the variants are genuinely the same product in different sizes or colours, because reviews and rank consolidate. They hurt when unrelated items are forced into one family to inherit reviews, which is against policy and tends to end with the family being split by Amazon rather than by you.
On most catalogues, images. They are the first thing a shopper processes, they carry the objections in less space than any sentence, and they are usually the least maintained part of the listing. Attributes come first only because they are free and decide eligibility.
There is no keyword count. There are fields with character limits: the title, five bullets, the backend search terms and the structured attributes. A term indexed once from any of them is indexed, so the practical limit is how much you can say without hurting readability.
No. Amazon does not score a listing on keyword density, so a term in the title does not get stronger by appearing in every bullet. Repetition costs readable space and, in the title, click-through rate.
Mostly in the bullets and the backend search terms. Long tail phrases rarely read well in a title, and they are usually variants of something already indexed, so they belong where they can be present without being read aloud.
No. It is against policy in the visible fields and in the backend, and enforcement tends to arrive as a suppressed listing rather than a warning. Compatibility phrasing, where the category allows it, is the legitimate version of the same idea.
Search the term together with the ASIN. If the listing appears, it is indexed for it. That says nothing about ranking, which is a separate question decided by sales velocity and conversion.
Monthly on the products carrying most of the revenue, driven by the search terms report rather than by a tool. Demand vocabulary shifts with seasons and competitors, and the report is the only source that reflects your own buyers.
The plain description is indexed, though it is replaced on listings that have A+ content, and A+ text is not indexed the same way. Treat the description as a supporting field rather than a main one, and never as the place your important terms live.
Amazon handles most common misspellings itself, so deliberately adding them is usually wasted space. The exception is a genuine alternative spelling that people use, which is a different word rather than an error.
The shared terms, yes, because the family competes as one. The variant-specific terms belong on the variant, which is what makes a size or colour findable on its own.
Yes, in two ways. Terms that attract the wrong shopper pull conversion down, and rank follows conversion. And a title stuffed to the limit reads as spam, which costs the click before the listing gets a chance to convert.
A notice that something on the account has breached a policy, ranging from a listing that broke a category rule to a report that an item was not authentic. They arrive in the same format whatever the severity, which is why the first step is reading which policy has been cited rather than reacting to the tone.
Rarely on its own. Suspensions usually follow a pattern, an unanswered notice, or a category of violation serious enough to act on immediately, which in practice means authenticity and safety. A single product policy notice is a listing problem, not an account one.
Three parts: the root cause, what you have already done about it, and what prevents it recurring. Write actions taken with dates rather than intentions, and keep it specific to this violation. A plan that promises to be more careful is the commonest reason an appeal is rejected.
Days rather than hours, and longer where documents are being verified. Submitting a second appeal while the first is in review does not speed it up and frequently resets the queue, which is why the first one should be the considered one.
High enough that a single incident does not put the account at risk, and trending in the right direction. The composition matters more than the number: three violations of one kind is a process problem you can fix, one of each is noise.
By getting the rights owner to retract it. Amazon records the retraction; it does not adjudicate the dispute for you. Approaching the complainant politely, with the invoices that show a legitimate supply chain, resolves more of these than any appeal does.
The weight of an old violation falls over time, so an account with a clean recent history recovers. That is why the trend is worth watching rather than the total, and why a run of small violations matters more than one old serious one.
It depends what was actioned. A suppressed listing stops selling; a warning on the account usually does not. Where selling privileges are suspended, funds are typically held as well, which is the part that turns an appeal into a cash flow problem.
Where the account offers it, yes, and prepare for it. They can tell you what the reviewer is looking for, which is information the notice itself does not contain, and that is worth more than another written attempt.
Three things. Listings created quickly and never checked against the category rules, suppliers whose invoices do not meet the standard Amazon asks for, and customer service processes that let a defect metric drift. All three are operational rather than clever, and fixing them is the whole of prevention.
The referral fee is a percentage of the total the customer paid, set by category, and it is charged on every sale however it is fulfilled. It is the only fee that is a straight percentage. Everything else, fulfilment, storage, returns and advertising, is charged per unit, per month or per event, which is why the referral fee alone never answers the question.
The referral fee is what Amazon charges for the sale. The FBA fee is what it charges for handling the item: picking, packing, postage, customer service and the return. A merchant fulfilled order pays the referral fee and not the FBA fee, and pays the handling in labour instead.
Almost always because of the charges that do not appear against an order: monthly storage, aged inventory surcharges, removals, returns that came back unsellable, and advertising counted as marketing rather than as cost of sale. Per-order maths looks healthy while the account loses money over a year.
On the volume your inventory occupies in the fulfilment network, charged monthly and higher in the final quarter of the year. Stock that sits long enough also attracts an aged inventory surcharge. Both are charged for time rather than for sales, which is what makes slow moving stock expensive.
Partly, and it depends on the category and the reason. A refund administration fee is typically retained. The larger cost is usually the unit itself and its condition when it comes back, which is why measuring your own return rate matters more than the fee schedule.
For most catalogues, yes. Advertising is now the cost of being visible rather than an optional lever, so it belongs in the unit economics. The figure to watch is total advertising cost of sales across the account, because campaign level returns can look excellent while cannibalising organic sales.
Mostly by changing the product rather than by appealing a charge. Dimensions and weight decide the fulfilment band, so packaging changes move real money. Turning stock faster cuts storage. Fixing the listings that drive returns cuts the most expensive cost of all. Fee disputes and reimbursements recover what was charged in error, which is worth doing and is not a strategy.
For referral and fulfilment fees on a well categorised product, close enough to plan with. It knows nothing about your storage duration, your return rate or your advertising, which are the three that decide whether the product is actually profitable.
A charge applied to FBA products whose stock cover runs persistently low relative to demand. The intent is to discourage inventory that runs out and disappoints buyers. The practical effect is that inventory planning has a direct cost attached to getting it wrong in either direction.
Yes, by country and by category, and the differences are large enough that a product profitable in one marketplace can lose money in another at the same price. Any expansion plan needs the arithmetic done per marketplace rather than assumed from the home one.
Walmart's own fulfilment network. You send stock in, Walmart picks, packs, ships and handles customer service and returns, and the listing carries the fast delivery tag shoppers filter on. It is Walmart's equivalent of FBA, with its own fee schedule and its own prep requirements.
A per unit fulfilment fee set by weight and size band, plus monthly storage on the space the stock occupies, charged at a higher rate in the peak quarter. The band boundaries matter more than the headline rate: a product a few ounces over a threshold pays the next band on every unit it sells.
On contested listings, usually, because the delivery promise is an input into the buy box and fewer Walmart offers carry it than on Amazon. On a listing you already win with no competing offers, you are buying convenience rather than placement, and the arithmetic is different.
The mechanics are similar and the economics are not. Fee bands, storage rates and the competitive picture all differ, so a product that is clearly better in FBA can be clearly better seller fulfilled on Walmart. Copying the Amazon decision across is the commonest mistake.
Yes, and most established accounts do. The split usually falls out by size and velocity: small, fast moving items into WFS, large or slow moving items fulfilled yourself where your own freight rates and storage costs are better.
Walmart handles the customer and grades the unit for resale where it can. In a high return category that convenience is paid for in units that come back unsellable rather than in a fee, which is why measuring your own return rate per product matters more than the schedule.
Longer than the freight, and variable. The clock runs from production through freight, receiving and check-in, and check-in slows down when everybody is sending stock in at once. Plan the reorder point against the slow end of your own measured range rather than the average.
Not directly. It improves the delivery promise and buy box share, and those improve conversion and sales velocity, which is what ranking follows. The effect is real and it arrives through conversion rather than through any preference for the fulfilment method itself.
Specific, and enforced. Labelling, packaging and unit preparation all have requirements, and stock arriving wrong is either handled at a cost or refused. Reading them once properly is faster than discovering them one shipment at a time.
It depends on the cash more than the strategy. WFS commits money to inventory earlier and gives a competitive listing sooner; seller fulfilment keeps the cash closer to the sale while demand is unproven. For a first product with no history, the cash argument is usually the stronger one.
Easier to be visible, harder to set up. There is less competition, so a decent listing can hold a first page that would be invisible on Amazon. The catalogue is stricter, the price rules are enforced rather than influenced, and the fix cycle is slower, so mistakes cost more time.
Considerably less, and considerably less competition with it. The ratio is what matters to a seller rather than the absolute number: a smaller share of a smaller audience can be worth more than an invisible listing on a larger one.
Most often because the same item is available cheaper elsewhere on the open web, which frequently means on your own website during a promotion. Walmart enforces price through visibility rather than through the buy box, and the comparison does not care that the cheaper seller is you.
Not directly. The search behaviour, the attribute requirements and the shopper vocabulary differ, and duplicated copy tends to underperform on both. The research transfers; the words should be rewritten.
Referral fees are broadly comparable by category. Fulfilment is where the difference lies, because WFS bands, storage rates and return handling are their own schedule, so the per unit economics have to be worked out separately rather than assumed from FBA.
Cost per click is lower in most categories, and the platform gives you less control and thinner reporting. That combination rewards being early and getting the structure right, and it frustrates anybody expecting the granularity Amazon offers.
If the supply chain can support it, usually yes, with one inventory plan across both. Stock committed to one marketplace is stock the other cannot sell, so the allocation decision matters more than the listing work once both are live.
Longer than the setup and shorter than Amazon took. Item setup and category approval are the slow parts, and the ranking builds from sales velocity the same way, from a lower base of competition.
Yes, and price and delivery speed weigh heavily in it. Because fewer offers carry a fast delivery promise, fulfilment choice moves buy box share more on Walmart than the equivalent decision does on Amazon.
Running it as a copy of the Amazon account. Same copy, same keywords, same fulfilment decision, same advertising structure, and no allowance for the price rules. It produces an account that looks busy in the dashboard and sells very little.
For several units of the same item, use a fixed price listing with a quantity. For several versions of one product, use a variation listing, which holds sizes, colours or styles under one item number with its own stock count per variant. Those are different features solving different problems.
As variations, in almost every case where the items are genuinely versions of one product. Sales history, watchers and search position accumulate on one listing rather than being split across six, and one listing is less work to maintain.
No. Variations are for versions of the same product, and forcing unrelated items into a family to inherit its history is against policy. The usual outcome is the family being split by eBay, which loses more than it gained.
The variant that leaves loses the sales history it earned inside the family, and there is no mechanism to carry it across. Merging is easy and splitting is not, which is why the family should reflect how the product is genuinely structured rather than how it is convenient to list today.
It ranks as one listing with the whole family's history rather than as several listings with a fraction each, and eBay's search rewards sales history and conversion. That consolidation is the ranking benefit; the title, item specifics and category still do the work of being findable.
The listing usually stays live with that option shown as unavailable, and conversion falls a little while it is. A listing that runs to zero across every variant ends, and ending a listing with history is the most expensive stock mistake on eBay.
No. Auctions are single item by nature, so variations are a fixed price feature. Sellers who need both usually run the auction for discovery and keep the fixed price listing as the one that accumulates history.
No, and the restrictions change. It is worth checking the category before building the listing rather than after, because rebuilding a family into separate listings is the direction that loses history.
eBay allows a large number, and the practical limit is lower: a selector with dozens of options converts worse than one with a handful, and stock management across many variants is where accounts start running out on their best sellers without noticing.
No. A bundle is its own product with its own identifier, not a version of its components. Listing it as a variation of one of the parts confuses buyers and tends to attract the wrong ones.
It depends on the product. Amazon suits new, branded, standardised goods with a clear identifier, where the buyer already knows what they want. eBay suits used and refurbished stock, parts, anything where condition varies, and products that need explaining, because the seller owns the listing rather than sharing a catalogue page.
They are broadly comparable on the sale itself, with category variation on both. The larger difference is fulfilment: most eBay sellers ship themselves, so the cost is postage and labour rather than a fee band, which makes the comparison a per product exercise rather than a headline rate.
Usually, and it is a different kind of competition. On Amazon several sellers compete for one buy box; on eBay several listings compete in search. Being outranked is recoverable in a way that losing the buy box is not.
Yes, and most established sellers do. Run them differently: separate copy, separate pricing strategy, separate fulfilment decision. Share the inventory plan, because a unit committed to one is a unit the other cannot sell.
Less naturally than Amazon, because eBay's demand skews towards buyers hunting for something specific, including used and hard to find items. A new branded product with a barcode and existing demand usually finds Amazon the faster route, and does well on eBay once the listing has history.
eBay, comfortably. Condition is a first class concept there, the buyer expects variation, and the seller controls the description and the photographs, which is what selling a used item requires.
More than most sellers expect. Service metrics and seller standing feed into search visibility across the whole account, so late dispatches on one product can dampen listings that had nothing to do with it. On Amazon the effect is narrower and mostly reaches the buy box.
No. Amazon reviews and eBay feedback are separate systems and neither carries across, so a listing arriving on the other marketplace starts from nothing and has to earn its history there.
eBay, generally. Listing is quicker, the catalogue is less strict, and there are fewer category gates. Amazon takes longer to set up properly and offers more traffic once it is done.
Copying the account across. Same copy, same price rules, same fulfilment model, same expectations. The two marketplaces reward different behaviour, and the account that treats them as one usually underperforms on both.
They solve different halves of the problem. Amazon supplies demand and charges for it; Shopify supplies the shop and leaves the demand to you. For most product businesses the sequence is to prove demand on the marketplace and build the direct channel with the margin it generates.
Per order, the direct store, because there is no referral fee. In total it depends entirely on what the traffic costs, which is the number missing from most comparisons. A store with expensive paid acquisition can be less profitable than the marketplace it replaced.
Yes, and most established brands do. Share the inventory plan, because a unit committed to one channel cannot ship from the other, and be deliberate about price: marketplaces compare against the open web including your own site, and on some of them a cheaper direct price can suppress the listing.
Yes, and that is the strongest argument for a direct channel. You get the email address, the purchase history and the ability to sell again with no acquisition cost. On a marketplace the relationship ends at the order, and the buyer belongs to the marketplace.
None, until somebody buys it or earns it. That is the single most important sentence in this comparison. Search compounds slowly, paid social works immediately and stops when the spend does, and email only exists once one of the other two has delivered customers.
On the marketplace, unless the brand already has an audience. Buying traffic to test an unproven product means paying to learn something the marketplace would have told you for a referral fee. A brand people already know should start direct, because demand is not what it is missing.
Not directly. What can hurt is price: marketplaces compare your listing against the same item elsewhere, your own site included, and a promotion on the store can affect the listing's buy box or visibility depending on the marketplace. Coordinate the promotions rather than the platforms.
The subscription, payment processing and apps are the predictable part and rarely the large one. Acquisition is the real cost, and it is the number to model first, because it decides whether the extra margin per order survives the traffic that produced it.
Not directly, and attempting it through package inserts or messaging breaks marketplace policy. What works is being findable when the customer looks for the brand later, which is a search and brand problem rather than a redirection one.
Whichever earns more per unit after every cost, including the traffic. That is usually the direct store for repeat purchase products with a warm audience, and the marketplace for products whose demand is entirely search driven. It is an arithmetic question and it is worth doing per product.
Making a page the source an assistant draws its answer from, rather than one of ten links a searcher might click. In practice it means answering questions plainly and near the top, keeping facts consistent across the site, and emitting structured data so a machine can tell which part of the page is the answer.
It overlaps heavily and the emphasis differs. Classic SEO competes for a position in a list; AEO competes to be the sentence that gets quoted. Almost everything that helps the second also helps the first, which is why it is not a separate programme of work.
Product markup with price, availability and identifiers; FAQ markup where the page genuinely answers questions; breadcrumbs; and organisation markup so the brand reads as an entity. Themes emit some of it, apps sometimes emit it twice, and duplicated markup is worse than none.
Both, and product pages are where most stores have the most to gain. A product page that states dimensions, materials, compatibility, delivery and returns plainly is answerable; one that lists features and hides the facts in a table image is not.
From the support inbox, the reviews and the marketplace question sections. Anything asked twice belongs on the page. That source beats a keyword tool because it is your own buyers using their own words.
Apps can emit markup, and they cannot decide which facts matter or write an answer that reads well. They also frequently duplicate markup the theme already emits. Use them for the mechanical part and check the output.
Yes. Clear answers, consistent facts and correct structured data are what rich results and featured snippets have always rewarded. The change is that the same qualities now decide whether an assistant quotes you, not just whether a snippet appears.
Structured data fixes are picked up within days to weeks. Being quoted more often is slower and harder to measure, because there is no rank position to watch. Track branded search and direct traffic alongside the usual metrics rather than expecting a single number to move.
It helps, provided the facts agree. Answer engines favour sources corroborated elsewhere, and a product whose specification matches across your store and every marketplace is more credible than one that contradicts itself between channels.
Facts that live only inside an image. A specification table saved as a picture is invisible to everything that reads the page, and it is the commonest reason a well built product page cannot be quoted.
Register as a seller with your business documents, pass category eligibility, connect a bank account, then build the catalogue and choose fulfilment. Approval is usually days. The setup that decides whether it sells is the affiliate commission and the content plan, and neither is part of the application.
A registered business, documents that match the name on the application, a bank account in that name, and a product category that is eligible in your market. The commonest delay is a mismatch between the documents and what was typed, not the review itself.
Whatever your contribution margin can carry, worked out before you publish a rate rather than copied from a competitor. Too low and no creator takes the product; set without knowing the margin and the channel grows while losing money on every order.
Somebody does. A shop with no video is a catalogue on a platform where nobody browses catalogues. The three routes are your own account, creator content through the affiliate programme, and live selling, and most brands use all three in different proportions.
It is worth it where the product demonstrates well on video and the margin can carry a commission. It is a poor fit for products that need explaining in text or that sell on specification. The commitment is a content rhythm rather than a launch, and a brand unwilling to make that commitment is better off not opening the shop.
Ship yourself or use the platform's service where it is available in your market. Either way dispatch times are enforced and feed seller performance, which affects how much reach the shop gets. Late dispatch is a distribution problem here, not only a service one.
Yes, and treat it as a budget rather than a favour. A creator holding the product makes a materially better video than one working from a product page, and the creators worth having are asked constantly, so the offer has to be easy to say yes to.
Amazon captures demand that already exists; TikTok creates it. That changes everything downstream: the product research, the imagery, the pricing and the pace of stock. Running it as a second Amazon account produces a catalogue nobody sees.
After something works organically. Paid amplification of a video that is already converting beats paid promotion of a product nobody has made a video about, and it costs less per order because the creative has been proven first.
Treating it as a listing exercise. The catalogue takes a week and the channel takes a rhythm, so the honest question before opening a shop is who is making a video this week, every week.
A decision about which of four levers you are pulling this quarter: traffic, conversion, average order value or repeat purchase. Everything sold as a growth tactic is one of those four with a different name, and at any moment one of them is the constraint that decides what the others are worth.
Write down all four numbers for the last quarter and compare each against what the category supports rather than against your own last year. The one furthest below what is normal is the constraint. Deciding it from the channel you are most comfortable in is the mistake that costs a quarter.
Almost always, because the traffic is already paid for and a conversion fix improves every future visit as well. Buying traffic into a page converting at half the category norm is buying more of the same disappointment at the same price.
For most product businesses, the marketplace, because the demand already exists and you pay for access rather than for creating it. Then search, which compounds slowly. Then paid, to accelerate what is already working rather than to discover whether anything does.
It is frequently the fastest of the four and the most ignored, because it feels less like growth than winning new customers. It needs no new traffic and no new trust: bundles, multipacks and offering the thing somebody would have bought anyway if they had thought of it.
It sets it. A brand whose customers buy three times can pay more for the first order than one whose customers buy once, and still make more money. That is why repeat rate is a growth lever rather than a retention metric.
A quarter, with one constraint and one measurement. Shorter than that is reading noise, and wider than that becomes a list of activities nobody can evaluate. Write down in advance what would make you stop.
Both, honestly. They are the fastest access to demand and they keep the customer relationship, which caps the repeat purchase lever. Most brands use them to fund the direct channel rather than choosing between them.
Working every lever at once. Four half-funded initiatives produce four inconclusive results and no way to tell which of them was working, which is how a year passes with a lot of activity and no compounding.
Whatever the contribution margin and the repeat rate support, which is a number your own accounts answer and no benchmark can. Anybody quoting a target before seeing your margins is guessing at the most important input.
As many as it takes to answer the questions that stop a purchase, which is usually six or seven: the main image, scale, detail, in use, what is in the box, a comparison where a range exists, and an infographic answering the commonest objection.
Most marketplaces require a plain white background, the product filling most of the frame, and no text, badges, props or watermarks. The rules are enforced with suppression rather than a warning, so they are worth reading for the category you sell in.
Indirectly and strongly. Ranking follows conversion, and images are the largest single influence on whether a shopper who lands buys. A better main image also raises click-through from search, which compounds the effect.
Whichever produces consistent, correctly specified images at the volume your catalogue needs. A studio is worth it for hero and lifestyle work and for a large first shoot. Ongoing variants and simple detail shots are frequently better in house once a repeatable setup exists.
No scale shot. Items arriving smaller than the buyer imagined is one of the largest return reasons in most categories, and one photograph with a familiar object beside the product prevents more of them than any amount of copy.
Yes, in the secondary slots, and they are usually the best performing images after the main one. Keep them to a few words answering a real objection, and check the marketplace's rules on which slots allow text.
It is a returns problem rather than a taste one. A product that arrives a different shade than the photograph promised comes back, and it costs the shipping both ways plus a review that will deter other buyers.
The same shoot, not the same crops. Marketplaces, product pages, advertising and social all want different framing, so the shot list should be planned against every channel before the day. Reframing a horizontal image into a vertical placement rarely survives.
When the evidence changes: a new objection in the reviews, a competitor doing something better, a variant added. A catalogue where the images have not been touched in two years is usually leaving conversion on the table.
On considered purchases and anything that needs demonstrating, yes, and it is the same brief: answer the objection. On a simple, obvious product it adds less than a better set of stills, so the money is usually better spent on the images.
Whatever you set, plus whatever the auction charges per click. The useful number is not the click price but the cost per order, which is the click price divided by your conversion rate. A cheap click on traffic that never buys is more expensive than a dear one that does.
There is no answer across categories, because the auction price follows what a customer is worth. The only meaningful ceiling is your own: contribution margin per order multiplied by conversion rate. Above that you are buying revenue at a loss, however good the number looks against an industry average.
Enough for the conversion event to happen often enough to learn from, concentrated on the products that already convert. Five campaigns funded at a fifth each produce five inconclusive results; one funded properly produces a decision.
Usually competition, seasonality or a relevance change on your side. It is an auction, so a competitor raising bids moves your price without you changing anything. What matters is whether the cost per order moved with it.
Where the margin can carry the auction price for the terms that matter, yes. Where it cannot, the arithmetic says so before any money is spent, and the honest answer is to work on the channels where demand already exists rather than to spend a quarter proving it.
Sometimes, and it is worth measuring rather than assuming. If nobody is bidding against you and your organic result is first, brand clicks are frequently traffic you were getting free. If competitors are bidding on your name, the calculation changes.
Usually a percentage of spend or a flat retainer, and it should be quoted after an audit rather than from a rate card. The fee matters far less than the waste: on an unmanaged account the money lost to irrelevant traffic is routinely larger than any management fee.
Waste comes out in the first month and shows immediately. Structural changes need a fortnight of stable spend to read, because every significant edit costs the campaign its learning. A fair verdict on the channel is about ninety days.
Google, directly, from your own account. The management fee is separate. Keeping the account in your own name matters more than the billing arrangement: it is your data, your history and your asset if the relationship ends.
Read the search terms report and add negatives. On most accounts that have not had it done recently, it removes a material share of the spend in an afternoon without touching a single bid.
If the product sells mainly on Amazon, start there: the intent is higher and the loop is closed, so the cost per order is usually lower and the measurement is cleaner. Google earns its place when you are building a direct channel, when the category has demand Amazon does not capture, or when you need reach earlier in the decision.
Per click, usually Google in most categories. Per order, usually Amazon, because the intent is higher and the conversion rate follows. Cost per click is the wrong comparison and it is the one everybody makes.
Yes, and it works in specific cases: a launch, a product with weak organic rank that needs velocity, or a term where the listing converts far better than your own site. You pay Google for the click and Amazon its referral fee on the sale, so the margin has to carry both.
Frequently because the purchase finished somewhere Google cannot see, on a marketplace listing rather than on your site. A brand judging Google on last-click site revenue alone systematically undervalues it, which is why blended reporting and holdout tests exist.
Less than people assume. They buy different moments: Google reaches somebody still deciding, Amazon reaches somebody already shopping. The genuine overlap is brand terms, where both can be paying for demand that already existed.
Amazon, usually, because it puts the product in front of shoppers already looking for the category and generates the sales velocity that ranking follows. Google is better once you know the product converts and you want demand the marketplace is not capturing.
Do not compare their ROAS figures directly; they are counted differently and both exclude what happened on the other. Judge contribution margin across the business over a period long enough to include the delay between attention and purchase.
On Google Shopping the feed is the campaign, and on Amazon the listing is. In both, product data does the work keywords do in a text campaign, and in both, most accounts never touch the raw export from their store platform.
The research transfers, the lists do not. Amazon search terms are shopping language and Google's include research and comparison intent, so a keyword set copied across brings the wrong half of the audience on one of the two.
Put it where the demand already is, which for most product businesses is the marketplace, and add the other once the first is steady and measured. Splitting a small budget across two auctions usually produces two campaigns too thin to learn from.
They do different jobs. Google captures demand that already exists and Meta creates it. A product people search for needs Google; a product nobody knows about needs Meta, and no amount of search budget fixes an awareness problem.
Usually Google on last click, because it is credited with the purchase that ends a decision Meta may have started. On blended numbers the gap narrows considerably, which is why the comparison is worth doing across the business rather than platform by platform.
Largely attribution. Someone who saw a Meta ad and later searched your brand is recorded by search. That is what last click means rather than a fault in either platform, and it is the commonest reason brands cut the spend that was filling their funnel.
Watch blended acquisition cost, branded search volume and total revenue rather than platform ROAS alone. When the answer genuinely matters, run a holdout: turn it off in one region or for one audience and measure what happens to the total.
One, funded properly. Two half-funded accounts produce two inconclusive results. Which one depends on whether the demand exists: if people search for the category, start with search; if they do not, start with Meta.
By constraint rather than by ratio. If searches for your category are going to competitors, that gap is the most expensive one. If search is captured and growth has stalled, the constraint is how many people know the product exists, which is Meta's job.
On Meta it is the main lever, because targeting is largely automated and the auction rewards the account with more distinct angles. On Google creative matters and structure matters more: keywords, match types, negatives and the feed.
They can, and being good at one does not imply the other. Search is a discipline of structure and exclusion; Meta is a discipline of creative volume and signal quality. The best accounts we take over are usually strong at one and neglected on the other.
It is a third auction with the highest intent of the three, and it frequently takes the credit for purchases the other two started. Any brand selling on a marketplace should read all three together, because judging each in isolation guarantees at least one of them is being undervalued.
Long enough for the conversion event to happen often enough to be read, and stable enough not to reset learning halfway through. For most eCommerce accounts that is a fortnight minimum per meaningful change, and ninety days before judging a channel.
Below your break even, which is your contribution margin as a percentage of price. That number is specific to the product: 15% is disastrous on a 12% margin and generous on a 60% one. Any target quoted without seeing your margins is a guess at the most important input.
Total advertising cost of sale: ad spend divided by total sales, organic included. It is the only one of the three ratios that says whether advertising is building a business or renting one, because it is the only one that can see the organic sales the advertising caused.
They answer different questions. ACoS is the right number for a bid or keyword decision; TACoS is the right number for a budget or channel decision. An account judged only on ACoS can look healthy campaign by campaign while becoming more dependent on advertising every month.
With sales rising, it means organic is growing faster than the spend, which is advertising doing its job. With sales flat it means you are spending less and selling the same, which may be efficiency or may be a top of funnel you have quietly switched off.
Take the selling price, subtract the cost of goods, the referral fee, fulfilment, and the share of returns and storage the unit carries. What is left, as a percentage of the price, is break even ACoS. Above it the marginal sale loses money.
Because ACoS knows nothing about your costs. It compares spend against advertising revenue, not against margin, so a product with thin margins can post an excellent ratio and lose money on every unit. Profit is contribution margin after every cost, and that is a different calculation.
Not on the ratio alone. A campaign at a high ACoS that is lifting a listing's rank is buying something the ratio cannot see, and cutting it is the commonest way an account is optimised into decline. Look at what happened to total sales and to TACoS before deciding.
It is the same relationship the other way up: 25% ACoS is 4x ROAS. It exists in Amazon conversations because people arriving from Google and Meta think in ROAS. Pick one and be consistent, because using both is how two people disagree about an account they both understand.
There is no universal figure, and the direction matters more than the level. A mature catalogue with strong organic rank runs a lower TACoS than one launching products, and a rising TACoS during a launch is expected rather than alarming. What matters is whether it falls as the listings mature.
All three, with contribution margin above them. The ratios are context; the margin is the verdict. A report that leads with ACoS invites the decision that improves ACoS, which is not always the decision that grows the business.
The report of actual shopper queries that triggered your ads, against the keywords or targets that matched them. It is the only source in the account for what people really typed, as opposed to what you bid on, and the gap between those two is where wasted spend lives.
The keyword report shows what you chose to bid on. The search terms report shows what shoppers actually searched. A single broad keyword can match hundreds of queries, and only this report says which of them were your product, a competitor's, or nothing to do with you.
Weekly for the campaigns carrying most of the spend, monthly for the rest. The cadence matters more than the depth: read weekly, a term burning budget is caught after seven days instead of after a quarter.
When it has enough clicks with no sales that the absence means something, which depends on your conversion rate. Converting at one in ten, forty clicks and no sale is a clear signal and ten is not. Negating too early removes terms that were converting normally for the category.
When it has converted more than once at an acceptable cost. Giving it an exact match keyword with its own bid takes it out of the broad match lottery and prices it at what it is worth to you.
Yes, and this is the step most accounts skip. A query that converts in advertising is a query the listing should rank for organically, and carrying the terms across is the whole of the loop between paid and organic.
Amazon withholds rare queries to protect individual searches, so the rows always sum to less than the total. That is expected rather than a fault, and it is worth knowing before somebody spends an afternoon trying to reconcile the difference.
No. It shows what your ads matched, not what the market searched. A term missing from it may be a term nobody searched, or one your bids never reached, and those are different problems.
Automatic campaigns, because discovering terms nobody thought of is their entire job. An automatic campaign that nobody mines is a discovery budget with nobody collecting the discoveries.
The collection and the obvious cases, yes: pulling the report, flagging terms past a click threshold with no sales, and surfacing converting terms not yet in the listing. What cannot be automated is the judgement about a term that is close to your product but not quite it, which is where the interesting decisions are.
Start with categories where several listings sell steadily rather than one dominating, read the negative reviews on the top five for a fixable, repeated complaint, and work out the margin after every fee including returns and advertising. The last step rules out more ideas than the first two combined.
Enough that the third, fifth and tenth listings all sell steadily, not just the first. A niche where the top seller does a hundred a day and the tenth does one has a single winner, and being tenth in it is worth nothing.
On their own, no. A listing with four thousand reviews at 3.9 is more beatable than one with eight hundred at 4.8, because the first has a documented list of what is wrong with it and the second has satisfied customers.
Enough to survive advertising and returns at the category's real rates, which is the calculation people skip. A 35% gross margin in a category with a 12% return rate and expensive clicks is not a 35% margin business, and the difference is knowable before ordering anything.
More than almost anything else. Fulfilment is priced in bands, so a few ounces over a threshold costs the next band on every unit for the life of the product, and small and light also wins on storage, freight and returns.
Mostly you cannot, so the question is what makes copying unprofitable: a trademark and Brand Registry, a genuine design or bundle, a supplier relationship that is exclusive in practice, or a brand people ask for. Everything else is a head start rather than a moat.
Avoid the ones you cannot realistically get ungated in, and treat the ones you can as an advantage: a gate that keeps you out keeps others out too, and a seller who has cleared it faces less competition inside.
For the demand estimate and the shortlist, yes, with the caveat that their sales figures are estimates and everybody with the same subscription sees the same list. The judgement that follows, about margin, defensibility and whether the complaints are fixable, is where the decision actually happens.
Enough to test the demand and small enough to be wrong about. The first replenishment carries the real information, so plan to reorder early rather than to order deep, and treat the first batch as buying data as much as inventory.
Choosing on demand alone. A high volume product with thin margins, expensive clicks and a high return rate is a busy way to lose money, and every one of those figures was available before the order was placed.
A programme that ties a registered trademark to a selling account and, in exchange, gives the brand control of its own detail pages plus a set of tools: A+ content, Brand Story, Brand Stores, Vine and brand analytics. It is control and tooling rather than a legal shield.
Yes. A registered mark in the marketplace you are enrolling for is the gate, though pending applications are accepted in some jurisdictions through the accelerator programmes. A brand with no mark at all cannot enrol, and that is the honest answer to most enquiries.
No. Anyone holding genuine stock can offer the same product, and the buy box is decided separately. What Registry stops is other sellers overwriting your title, bullets and images, which is the complaint unregistered brands raise most often.
For most brands, the tools rather than the protection. A+ content and a Brand Store change conversion, and Vine solves the launch problem where nothing sells without reviews and reviews come from sales. The enforcement route matters when it matters and is used far less often.
Days once the trademark is in order, and the delays are almost always about proving the brand rather than the review itself: the account name not matching the mark, or images that do not show the brand on the product.
A programme that puts a new product in front of Amazon's own reviewers in exchange for units. It is the sanctioned way to get early reviews, it is available to registered brands, and the reviews are honest, which means a weak product gets honestly reviewed as one.
Yes, and it is the most ignored thing Registry unlocks. Search term share says what shoppers typed and which listings they clicked and bought, which is market data Amazon gives to nobody else, and repeat purchase behaviour is the only view of loyalty a marketplace seller gets.
You can report genuine intellectual property infringement, and the report carries weight because the brand is verified. Reports that are not well founded damage the account making them, so the tool being easy to use is not a reason to use it loosely.
It is the route by which a brand reports them and gives Amazon what it needs to act, and it works better with Transparency or Project Zero on top for a brand with a real counterfeit problem. Registry alone is the reporting channel rather than the prevention.
Correct the titles and images on every ASIN, because those changes stick now in a way they did not before. Then A+ content on the products carrying most of the revenue, the Store, and Vine on anything launching.
The Add to Basket button on a product page that several sellers offer. One seller holds it at a time and takes the overwhelming majority of that page's sales; every other offer sits behind a link most shoppers never open.
The share of page views where your offer held the button. Seller Central reports it per product and across the account, and it is the number to read when sales fall without traffic falling.
No, and this is the most common misunderstanding. Price counts as the landed price, item plus delivery, and it is weighed against fulfilment speed, stock and seller performance. A slightly dearer FBA offer beats a cheaper merchant-fulfilled one routinely.
No, but it is the strongest single input, because the offer is competing on Amazon's own delivery promise rather than yours. A merchant-fulfilled seller can win with genuinely fast, genuinely reliable delivery and good metrics, and has to be better on the rest to do it.
Almost always because the other side moved: a competitor restocked or repriced, a faster seller arrived, or your own stock dropped low enough that Amazon began splitting the page. Check your stock and the other offers before changing your price.
Amazon suppresses it when it judges the price too high against its own reference, and then the page shows See All Buying Options instead of a button. Everybody's conversion falls, so it is worth catching quickly.
In effect, yes. It rotates between offers that are close on every measure, so two comparable sellers split the page over a day. A percentage in the seventies is often that rather than a fault.
If their stock is genuine you cannot stop them offering it, so the answer is to be the offer Amazon prefers: fulfilment, stock cover, landed price and performance. Where the stock is not genuine, that is a Brand Registry matter, and where it keeps arriving, it is usually a distribution contract matter.
Not for the same button. Conditions have their own buy boxes, so a used offer does not take the new one, though it does sit on the same page and can take the sale from a shopper who was happy either way.
On a competitive catalogue, yes, with a floor that reflects your real costs including fees, returns and advertising. Without a floor a repricer eventually finds the price at which every sale loses money, and it finds it faster than you will notice.
Yes. Featured Offer is Amazon's name for the Add to Basket button, and the figure Seller Central reports as Featured Offer percentage is the buy box share; buy box is what sellers call it. The same measures decide both because they are one thing.
The separate seller-eligibility step, yes. From July 2026, rolling out across its stores through the end of the year, Amazon no longer decides first whether a seller is eligible on account performance and then ranks only the eligible; every seller's offers are ranked. Performance still counts inside that ranking, so a poor late-shipment or defect rate still costs the button, just not on every listing at once.
Not as such. Amazon says the fulfilment method, the Prime badge, customer reviews and star ratings are not factors. What is scored is what those things usually bring: delivery speed, free shipping, free returns and stock. FBA wins the button so often because it supplies all four by default, not because it is FBA.
The front page of Amazon's seller console: today's orders, recent messages, account health at a glance and whatever Amazon is promoting. It is a heartbeat rather than a report, and daily variance is wide enough that acting on it is usually a mistake.
The business report by ASIN. Sessions, page views, unit session percentage and buy box percentage per product separate the three ways a product can fail: nobody arrived, they arrived and did not buy, or you did not have the buy box.
Daily, briefly, and not for the score. What matters is whether something new has appeared, because a notice answered the day it lands is a reply and the same notice three weeks later is an appeal.
The share of visits to your product page that ended in a unit sold. It is the conversion rate for that listing, and it is the number to read before spending anything on more traffic: sending visitors to a page that does not convert is the most expensive mistake available.
Stock in Amazon's warehouses that cannot be sold because of a listing problem rather than a physical one: the listing was closed, deleted, or is missing something required. It is silent, it costs storage while it sits, and it is usually a few minutes to fix.
The payments report, which itemises the fees rather than netting them into a payout figure. Reading it once a month is how spreadsheet assumptions get corrected before they turn into a category that quietly loses money.
Read the reasons rather than the rate. A rising return rate on one product with the same reason repeated is either a listing that oversold it or a product that underdelivers, and knowing which is the difference between rewriting a page and changing a supplier.
A set of reports available to brands enrolled in Brand Registry: search term share, what shoppers bought instead of you, and repeat purchase behaviour. It is the only view of the wider market Amazon gives a seller, and most enrolled brands never open it.
The pulling, yes, through reports and the API. The reading, no. The value in these screens is in noticing that two numbers disagree, which is a judgement rather than a threshold.
Long enough for the recent past and not long enough for a year on year comparison in every report, so pull anything you want to compare later on a schedule and keep it yourself. The reports also finalise on a delay, so the last two days of any window are provisional.
An instruction to Amazon to take stock out of its fulfilment network. You can ask for it to be returned to an address, disposed of, or liquidated, and each has a different per unit cost and a different recovery.
Return when the goods are worth more than the shipping and you have somewhere to sell them. Disposal when they are genuinely worth less than getting them back. Liquidation when you want a small recovery with no handling and do not mind where they end up.
Units Amazon has graded as not sellable, mostly customer returns plus anything damaged in the network. They sit in your account waiting for an instruction and pay storage while they wait, which is why standing removal settings are worth turning on.
Days rather than hours, and longer when everybody wants one at once: before the storage assessment dates and around peak. Place it a fortnight before the date you are trying to beat rather than in the same week.
Usually, unless you can show it will sell through soon. Compare the storage between now and a realistic sell-through date against what removal costs and recovers. Holding aged stock through another peak in hope is a payment for optimism.
Yes, once it has been returned to you, inspected and repacked to the requirements. Sending it straight back without inspecting it is how the same units become unfulfillable a second time.
No. It is an inventory instruction, not a performance event. What does affect you is leaving unfulfillable stock and stranded listings sitting there, because both are visible signs of an account nobody is running.
Yes. Units lost or damaged in the network are reimbursable, and a removal is a bad moment to discover you never claimed. Check what is owed first, then remove what is left.
It recovers a fraction of the value with no handling, which beats disposal and loses to a return where a real second-hand market exists. For anything with resale value, a return plus a clearance listing elsewhere usually recovers more.
Order shallower and reorder more often, plan the tail of seasonal stock before it ships, and review returns rather than letting them accumulate. Removal orders are a symptom of purchasing decisions made months earlier.
Amazon's own ten character identifier for a detail page. Amazon assigns it, not the seller, and every offer on that page shares it: one product, one page, however many sellers.
A barcode, a GTIN, identifies the product in the world and is issued by GS1. An ASIN identifies the page on Amazon and is issued by Amazon. The barcode is how Amazon matches your product to the right page in the first place.
Only to create a new detail page. If the product is already listed, you add your offer to the existing ASIN and no barcode is needed, which is the step new sellers most often skip before creating a duplicate.
You can, and it is a false economy. Amazon checks the number against the GS1 database, and a barcode registered to another company's name gets the listing removed, sometimes months later with the stock already in a warehouse.
Permission to list without a barcode, granted per brand and per category, for products that genuinely have none: handmade goods, some private label lines, parts sold by brands that do not barcode them. It is granted on evidence, usually images showing the brand on the product and packaging.
Your own code for your own stock. You invent it, it never leaves your account, and two sellers on the same ASIN have different ones. Never reuse an SKU for a different product: sales history, stock and returns all hang off it.
The label FBA puts on a unit so Amazon knows whose it is inside the warehouse. It matters most when stock would otherwise be commingled and when a return needs to come back to the right seller.
Length and geography. A UPC is the twelve digit form common in North America and an EAN the thirteen digit form common in Europe. Both are GTINs and Amazon accepts both.
Yes. Walmart and eBay use the same standards to match your offer to their catalogues, so a product identified properly once is easier to list everywhere afterwards. The ASIN is the part that does not travel, because it is Amazon's.
The reviews and sales history split across two pages, and eventually Amazon merges them or the brand reports it. Neither outcome is under your control, so search the catalogue thoroughly before creating anything new.
Finish the listing completely, secure stock cover for at least the first months, get early reviews through Vine and the request a review button, and advertise on exact match against the terms the product genuinely is. The order matters more than any single tactic.
Fewer than most people think, and the rating matters more than the count. A handful of genuine reviews above four stars converts far better than none, and a product sitting at 3.5 with thirty reviews is harder to rescue than one with three.
For a registered brand launching something new, usually yes. It is the sanctioned way through the opening problem, and the reviews are honest, which means a weak product gets honestly reviewed as one. Learning that in week two is cheaper than learning it in month six.
No. Anything offered in exchange for a review is against policy, including inserts that offer a gift, and the penalty falls on the account rather than the product. The request a review button is the compliant way to ask.
Worse than your target, and that is expected: a launch buys ranking rather than profit. What matters is the trend across weeks, cost per sale falling and organic sales appearing. Judging the first fortnight as a verdict is how launches get killed early.
Usually, and not so deep that returning to the target price collapses the conversion rate. Plan the way off the launch price before you start, and prefer a coupon or a discount badge to the same money taken off the price, because the badge is visible in search results.
Enough to survive the launch itself plus the lead time on a reorder. Running out in week three loses the ranking the advertising just paid for, and regaining it costs more than the stock would have.
Exact first, on the few terms the product genuinely is. Broad and automatic campaigns are discovery tools, and discovery on a product with no history spends money finding out things the listing already told you.
Weeks rather than days to see whether it is working, and months for organic rank to settle. If cost per sale is not falling and no organic sales have appeared by the end of the first month, the problem is the listing or the price rather than the bids.
Advertising a half-finished listing. The conversion rate recorded in the first fortnight is the one Amazon uses to decide whether to show the product organically, so paying for traffic to a page that is not ready damages the thing the launch was for.
Yes. Amazon requires commercial liability insurance within 30 days once gross proceeds from Amazon.com sales exceed USD 10,000 in any month, with at least USD 1 million of cover per occurrence and in the aggregate. From 2 November 2026, sellers in safety-enhanced categories such as children's products, cosmetics and ingestibles, and lithium batteries need that cover whatever their sales.
Commercial general liability, which covers product liability. Amazon has to be named as an additional insured and has to be able to verify the policy, so a policy that satisfies you can still be refused on a paperwork point.
Not to satisfy Amazon. You may well want it anyway if you put your own brand on anything that goes in, on or near a person, because a private label seller is the responsible party in a way a reseller is not.
No. Amazon accepts individual and sole trader accounts. Incorporating is about separating the business's liabilities from your personal finances, and whether that is worth the administration depends on your jurisdiction and your profits.
Yes, and it is a verification event: the account details have to be updated and re-verified. It is doable while trading and it is not the week to also be launching a product.
The account, the bank account, the tax registration and any document you upload. Verification delays are almost always a mismatch somewhere in that list rather than a judgement about the business.
No. It is the general shape of two questions, and the specifics depend on where you are and what you sell. The last step is somebody qualified in your own jurisdiction, and it is much cheaper before a problem than after one.
It depends on the category: safety marking, test reports, ingredient or material declarations, and supply invoices for authenticity. The accounts that survive an inspection are the ones that had them ready rather than the ones that went looking.
No. Liability insurance covers harm caused by a product, not commercial losses from a policy decision. Nothing insures against a suspension, which is why account health is managed rather than covered.
Each sets its own terms, and the underlying documents are the same, so a seller organised for Amazon is usually most of the way there for Walmart and eBay. Read each agreement rather than assuming the thresholds match.
Sellers listing children's products, cosmetic and ingestible products, lithium battery products and other safety-enhanced categories must carry USD 1 million of commercial liability cover regardless of the USD 10,000 monthly threshold, and sellers based in mainland China must obtain new policies through the Amazon Insurance Accelerator.
One system that knows what stock exists, which channel sold it and where it ships from, with every marketplace reading from that single number instead of keeping its own. The distributed part is the stock and the fulfilment; the management is having one authority over both.
Make one system the owner of the stock number and have every channel read from it, then shorten the sync interval until a buffer is no longer needed. Buffers prevent oversells by capping sales on exactly the products with the thinnest cover.
Roughly above two channels, a few hundred SKUs, or any product held in more than one place. Below that a disciplined spreadsheet and one responsible person genuinely works and costs nothing.
How often it syncs, what happens when a channel's API is unavailable, whether it maps your existing SKUs rather than forcing a rename, and whether returns go back into stock automatically. The last one gets skipped and is what causes numbers to drift.
Yes, through multi-channel fulfilment, and it needs a real integration: something has to raise the fulfilment order and pass tracking back before the other channel's dispatch deadline. At any volume that is not a manual job.
It is simpler to run and ties up more money, and it is a legitimate choice. What causes trouble is drifting into it without deciding, so that nobody can say which pool a given unit belongs to.
Often enough that the buffer can come down. The right interval is the one where two channels cannot both sell the last unit, and on fast-moving lines that is minutes rather than hours.
More than the order. A cancellation is a performance metric on every marketplace, it can affect the buy box on Amazon and seller standing on eBay, and a run of them attracts attention that a single one does not.
They are the part most setups handle worst. A return that goes back into sellable stock on one channel and stays invisible to the others makes the numbers drift a little every week until somebody counts by hand.
It should be. A store reading the same stock number as the marketplaces is one more channel; a store keeping its own numbers is a fourth thing to reconcile, and it is usually the one that gets reconciled least often.
Changing your price automatically in response to competing offers, to win or keep the buy box on Amazon or the equivalent decision elsewhere. It is a rules engine, and the rules rather than the tool decide whether it makes money.
Landed cost plus every fee the sale actually incurs: referral, fulfilment, storage for the time it really sits, returns at the category's real rate, and advertising at what your clicks cost. Cost plus a round margin is how floors end up below break-even.
It will if every seller's rule is to undercut with no floor. Yours cannot control theirs, so the protection is a floor you will not cross and the willingness to lose the buy box below it.
Yes, and this is where most of the profit is. When a competing offer goes out of stock or withdraws, a rule that raises your price captures the margin; matching only downward leaves it on the table permanently.
No. An offer from a seller with poor metrics or no real stock is a price that is not genuinely available, and matching it gives away margin against a competitor who cannot deliver anyway.
Minimum advertised price, a term some brands set for their resellers. Where it applies your floor is contractual as well as commercial, and breaching it risks the supply rather than the sale, which is a much more expensive outcome.
Yes. Amazon compares your price against the rest of the web, so a deep discount on your own store can suppress the buy box on the marketplace. Price across channels rather than per channel.
Buy box share should move with the price changes. If price fell and share did not, price was never the problem, and the repricer is spending margin on the wrong diagnosis. Check how many lines sit at their floor: a handful is competition, most of the catalogue is a mistake.
No. A product nobody else offers has nothing to react to, and a repricer on it can only lose you margin. Turn it off for exclusive lines and run it where there is genuine competition.
On a small catalogue with few competitors, yes, and it is cheaper. It stops being possible somewhere around the point where you cannot check every competitive line daily, which arrives sooner than most sellers expect.
Through the account settings, after the account is genuinely finished with: stock removed, funds disbursed, the returns window closed and your reports downloaded. Closing cannot be undone in a hurry, so most people asking this actually want holiday settings or a plan downgrade.
Holiday settings make listings unavailable while the account, its history and its verification stay intact. Closing ends the selling relationship and takes the listings and history with it. One is a break, the other is a decision.
No, and this surprises people. Holiday settings take merchant-fulfilled listings inactive; anything already in Amazon's warehouses keeps selling, because Amazon is doing the fulfilling.
Downgrade from the professional plan to the individual one. The subscription stops, the account stays, and you pay per item instead and lose some tools. If cost is the reason, this is usually the right answer.
Nothing good, unless you remove it first. A removal order takes days and cannot be raised on a closed account, so stock left behind at closure becomes much harder to recover.
Disbursements run on a cycle with a reserve behind them, so closing while funds are held means chasing a balance on an account you can no longer use normally. Let the money settle first.
Sometimes, and it is not a formality. What is not allowed is opening a second account to sidestep the first, and Amazon links accounts on details most people would not think to change.
No. A closure is your decision. A deactivation is Amazon's, for a performance or policy reason, and it has its own appeal route. Closing an account to escape a problem does not resolve the problem.
Look at what it costs and earns per product first. Accounts that lose money usually lose it on a handful of lines, in a category whose click costs the margin cannot carry, or on stock decisions made a year earlier.
Payment and settlement reports, tax documents, order history and anything your accountant will ask for later. All of it is easier to have than to request from a closed account.
Only if it breaks the community guidelines: profanity, personal information, promotional content, an obvious conflict of interest, or a review that is really about delivery rather than the product. A negative opinion about the product is not removable by anyone.
A product review sits on the detail page and belongs to the product, shared by every seller offering it. Seller feedback sits on your profile and is about your dispatch and service. The removal grounds are different, and feedback has more of them.
More often than a review. Feedback about an FBA fulfilment failure, a product review left in the feedback field, obscenity or personal data can all be removed. It is worth checking monthly, because a removable one left in place drags a metric for nothing.
No. Anything offered in exchange for changing or removing a review is against policy, including a conditional refund. You can refund unconditionally and say so, and many customers update the review themselves, which is allowed because it is not a trade.
Public replies were withdrawn for most sellers, so the practical reply is private: contact the buyer about their order, solve the problem, and let the next set of reviews say something different.
The request a review button on each order, or an approved tool that presses it for you. What is not allowed is your own wording asking for a positive review, or an insert that offers anything in return.
Inserts that offer anything in exchange, or that steer unhappy customers to contact you instead of reviewing, are against policy and are detectable. A plain insert with care instructions and no request is fine.
It depends entirely on volume. A 3.9 with sixty reviews moves with the next twenty; the same average on six hundred barely moves at all. That is the argument for acting early rather than cleverly.
Only if it repeats. One is noise. Four saying the same thing is a product or listing problem, and no amount of new reviews will outrun a fault the reviews keep naming.
Not the ones selling reviews, which are against policy, detectable, and cost the account rather than the review. Tools that automate the compliant request are legitimate and cheap.
Four routes: wholesale from authorised distributors, private label from a manufacturer under your own brand, arbitrage from retail discounts, or manufacturing your own design. They are different businesses with different capital needs and different failure modes.
Arbitrage teaches the mechanics cheaply and does not scale. Wholesale is the most straightforward real business if you have the capital and can get distributor accounts. Private label pays best and asks for the most upfront.
Trade shows, brand websites and their distributor lists, and industry directories. Verify the distributor is authorised by the brand, because a supply chain you cannot document is worth less than whatever discount made it attractive.
Thinner than private label, and fought over with other resellers on the same listing. That is the trade for proven demand and nothing to invent, and it makes the buy box and your cost base the whole of the game.
Not to sell, but without one you cannot enrol in Brand Registry, which means no control of your listing content, no A+ content and no Vine. It is what makes a private label product defensible rather than just unbranded.
As a way to learn and to make some money, yes. As a business, it does not scale, the sourcing never stops, and gating keeps it out of many of the categories worth being in.
For a paid sample, their real minimum order, who else they supply into your marketplace, and the compliance documents your category needs. The documents are cheap to get alongside an order and slow to get afterwards.
Order shallow enough to be wrong about, inspect before it ships if you can, and treat the first batch as buying information as much as stock. The first reorder carries the real decision.
Yes, and from the supplier rather than a retailer. Authenticity checks are routine, they come with a short deadline, and an invoice you cannot produce turns a routine request into a removed listing.
Usually, and it is one of the better reasons to add a channel: the sourcing work is already done, so a second marketplace is mostly listing and fulfilment rather than a new supply chain.
A set of reports available to brands enrolled in Brand Registry showing what the market did rather than what you did: search term share, what shoppers bought instead of you and alongside you, and repeat purchase behaviour.
Brands enrolled in Brand Registry, through Seller Central. It costs nothing extra, and in most accounts nobody has opened it, which makes it one of the cheapest advantages available.
For a given search term, the top clicked products and their share of clicks and purchases. It is the only view Amazon gives of how shoppers chose between you and the competition on a specific term.
The problem is after the click: the page, the price, the reviews or the delivery promise. It is not a ranking problem, and advertising more only buys more of the same clicks that were not converting.
The funnel for each query from your side: impressions, clicks, add to baskets and purchases with your share of each. It says exactly where the funnel breaks, which decides whether to change the image, the page or the price.
The report showing which products customers bought alongside yours. It is the practical list for bundle ideas and for sponsored placements on complementary pages, because those are purchases that already happen together.
It is the only view of loyalty a marketplace seller gets, and it decides what a first order is worth. Paying above margin to acquire a customer is a mistake on a one-off product and an investment on something people rebuy regularly.
Monthly, with three questions: which terms give clicks and no purchases, which competitor is taking purchase share on our best term, and has repeat purchase moved. Everything else is interesting rather than actionable.
No, it is periodic and it lags, so treat it as a monthly review rather than a dashboard. That is a good fit for what it is used for, which is deciding what to change rather than watching what is happening.
It shows share of clicks and purchases on a search term, not another seller's revenue. That is enough to know who is winning a term and by how much, which is the decision it exists to inform.
Yes. Reviews and ratings pool to the family, so a new child added to an existing family shows the family's review count rather than starting at zero. That is the main reason sellers build variations, and it is also what makes splitting one later painful.
Most likely because it is not the child Amazon shows by default. The featured child receives the traffic, which makes it sell better, which keeps it featured. Children that are genuinely different products rather than versions of one product tend to disappear inside a family for exactly this reason.
No. Children have to be genuine variants of the same product under a theme the category defines, such as size, colour or count. Grouping unrelated products to pool reviews is a listing violation, and enforcement can arrive months later by splitting the family without warning.
Yes, but not cleanly. Reviews follow the children that earned them, so a product that has been showing a family's review count can end up with very few of its own. It is a decision better made deliberately when the family is built than discovered when it is undone.
Yes. Brand Registry is the eligibility requirement and there is no way around it, which in turn means a registered trademark. It is the same gate that unlocks A+ content and Sponsored Brands, so it tends to be the change that matters rather than any one feature behind it.
Not directly. Store pages do not compete in search results and having one does not lift a listing's position. What a Store can do is hold a shopper who has already found one product and show them the rest of the range, which affects what they buy rather than whether they find it.
Mostly through the brand byline on a product page, which is the highest intent route and costs nothing. Beyond that, Sponsored Brands campaigns can send a click to a Store page, and a Store page has its own URL that can be used as a landing page for traffic from outside Amazon. Organic discovery inside Amazon is close to nothing.
Quarterly is a reasonable rhythm for most catalogues. The failure mode is not a Store that is slightly out of date, it is one built once and never opened again: discontinued products still listed, a seasonal image half a year out of season, and categories that no longer match what the brand sells.
No. Vendor Central is invitation only and the invitation comes from Amazon's retail team. Seller Central is open to anybody who can pass the account verification, which is why almost every brand starts there whether or not it is where they intend to stay.
No, and this is the difference that catches most brands out. A vendor agrees a cost price with Amazon and Amazon sets the retail price, including pricing below what the brand charges elsewhere. A seller on Seller Central keeps that decision.
Usually not on margin, and the honest comparison is harder than it looks. Wholesale margin is lower than retail margin after fees, but vendor cash flow is larger and more predictable, and against that sit chargebacks, allowances and price protection, which are deductions that do not appear in a simple margin calculation.
Yes, and some do. It is the most operationally complex of the three arrangements because one catalogue is being run under two sets of rules with separate reporting, and the two can end up competing for the same buy box on the same product.
Sponsored Products advertises to people who are searching, so the intent arrives before the advertisement does. DSP buys display and video placements aimed at audiences, on and off Amazon, and almost nobody it reaches is searching at that moment. That changes what the advertising is for and how long it takes to judge.
Effectively yes, and it is materially larger than Sponsored Products. DSP has historically been bought through a managed service with a spend commitment, and self service access is not universally available. The practical consequence is that it cannot be tested cheaply the way search advertising can.
Usually because of attribution rather than performance. Somebody sees a display advertisement, does not click, and buys days later after searching by name, so the search campaign takes the credit. DSP is judged on view through attribution over a longer window, and the honest measure is what happened to total sales and branded search rather than the campaign's own reported return.
When the listings do not convert search traffic, because display traffic arrives with less intent and will convert worse. When there are still profitable search terms held back by budget, because that money returns faster. And when the catalogue has no repeat purchase or range, because DSP's best economics come from audiences built on what people bought before.
Two automated campaign types Amazon Ads made available in the DSP globally in March 2025. Performance+ is for immediate conversions, measured on cost per acquisition and return on ad spend; Brand+ is for long-term awareness and engagement. The advertiser sets the goal, budget and guardrails, and Amazon's models choose audiences and bids.
There is no useful universal figure. It varies by category by a factor of several, and by traffic source within a single listing: a page fed by branded search converts far better than one fed by broad category browsing. The comparison worth making is the listing against itself, this month against last.
In the business reports, as unit session percentage: units ordered divided by sessions. It is per child ASIN rather than per variation family, which matters, because reading it at family level adds up children that behave differently.
The three usual causes are all outside the listing. A competitor moved on price, so your price became relatively worse. Advertising broadened and started bringing sessions that were never going to buy. Or the listing went out of stock for a few days, and the sessions with nothing to buy are still in the monthly average.
The main image, by a distance, because it is the only element that has to work at thumbnail size against competitors and most of the decision happens there. After that, price relative to what is on the same screen, then reviews, which behave as a threshold rather than a slope.
A reseller competes with other sellers on a listing somebody else created, for demand that already exists. A private label seller creates the listing, is the only seller on it, and has to establish that demand exists for their version at their price. One removes competition on the page and hands over every job the page needs doing.
Most first budgets underestimate by roughly half, because they are built around the inventory. Beyond the first order there is freight, duty, photography, a trademark application, barcodes and packaging, and then the launch advertising, which is the cost of being visible at all for the first months rather than a marketing extra.
Months rather than weeks, and then months more before the numbers mean anything. Sampling involves iteration and each round is a shipping cycle, manufacturing and freight add weeks each, and the trademark that Brand Registry needs takes months and runs in parallel.
Not to sell, but effectively yes to compete. Registry needs a registered trademark and unlocks A+ content, Sponsored Brands and the brand reporting. Launching without it is possible and means launching with fewer tools than the listings you are competing against.
Because delivery speed is decided by distance, and the network places inventory towards where it expects that product to sell. The split is a forecast made visible, using history for that product and category, and it is done as stock enters because moving it later costs Amazon money.
It depends on the shipment rather than being a general answer. It buys one destination and the freight and handling savings that come with it, but stock still gets distributed afterwards at Amazon's pace, and the delivery promise during that window is what one location supports. For a large shipment ahead of a peak it often pays; for a small slow moving product it often does not.
Not directly. What a seller controls is what goes in, when it arrives, and how accurately it is described. Sending earlier is the largest lever, because distribution takes time and stock arriving the week demand starts will spend that week in the wrong places.
Yes, and it is a common blind spot. Total inventory can look comfortable while a product is effectively unavailable for a third of the country, which lengthens the delivery promise there and costs conversion without anything on the listing changing.
Shipping your own orders at Prime speeds with tracking that scans, a cancellation rate close to zero, approved carriers and weekend shipping, proven through a trial period. It is a performance bar rather than a setting, and losing the badge later is worse than never applying, because the listing has been converting at Prime rates and stops.
For bulky or heavy items and for slow moving stock, usually yes, and the gap is not close. For small light products with steady velocity, usually no: FBA's per unit fee is competitive, the storage is trivial and returns handling is included rather than negotiated.
Yes, and it is what most brands of any size end up doing: fast moving small lines in FBA, bulky and slow ones in a third party warehouse, with that warehouse also serving every other sales channel from one pool of stock.
That it can only serve one channel. Units in an Amazon warehouse cannot fill an eBay, Walmart or Shopify order without being removed first, so a brand selling in several places carries duplicate safety stock. It never appears in any single channel's numbers.
The same marketplace with a different buyer: companies, schools and institutions purchasing on account rather than individuals. A seller does not build a second catalogue. The same listings become visible to business buyers, with quantity pricing and tax exempt purchasing available on top.
Practically, yes. Tiered pricing is the mechanism that makes a business buyer choose you over a distributor, and a listing without tiers competes on its consumer price against sellers who have set them. It gives away unit margin by design, which works when the order size is larger.
Consumables first, because the defining behaviour is reordering and repeat purchase is what makes thinner margins work. After that, tools, parts, safety equipment and anything bought against a specification. Gifts and impulse products have no procurement case and return very little.
Per unit, yes, and the question is whether order size covers it. Contribution per order is the figure that matters, because fulfilment cost does not scale with the order the way revenue does. That only holds if the product ships as a case pack rather than as twenty separate units each paying a fee.
Split the spend three ways: terms that produced orders, terms that produced clicks and no orders, and terms that produced neither. The second group is the money. Work down it by spend and decide for each whether the term is wrong for the product or the product is wrong for the term.
Because it averages campaigns that are working with campaigns that are not, so it moves when the mix moves and can improve while the account gets worse. The figures worth reading are per campaign and per search term.
Often duplicate targeting, where the same term is bought in several campaigns and the account bids against itself. It looks like competitive pressure and the competitor is you.
One at a time, with a fortnight between, on any account with normal volume. Forty changes in an afternoon produces an account nobody can reason about a month later, and bid changes in particular take long enough to show that a seasonal movement gets credited to the wrong cause.
Yes, through Multi-Channel Fulfilment. The order is sent to Amazon, picked from the same inventory that serves the marketplace and shipped to your customer, so one pool of stock serves every storefront instead of each channel carrying its own.
Because FBA's fee is subsidised by the referral fee the marketplace takes on the sale, and there is no referral fee on a Shopify order. The comparison worth making is MCF against a third party warehouse rather than against your FBA rate.
It can ship unbranded, which is the usual choice, but there is no inserting a card or controlling the packaging. For a commodity product that is fine. For a brand whose direct channel exists because the unboxing is part of the product, it is the wrong tool.
Channels compete for it. A good week on TikTok Shop can empty the stock Amazon was ranking on, and losing availability costs ranking that takes far longer to rebuild than the stock takes to replace. Forecast the combined demand rather than the marketplace demand.
In Seller Central only. Go to Payments, then Transaction View, filter the date range and open an order's total to see the fulfilment fee Amazon charged for it; the current per-unit rates are on the MCF fee schedule. Shopify records the order and the sale, never the fee, so a margin by channel is reconciled from the Payments report or through an accounting connector that reads it. Sellers on the Shopify forums asking how to fetch MCF fees into Shopify were asking for something Shopify does not hold.
Not first. Price moves conversion, conversion moves rank, and rank moves the volume the unit economics rested on, so recovering the position costs more in advertising than the price rise protected. It is the right move once the category has moved, or when the alternative is selling below contribution.
Advertising efficiency, which is usually where the slack is. Fulfilment and storage costs, especially size bands and how long stock sits. The return rate, which is real money recorded as a percentage. And the mix, by shifting attention to the lines least exposed.
Contribution per unit after every fee on the new landed cost, not gross margin, and the volume at which that contribution stops covering the working capital. Some products do not survive a duty increase and the right answer is to stop buying them rather than price them into irrelevance.
No, and that is why moving first is risky. A competitor sourcing from elsewhere absorbs nothing, so raising your price while theirs holds hands them the comparison on the same screen. Watching what the category does is worth more than reacting on the day the rate changed.
The arrangement where TikTok holds your stock and picks, packs, ships and handles returns, in exchange for a fee. The reason it matters is not the logistics: the platform rewards it with a faster promise and a badge, and on a channel that converts on impulse that removes the last hesitation.
On a low priced impulse product usually yes, because the transaction turns on removing friction and a lost sale costs more than the fee. On a considered higher priced purchase often not, because the buyer will wait and the fee is a larger absolute number. Run both on comparable products and compare conversion rather than cost.
Enough that it is affordable if the channel does nothing, because social commerce demand is spiky in a way marketplace demand is not. Under-committing means the one video that lands sells through in hours and goes out of stock exactly when the algorithm was ready to keep showing it.
Yes. Dispatch deadlines and cancellation rates apply either way, and the consequences escalate from reduced visibility to restrictions. Platform fulfilment does not lower the bar, it meets it for you, which is an operational argument rather than a commercial one.
The predictable prohibitions apply, and the list extends further than most sellers expect: anything making a health claim, anything with a battery or pressurised or flammable, and anything where you do not own the brand and hold no authorisation. The lists also differ by market.
Usually the words rather than the product. Automated review reads the title, description and images together, and a phrase implying a restricted claim stops a listing whose product is fine. The other two causes are the wrong category, which inherits that category's rules, and something restricted appearing in a lifestyle image.
Read the platform policy for that specific market, look for comparable products already selling, and then list a single unit before committing to a purchase order. The review process is the only definitive answer and it costs nothing to ask it early.
Accumulating them is not free, and the pattern draws attention rather than any single rejection. Change one thing at a time and resubmit, so that a pass tells you which change mattered.
TikTok's automated campaign type: you supply a budget and an objective, and the platform decides placements, audiences and bids. It is the same trade every advertising platform eventually offers, control given up in exchange for optimisation against live behaviour.
Almost always too little conversion signal. Automated campaigns optimise towards outcomes they can observe, and a product with a handful of sales gives them nothing to observe. They spend the budget anyway, because that is what they were asked to do.
The budget, which is the only hard limit. The objective it optimises against. The creative, which is the biggest remaining lever because automation decides who sees a video and cannot make it better. And the catalogue: which products, at what price, with what stock behind them.
Longer than the learning period, which is genuinely unrepresentative, and then on account level total sales and blended advertising cost of sales rather than the campaign's own reported return. Watch organic at the same time: advertising growing while organic falls by a similar amount is a transfer, not a gain.
Because it removes the two things that stop an online purchase: not being sure what the product is really like, and not being able to ask. A host holding the product and answering a comment is doing what a shop assistant does.
Three, realistically: a host, somebody reading and answering comments, and somebody watching stock and orders. One person doing all three produces a stream where questions go unanswered, which is the one thing the format was supposed to fix.
An audience that already follows and has been told, paid promotion of the stream itself, and the platform's own distribution, which surfaces streams that are already performing. That last one is why the first fifteen minutes matter disproportionately and why promotion at the start is better value than the same money spent later.
Usually not yet. Live rewards an audience and ordinary content is what builds one, so a brand with no content presence is buying viewers for every session indefinitely. Content first, then live as the thing that converts the audience it produced.
Ones that answer an objection. A demonstration answers whether the product is any good, a before and after answers whether it will work for the viewer, and a straight answer to a customer question converts the people closest to buying. The format follows from the objection rather than from a content calendar.
Lighting, framing and audio matter. Production values that signal a budget work against you: a video that looks like an advertisement is processed as one and scrolled past. Somebody holding a product in a kitchen repeatedly outperforms a studio set.
Often enough to have a comparison. Twice a week is learning at a rate that makes every conclusion a guess, and the platform rewards volume regardless. Sustainable production rate is the real constraint on most brands.
Not views. Watch through rate tells you the opening worked; attributed sales tell you the video worked. A video can pass the first and fail the second, which is the most informative result there is: a strong hook pointed at the wrong objection.
They do different jobs. In house is unbeatable for speed and for answering customer questions, because nobody has to be briefed. Creators buy credibility and range, which is why their videos convert better, and a brand at any real volume ends up using both.
The operation rather than the artistry: a schedule that does not depend on one person's week, a creator roster already sourced and contracted, and the discipline of testing formats. An agency producing polished brand films is producing the thing the audience scrolls past.
What gets handed over when it ends: raw footage or only finished files, who holds usage rights, and what happens to the creator relationships. And who writes the briefs, because an agency that briefs alone drifts away from what the product does, since they do not answer the customer emails.
Output has gone irregular, which means it depends on spare capacity. Or the last twenty videos share a format, which means you have hit somebody's range limit. Or cost per sale has started mattering more than cost per video, which is when spending more per video becomes rational.
One issue per case, identifiers first, the action you want stated explicitly, and the evidence attached rather than described. It is read by somebody working from a script under a handling time who has your message and nothing else, so write for that.
Reopening sends it back to the same queue with the same script, which is why the second reply repeats the first. Where the original was badly framed, a new case with one issue and the right evidence usually outperforms arguing with a reply to a bad one.
When you can name what was missed: the policy that permits what was refused, the evidence already supplied and not addressed, or an inconsistency with a decision on an identical case. Dissatisfaction with the answer is not a reason for review.
A separate log with case ID, subject, the ASIN or shipment, the money or risk attached, the date opened and the date to chase. The chase date is what stops cases expiring silently, and the money column is what decides which ones get escalation effort.
You create the listing, send stock into Amazon's network on a shipment plan it generates, and Amazon stores it. When an order arrives it picks, packs and ships the unit, handles the customer contact and processes any return. You are charged a per unit fulfilment fee set by size and weight band, plus monthly storage on the space your stock occupies.
Amazon said US FBA fees would rise by an average of about eight cents per unit from 15 January 2026, and that no new fee types were being introduced. It is a small change relative to previous years. The numbers that decide profitability for a given product are still its size band and how long it sits in storage.
Yes, through Multi-Channel Fulfillment, which fulfils orders from non-Amazon channels out of the same stock pool. Amazon ships those orders in unbranded packaging by default, with standard delivery inside three business days and expedited inside two, so the trade is branding and speed against holding a second pool of stock somewhere else.
No. Seller Fulfilled Prime shows the Prime badge on products you ship yourself, after a thirty day trial you have to pass and with continuing requirements on late shipment, valid tracking and cancellation rates. Performance is assessed separately by product size tier, so qualifying for standard size items does not depend on how you handle oversize ones.
Often not, because the fulfilment fee is a fixed amount per unit and a low price leaves little room for it. The test is the fee as a share of the price after the referral fee, not the fee on its own. Where the numbers are marginal, packaging that moves the unit into a smaller size band changes the answer more reliably than a price increase does.
It becomes unsellable inventory, and it keeps costing storage until you deal with it through a removal order, disposal or liquidation. Each of those has its own cost and its own timescale, and none of them is automatic, so the stock sits and is charged for until somebody acts.
The account itself is $0.99 per item sold on the Individual plan or $39.99 a month on the Professional plan, and referral fees are taken from each sale. The real starting cost is stock: a reseller can begin with a few hundred dollars of inventory, while a private label product with its own packaging and a first production run is usually several thousand before the first unit arrives at Amazon.
Amazon's own list is a government issued ID, an email address, an internationally chargeable credit card, a bank account and routing number, a business licence or registration where you have one, proof of residential address from the last 180 days, and your tax information. Identity is verified from a photograph of you with your ID or on a video call, and Amazon says that usually takes three business days or less.
The Individual plan suits somebody selling a few items to see how it goes. A business registers on the Professional plan: it is cheaper past about forty units a month, and advertising, Amazon Business, bulk listing tools and the programmes worth having all require it.
Only within Amazon's dropshipping policy, which requires you to be the seller of record on every document the customer sees and forbids shipping from another retailer. In practice that rules out the arrangement most people mean by dropshipping, and the accounts that try it tend to be suspended on the paperwork rather than on the sales.
Not to register: an individual can open an account with their own identity and tax details. A registered business is asked for where it exists, and it becomes worth having quickly, because distributors want a resale certificate, Brand Registry wants a trademark owner, and liability sits somewhere either way.
With advertising running and a price near the floor, usually within the first fortnight. Without advertising a new listing can sit for months, because it has no rank and nothing to earn one with. The first sixty days are a sequence of paid sales, reviews and rank, and each depends on the one before it.
If you are a registered business, eBay asks you to select the business account at registration and to give the legal business name, the address, identification for the people who control the business, an EIN, and a business checking account in the same name for payouts. An individual account is for a person selling under their own name and Social Security Number.
In most categories, 13.6% of the total sale amount up to $7,500 per item plus $0.30 per order under $10 or $0.40 per order over $10, with 250 free listings a month and $0.35 per listing after that. A Store subscription from $4.95 a month raises the free listings, Promoted Listings add the ad rate you choose only when an item sells, and Top Rated Plus takes 10% off the final value fee. These are eBay's published US fees in September 2026.
eBay reviews every account monthly and raises the limit automatically on sales volume and buyer feedback, so the first increase usually follows the first full month of sales. You can request an increase sooner from the message eBay sends as the limit approaches, and eBay may ask for business or contact details before granting it.
An account at least 90 days old with 100 or more transactions and $1,000 or more in sales to US buyers in the past 12 months, a defect rate of 0.5% or less involving no more than three buyers, no more than two cases closed without seller resolution, no more than five late shipments, and tracking uploaded on time on at least 95% of transactions. Adding one day handling and 30 day free returns to a listing makes it Top Rated Plus, with a 10% final value fee discount.
It is a listings calculation. Without a Store you pay $0.35 for every listing beyond 250 in a month; a Basic Store at $21.95 a month includes 1,000. The subscription is cheaper than the insertion fees from about the 320th active listing, and each tier above adds more free listings and lower fees beyond them, so the right tier is the one whose allowance covers the catalogue.
It does the technical part: a sitemap at /sitemap.xml that updates itself, canonical tags on every page, a robots file that keeps crawlers out of the cart, checkout, search and filtered collection URLs, and title and description fields on every product, collection, page and post. It does not write collection content, product copy or the structured data the theme leaves out, and those decide whether the store ranks.
No. The path prefixes are fixed by Shopify: collections at /collections/, products at /products/, pages at /pages/ and posts at /blogs/. The handle after the prefix is yours to set, and changing it creates a redirect from the old address if you tick the option when you save.
At the root of your domain, at yourstore.com/sitemap.xml. Shopify generates it and updates it whenever a product, collection, page, image or post is added. Submit it once in Google Search Console under Indexing, and if the store uses international domains, submit each domain's sitemap separately.
Because on most themes a collection is a product grid under a heading with no text to rank on, the same product sits in several collections with several paths, and the filter and sort controls generate URLs the robots file has to block. A collection ranks once it carries content of its own, one collection answers each category term, and the filtered versions are kept out of the index.
Some do one job well, such as image compression or review markup. None of them can do the work that decides ranking, which is writing collections and product pages that are unique to the store, and each one adds a script the theme loads on every page. Install one for a specific job, remove its theme code when it goes, and never install one to build a sitemap you already have.
On its own retail offers Amazon prices dynamically, moving prices with demand, stock and the prices it sees from competitors on and off Amazon. For sellers it enforces a Marketplace Fair Pricing Policy: an offer priced significantly higher than recent prices on or off Amazon can lose the Featured Offer, the listing, or in repeated cases the account, and the Pricing Health page in Seller Central shows which offers are affected.
Yes, on its own offers, and it gives sellers the same capability. Automate Pricing, free with the Professional selling plan, changes a seller's price under a rule the seller chooses: against the Featured Offer, against the lowest price in the store, against the lowest external price, or on sales velocity. It needs a minimum price and Amazon advises a maximum too.
No. Matching the lowest price by default races the listing to the floor and hands the margin to nobody. The Featured Offer is decided on price together with fulfilment, delivery speed, seller metrics and stock, so a seller with better metrics can share it at the market price, and a brand that owns its listing and its distribution can hold above it. Leading on price is for launches and for a real cost advantage.
Since 2 June 2025, as announced in Seller Central, a coupon costs an up front $5 per coupon plus 2.5% of the sales redeemed through it, on top of the discount itself. Before that the fee was charged per unit sold. The referral and fulfilment fees on each order are unchanged by the coupon.
A discount of at least 10% off the non member, non promotional price, on a product in new condition with a rating of three stars or more or no rating yet. The everyday version is offered on Prime and non Prime offers; the event versions, for Prime Day and the holiday period, carry a flat campaign fee that Amazon publishes ahead of the event.
Amazon's announced cutoffs are 14 October for AWD, 21 October for FBA shipments with minimal shipment splits, and 28 October for FBA shipments with Amazon optimised splits. A standard delivery has to be scheduled at least 7 days before the cutoff and a partnered carrier pickup at least 14 days before.
Amazon has said the event falls on the same timing as 2025, in early October. Deal submissions closed on 8 September 2026, and the inbound cutoffs were 2, 9 and 16 September by shipment method.
The same as Prime Day 2026: an upfront fee of $100 per promotion and a variable fee of 1.5% of promotional sales, capped at $5,000. Submitting before the early deadline saved $50 on the upfront fee.
A 3.5% fuel and logistics related surcharge on FBA fulfilment, an average of $0.32 per unit, from 15 October 2026 to 14 January 2027. It applies to every FBA unit shipped in the window, deal or no deal.
No. Amazon excludes Prime Big Deal Days promotional prices from the 30 day and 60 day lookback windows, so an October deal price does not become the reference price a November deal is judged against.
Capacity is set monthly from the Inventory Performance Index, Amazon's forecast for your products and the space Amazon has, and Amazon has said limits may be lower while its centres focus on receiving holiday stock in September and October. Sellers report limits well below last year's fourth quarter volume this year.
You can request more by naming a reservation fee per cubic foot, refunded as performance credits against sales from that space. Sellers report the amount that can be requested is capped, so treat it as one lever among five rather than the answer.
Amazon states that AWD auto replenished inventory is not subject to FBA capacity limits during peak. The AWD inbound cutoff for Black Friday is 14 October 2026, a week ahead of the first FBA cutoff.
Amazon announces the threshold each quarter and applies limits to sellers below it. The score is built from excess inventory, sell through, stranded inventory and in stock rate; clearing stranded stock and pricing slow units to sell are the two you can move inside a month.
Amazon says inventory arriving after the cutoff dates is not guaranteed to be processed in time for the event, and that receiving is no longer the priority in November and December. It is received, but not necessarily before the sales it was shipped for.
In 2026 the fees are unchanged from Prime Day: $100 upfront per promotion and 1.5% of promotional sales, capped at $5,000. Submitting before 5 August saved $50 on the upfront fee. The peak fulfilment surcharge of 3.5% applies separately to every FBA unit from 15 October.
The Black Friday week and Cyber Monday window runs from 8 July to 20 October 2026, with an early submission discount deadline of 5 September. Prime Big Deal Days submissions closed on 8 September.
Amazon excludes Prime Big Deal Days promotional prices from the 30 day and 60 day lookback windows, so the October deal price does not become the reference a November deal is judged against. A plain discount run alongside should be treated as inside the lookback.
No. A product belongs in a deal when it still holds margin after the deal, the fees and the surcharge, when it has the stock to last the event, and when a deal can move it onto page one and keep it there. A product that fails those tests sells better at its everyday price on a busier day.
Yes, for the week rather than the month: bids in position before the event, budget released on the day so campaigns do not run out early, and bids brought back down the day after. A deal without advertising depends on the deal page, which in November is the busiest page on Amazon.
As reported from Walmart's holiday guidance, 1 September for the October deals event and 15 September for the annual holiday event, with the WFS peak period running from 1 October to 31 December. Check Seller Center for the dates as published to your account.
By nominating it through Seller Center's promotions tools against a deal price and a stock commitment. There is no per deal fee as on Amazon; Walmart's category teams choose from the nominations, and the price cut and the stock behind it decide whether a product is chosen.
Walmart's fulfilment network is smaller, with fewer centres and receiving doors, so it cannot absorb a late surge of inbound the way Amazon's can. The inbound dates are set so that receiving is finished before the centres turn to orders in October.
Set them in September so the campaigns already hold position, then release budget on the event days rather than raising bids into the November auction, which is when most of the year's new bidders arrive. Bring them back down in the week after Cyber Monday.
The last WFS shipping date that delivers before 25 December, which Walmart publishes each year in Seller Center. Orders after it ship with a January delivery promise, and the selling season for gifts ends on that date, not on the 25th.
Black Friday falls on 27 November 2026 and Cyber Monday on 30 November. TikTok Shop runs platform campaigns around both, with an early holiday event in October and a December gifting window. Registration windows, discount depths and stock commitments are on the campaign pages in Seller Center and change each year.
A shop above the health thresholds for shipping, cancellation and returns, products at or below the campaign's required discount, and the minimum stock commitment the campaign names. Check the campaign page in Seller Center for the current figures rather than a summary from elsewhere.
September. A creator needs the sample in October to post in November, and the videos that sell in the week of Black Friday were posted before it. Existing sellers of the product get the holiday sample and deal price first; new creators in the category get a higher commission for the quarter.
The orders placed after the stock is gone become cancellations, and cancellations are part of the shop health score that decides campaign eligibility. Stock the product to what its videos can sell, and pause the affiliate offer before the stock runs out rather than after.
Yes, as a scheduled slot with a host who knows the product, the deal price on screen and flash deals set up inside the live. Lives that hold viewers with something happening convert; a live that reads the product page aloud does not, whatever the day.
Top Rated status expects two working days or less, with tracking uploaded and the carrier's first scan inside the handling time. Holding it in the week after Cyber Monday means daily carrier pickups and labels printed the day the order arrives.
Yes, under its published seller protections: a late delivery caused by the carrier is removed from the seller's metrics where the package was scanned on time, delivery estimates adjust for known delays, and related defects and feedback are eligible for removal. The condition is the on time first scan.
As reported from the USPS notice, USPS peak pricing runs from 4 October 2026 to 17 January 2027, with increases by service and weight. The private carriers set their own surcharges and windows. Read the table for the services you use and decide in October whether the price moves.
The carriers publish last shipping dates by service each year, and eBay's delivery estimates stop promising 25 December once a service's date has passed. Treat the earliest of those dates for the services you offer as the end of the gift selling season.
As a January routine: inspect, relist or write off, refund inside the window eBay's money back guarantee sets, and count the returns by product. The product that comes back most is the one whose listing needs a better description or better photographs before next year.
September. Black Friday is 27 November and Cyber Monday 30 November 2026, and the eight weeks before them are for running the campaign, not building the store. Speed, checkout, stock, the campaign, the shipping promise and support templates are all September work.
Shopify reported $14.6 billion across its merchants over the Black Friday to Cyber Monday weekend in 2025, up 27% on 2024. The figure is Shopify's own announcement.
Automatic discounts where the offer allows it. A code is mistyped by a share of everybody who tries it and generates the support tickets that follow; an automatic discount applies on its own and can be tested in September to confirm it stacks, or refuses to, the way it should.
A feature from Shopify's Spring 2026 Editions that plans and runs a marketing campaign across email and the store from a single brief. It schedules the sequence; the merchant still decides the offer, the audience and the dates.
Checkout paths nobody tested (a discount code with an express payment, a failed payment), stock that was not true across channels on the morning of the sale, and a support inbox with nobody answering on the Saturday. All three are fixed in September, none can be fixed on the day.
The better question is how to divide the quarter. Sort the thirteen weeks into build weeks, event weeks, gift weeks and the tail, give each a share by marketplace from last year's sales pattern, set the Amazon event weeks as daily figures, and hold a tenth back until Prime Big Deal Days has shown what converts.
Raise them in October so the campaigns already hold their placements, and change the daily budget on the event day rather than the bids. Bids raised on the morning of the event enter an auction where every other seller did the same, and the budget is spent learning what the build weeks would have taught.
Yes, and it is usually the cheapest traffic of the year. The auction empties after the last shipping date while buyers keep spending gift money on themselves, and the tail is where a January stock position is cleared at a reasonable cost per click.
Against each platform's own return on ad spend norms rather than Amazon's. TikTok Shop's attribution and creator commissions, Walmart's smaller auction and eBay's percentage of sale model each put different things inside the number, so a single target across all four misreads three of them.
Not on Amazon in 2026: Prime Big Deal Days promotional prices are excluded from the 30 and 60 day lookback windows, so the October deal does not become the reference for the November one. A plain discount run beside the deal should be treated as inside the lookback.
No. Amazon judges a deal against the price over the lookback window, not the price on the day, so a price raised for a few weeks does not move the reference. Hold the real everyday price through October instead, with no coupons stacked on it.
Most often because its price was beaten on another marketplace, usually the seller's own Amazon deal. Walmart can suppress a listing priced above the same item elsewhere, so deal prices have to be matched on Walmart on the same days.
It raises the floor: 3.5% on FBA fulfilment from 15 October to 14 January, about $0.32 a unit on average. Rebuild the unit economics with it in before pricing any deal, because a deal priced from the summer's margin is priced from a margin that no longer exists.
Pause the rules that match competitors during the event weeks and keep a per product floor that includes the surcharge and the deal fee. A competitor's Black Friday price is a deal with a fee behind it and an end date, and a repricer that follows it sets a new reference on the way down.
Run the returns desk: inspect every return, refund inside the platform's window, and count returns by product and reason. On Amazon, reconcile the season's reimbursements at the same time, because peak receiving and holiday refunds are where the errors happen and the claim windows are finite.
14 January 2027 for holiday 2026. The fourth quarter's real margin by product can only be read once the surcharge has ended and the returns are in, which makes the second half of January the time for the numbers.
Decide per product from a table: units, months of cover at the January rate of sale, storage cost per month, the price that clears it in six weeks, and what it would carry next season. Seasonal stock is usually held and the storage bill is the argument; everything else takes the January price while the buyer is still spending.
In February, from January's figures by product and marketplace. The purchase orders for next October are placed in the summer, and the summer's decisions come from what was written while the reasons were still fresh.
By product and by marketplace, on five figures: units against forecast, margin after every fee including the surcharge and deal fees, advertising cost per unit in event weeks against build weeks, return rate, and stock left over. The product that held on all five is the one to order more of.
Amazon's notice says switching back as your default will no longer be available, so yes, for accounts that receive it. Shortcuts and saved searches carry over. Amazon has not published a single date; the notice appears per account.
Split between the workspace they belong to and Analytics, under their current names. Open the site map from the navigation menu or the footer, find the report by name once, and bookmark it. The sales dashboard draws by day and hour by default; the period is set on the dashboard.
Catalog became Products and Inventory became Supply Chain. Amazon's tooltips on the new menus show the old name against the new one, and breadcrumbs show the path, as in Products > Manage Products > Manage A+ Content.
Not entirely, but action cards can be hidden for seven days and recommendation cards permanently. Hide every recommendation card and what remains is the list of tasks the account actually requires, with critical badges on top.
The left navigation pane overlays the input fields on some pages. Collapse the pane and it stays collapsed for the session. Sellers reported it in Amazon's navigation feedback thread and it is on Amazon's list.
Open the site map. Amazon added one in early September 2026 showing every tool in the new Seller Central organised by workspace, reached from the navigation menu or the footer of any page. Every sub-navigation page now carries breadcrumbs at the top (Products, then Manage Products, then Manage A+ Content, for example), and hovering a menu item shows where a renamed tool went: Catalog is under Products, Inventory under Supply Chain. Bookmarks for one-click access to the pages you use most were announced as the next addition.
Open the Seller Forums and sort by newest. A real incident produces a thread within the hour with replies from sellers across categories, and Amazon's community managers post in it once it is acknowledged. If that thread exists and matches what you see, the cause is Amazon and the orders will appear when reporting catches up.
The buy box first: open the product page logged out and see whose offer has the buy button and at what price. Then account health, the notifications inbox and the case log for a review, verification request or performance notification. Then the campaign manager for a budget that ran out.
Not until the check is done. During an Amazon reporting delay the orders exist and have not been shown yet; a price cut made against missing data is a real cut against real sales. Decide on price from the buy box and the unit economics, not from the sales graph on the day.
For an advertised product, usually because the daily advertising budget runs out at that hour. The sales graph shows a wall at the time the campaign stopped. Raise the budget, spread it, or accept that sales after that hour are organic only.
Most reporting delays clear within hours and the orders appear with their original timestamps. The forum thread for the incident is the place to watch; Amazon's community managers post when the incident is resolved.
For one of the reasons eBay publishes: a new or infrequent seller, information being confirmed about the account, a buyer dispute, unusual activity, a sudden increase in volume, a high priced item, the thirty days after a restriction, or an Authenticity Guarantee sale awaiting inspection. The payments page names the reason on the held transaction.
For a private seller, two days after delivery confirmation with an eBay label, seven days with another label, or thirty one days after the buyer paid with no tracking. For a business seller, twenty four hours after delivery with an eBay label or fifteen days after payment otherwise.
A seller who has not reached at least ten completed sales with a combined total of $150 or more in the last five years. Below that, an account of any age is treated as new for payment holds.
Ship with an eBay label and upload tracking the moment the item ships. Delivery confirmation from the carrier's scan or a QR code scan releases the hold on the fast schedule. Messaging eBay does not release a hold.
Not on its own. Holds end on delivery confirmation or when the waiting period expires. Feedback is not one of the release conditions on eBay's policy page.
They can offer, and the seller should decline. Seller protection against item not received claims requires tracking uploaded by the seller from an eBay label or an integrated carrier showing delivery to the address on the order. A buyer's own label gives the seller none of that.
Protection applies to delivery to the address on the order. Ask the buyer to cancel and repurchase with the right address rather than shipping to one supplied in a message.
eBay's policy forbids arranging a transaction outside the site and sharing contact details before a sale; enforcement ranges from a warning to suspension, and the transaction has no eBay protection or Money Back Guarantee. eBay asks sellers to report the offer.
It stops change of mind returns. The Money Back Guarantee still covers items not as described, damaged or not received. A not as described case on a no returns listing is decided by eBay on the evidence, and an accurate listing with its own photographs is the defence.
If the item shipped with an eBay label, eBay reimburses the seller automatically, removes any defects and excludes the case from service metrics. That is one of the reasons to buy the label through eBay.
Most often because the sessions are not people. Filter Shopify analytics to human traffic, then read sessions by referrer and location. Bot traffic arrives direct or from unknown referrers, from countries you do not sell to, at odd hours, on a few pages. Judge the store on the human sessions that remain.
A session is tracked with a cookie on the visitor's device and ends after thirty minutes of no activity, and at midnight UTC. One visitor can have many sessions, so sessions run higher than visitors.
It costs bandwidth and distorts the analytics; it does not cost sales. Filter it out in analytics and, if the load matters, block it at the edge rather than with an app, which adds weight to every real visitor's page.
It depends on the traffic source and the product, so the useful comparison is the store's own funnel: sessions that add to cart, reach checkout and purchase, with bots filtered. The shape of that funnel says where the leak is, which matters more than a benchmark.
At six places in order: a landing page that is not the product promised, an offer that does not say what is different, a price that grows at checkout, a product page that does not answer the buyer's questions, shipping cost and time hidden until the cart, and a checkout untested on a phone.
Carriers audit packages in batches and bill the differences after the fact, and Amazon passes the charge through when it arrives. One email covers a period, often a quarter, and each line in the transaction view links to the order it belongs to.
In the payments transaction view: Transaction Type: Shipping Services, then Product Details: Adjustments. Each line links to its order ID and shows what the carrier billed against what the label declared.
Open a case with Seller Support with the order ID, the adjustment line, and photographs of the packed parcel on a scale and against a tape. FedEx Ground Economy claims go to FedEx directly, with the seller's name and address matching the label. Group a pattern into one case rather than fifty.
Carriers bill the greater of the actual weight and a weight assigned to the package's volume. A light item in a large or puffed package is billed for the volume, and a label bought against the item's weight is adjusted to the package's.
No. A Buy Shipping label with tracking uploaded on time protects the order against item not received claims, which is worth more than the adjustments. Measure the packed parcel, store the real dimensions on the SKU, round up to the inch, and dispute patterns with photographs.
Because the carrier audits the parcel after it ships and Amazon passes the charge through when the carrier assesses it, which can land in a later settlement than the order. In September 2026 sellers on the Seller Forums reported USPS correction charges for labels bought between 1 January and 31 March, applied between 7 and 10 September, with the notice naming the same three causes Amazon lists: a weight or dimensions that did not match the label, or a return address that was wrong or invalid. The order is still the evidence: open it in Transaction View, filter Shipping services purchased through Amazon, and search Adjustment to see the carrier's audited figures against the label.
Amazon's moderators have said there is no fixed deadline, but that the information gets harder to reach as time passes. Dispute while the order, the label and a photo of the packed parcel on a scale and beside a tape are still to hand, and send them with the case to Seller Support; a dispute with no photo is a request the carrier can refuse on its own audit.
No. A rate change applies at purchase, to the price you see on the label. An adjustment is the difference between what the label said and what the carrier's audit found: weight, dimensions, service or return address. If every parcel is the same size and the same weight and the charges still arrive, dispute them with one photo set that shows it; several sellers reported dozens in a year on identical mailers, and that pattern is the case to put to Seller Support rather than a reason to stop buying labels.
A pure white background (RGB 255, 255, 255), the product filling at least 85% of the frame on its longest side, no text, logos, graphics, watermarks, borders or props, the actual product as sold, between 500 and 10,000 pixels on the longest side with 1,600 or more recommended for zoom, in JPEG, PNG, GIF, WEBP or TIFF.
Between 500 and 10,000 pixels on the longest side. Amazon recommends 1,600 pixels or more so the product page's zoom works, and a square or near square frame fills the search results thumbnail best.
Retouching such as background removal and colour correction is treated as ordinary editing. Photorealistic AI generated people must be disclosed through the AI generated and AI generated people checkboxes when uploading A+ Content, for new uploads going forward. Every image must still show the product accurately.
The main image failed a check: a background that is not pure white, the product under 85% of the frame, text or a prop, or a file outside the pixel range. The account names the reason on the suppressed listing; upload a compliant main image and the listing returns once the check passes.
No. Text, logos, badges, graphics and watermarks are not allowed on the main image. They are allowed on the other images, which is where a benefit callout or a comparison chart belongs.
Since 20 June 2025, 36 inches long by 25 wide by 25 high and 50 pounds for FBA in the US. A box over 50 pounds must contain a single item that itself exceeds the limit. AWD shipments keep the older 25 by 25 by 25 inch, 50 pound limit.
A 40 by 48 inch, four way access wooden pallet, GMA grade B or better for most products, GMA 1A for grocery, GMA 1A or 1B for health, personal care and beauty. No plastic, cardboard or particle board.
No more than 1,500 pounds in total, and a pack height ideally no higher than 50 inches: 45 inches of boxes on a 5 inch pallet. Pallets should be stackable, stable and flush on all sides.
An FBA box ID label on a flat side of every box, not across a seam, and on a pallet the box labels face outwards with the shipment ID label on all four sides of the stretch wrap. Every box on a pallet belongs to one shipment ID.
Amazon may refuse the delivery at the seller's expense, or accept it and charge for the correction: unplanned prep, an overweight box, a rebuilt pallet. A refused shipment can be returned or disposed of, and disposal is charged and final.
A no cost programme for professional sellers with a consistently high Account Health Rating. When an issue would otherwise deactivate the account, a specialist contacts the seller, explains the issue and how to fix it, and the account is not deactivated as long as the seller can be reached within 72 hours and works with Amazon to resolve it.
A professional selling plan, an Account Health Rating of 250 or higher for at least six months with no more than ten days below 250, and a valid emergency contact number in Seller Central. Enrolment is automatic and confirmed by email.
Usually because a violation dropped the rating below 250, which disqualifies the account. The programme protects a good account from a single mistake; it is not designed to hold an account that has accumulated several. Amazon can also remove a seller for suspected fraud or illegal activity regardless of rating.
No. Fraud, deception, illegal or harmful activity and regulatory failures are excluded, and Amazon can deactivate immediately for those. It also does not cover listing suppressions, payment holds or verification requests, which have their own processes.
That the brand, the category or the specific product is gated and needs an application before it can be listed. The message names which one, and the application is for that gate: a category approval does not open a gated brand, and a brand approval does not open a restricted ASIN.
As the application states it: at least one invoice from a manufacturer or distributor, dated within the last 180 days, showing at least ten units of the brand, category or product, with the seller's name and address matching the account, nothing redacted except pricing, and the issuer's name and contact details on it.
The reasons sellers report most: a retail receipt instead of an invoice, a distributor the brand has not authorised, a name or address that does not match the account to the legal suffix, an invoice older than 180 days or over redacted, or a brand that requires a letter of authorisation it has not given.
Usually because the selling account is not linked as the owner in Brand Registry, or the restriction was applied before registration completed. Confirm the role assignment in Brand Registry; where an application is still demanded, a letter of authorisation from the brand to the seller and the trademark and registry documents are the evidence.
An application with a compliant invoice can be approved in hours to a few days. An application argued through Seller Support, such as a brand owner without a manufacturer's invoice, takes as long as the case does, which is why the gate is checked before stock is bought.
Yes, in most categories, under the condition grades Amazon defines: Used, Like New or Open Box; Very Good; Good; and Acceptable, with the last two only where the category guidelines allow. Beauty, grocery, health and personal care and safety certified products are new only.
Clothing and shoes are governed by their own category guidelines and programmes rather than the general used grades. The condition menu on the item's catalogue page shows what the category permits; if a used grade is not offered there, it is not allowed.
In Amazon's words, an apparently untouched item in perfect condition, with the original packaging intact and pristine, no signs of wear on the item or its packaging, instructions included, and suitable for presenting as a gift. The protective wrapping may be missing.
Not against a new offer. A used offer appears in the other sellers list and in the used and new box where there is no new offer or the used price is far enough below. Its competition is the other used offers, graded and priced.
Amazon's programme for refurbished products, with its own qualification, quality standard and placement. A business selling refurbished electronics, appliances or tools in volume sells under it rather than as used.
Yes, four ways: a supplier ships stock straight into FBA and Amazon fulfils it; dropshipping within Amazon's policy, where the seller is the seller of record on every order; print on demand through Amazon's programmes or a supplier; or a third party warehouse or Multi Channel Fulfilment shipping under the seller's name.
Under conditions: the seller must be the seller of record, and invoices, packing slips and packaging must identify the seller and no third party. Ordering from another retailer and having them ship to the customer with their paperwork in the box is prohibited.
Supplier to FBA, because nobody else sits in the chain between the product's cost and the customer. Dropshipping is the thinnest, since the supplier charges for the fulfilment the seller is not doing.
Programmes such as Merch on Demand and Kindle Direct Publishing, and print on demand supplier integrations, that make the item when it is ordered. The seller supplies the design and the listing and earns a royalty or a fixed spread.
Fix conversion first: the main image, the title and bullets, the price against the page's other offers, and reviews. A listing that converts turns every existing session into more sales without buying a single click, and makes the advertising that follows pay.
Yes, after the listing converts. Advertising is lever seven of twelve. Run first, it buys traffic to a page that loses it and measures the conversion problem at the price of a click each.
From the business report's sessions and unit session percentage by product. Low sessions with a good conversion rate is traffic; high sessions with a poor rate is the listing or the price; both low is the listing first.
Days out of stock reset sales velocity, which organic rank follows, so the listing returns lower on the results page than it left. Stock is a sales lever, and the forecast is how it is pulled.
Meet the prerequisites Amazon states: fulfil outstanding orders, wait 90 days after the last sale for the A-to-z claims period, reach a zero balance, resolve every transaction and refund, confirm your bank details, and resolve every performance notification. Then use Close Account under account settings.
Because one of the prerequisites is not met: a sale inside the last 90 days, a balance that is not zero, an open order or refund, stock still in FBA, or an unresolved performance notification. The 90 days run from the last sale, so listings have to be closed first.
It has to leave first, through a removal order to return it or dispose of it. Storage is charged for every day it stays, and the account cannot close with stock in a fulfilment centre.
Downgrade to the individual plan and set vacation mode unless the business is ending or the account is being replaced. Downgrading stops the monthly fee and keeps the history, feedback and Brand Registry roles; closing gives them up for good.
Not the same one. Opening another account after closing raises Amazon's rules on multiple accounts and needs Amazon's agreement, and the history of the closed account does not carry over. That is the main reason a downgrade is usually the better choice.
Amazon publishes none, and the averages quoted online are other sellers' accounts in other categories. Your own average is in the campaign manager and the search terms report, by term, and that is the only one that predicts what your next click costs.
The cost is set by the next competing bid plus an increment, up to your bid as a ceiling, and placement adjustments multiply the bid for top of search and product pages. A cost near the bid means the auctions are competitive; a cost well below it means they are not.
More bidders on the term, top of search placement, broad match entering stronger auctions, a new ad with no conversion history, the fourth quarter, and the category's margins. Brand terms are cheap because competitors rarely bid them; the category's head term is dear because everyone does.
Lower the bids on terms whose cost per order is too high, and remove with negative keywords the terms that take clicks without orders. A cheaper click on a term that does not convert is still wasted, and a dear click that converts is worth keeping.
No. Cost per order, read as advertising cost of sales, is: it is the click price divided by the conversion rate. A product converting one click in five can afford four times the click price of one converting one in twenty.
As a multiple of the profit a new owner would keep, measured for smaller businesses as seller's discretionary earnings over the trailing twelve months (net profit with the owner's salary and one off costs added back) and for larger ones as EBITDA. Inventory is bought separately at cost.
Brokers report 2.5 to 4 times seller's discretionary earnings for most, with strong brands above and small or risky businesses below. At the 2021 peak the same businesses fetched 4 to 6 times; the aggregators' retreat and more careful buyers compressed the range.
Three or more years of consistent profit, no single product or term carrying the sales, revenue beyond Amazon, a registered brand people search for, a clean account, written supplier agreements with a second supplier, and books that tie to the settlements by month.
No. Inventory is paid for separately at cost on top of the multiple. Too much stock is stock a buyer did not want to pay for, and a stock out during diligence resets rank at the worst time, so the level is planned for the sale.
A year before. The price is paid on the trailing twelve months, so that year is the one being sold: books in order, advertising run for profit, the account clean, stock sized for the process, and any launch done early or not at all.
Units Amazon holds that cannot be sold at that moment, for one of three reasons the FBA inventory report names: reserved for customer orders that have not shipped, reserved for transfer between fulfilment centres, or reserved for processing such as counting or measuring on receipt.
Customer orders clear in a day or two per unit. Transfers take days to two or three weeks, in waves after a shipment lands. Processing takes hours to days. Units reserved in transfer or processing for weeks at a steady figure are stuck and worth a case.
The listing sells from available units only. Units in transfer or processing are not available, so a product can be out of stock on the page with stock in the network. Time inbound shipments on the available count rather than the total.
Not for units that are reserved, but for units that never came back: stock that entered processing or a transfer and was never returned to available is lost in the warehouse or in transfer and reimbursable after the reconciliation window, with the inventory ledger as the evidence.
Yes. Grocery and gourmet food is a gated category, applied for through the listing workflow with manufacturer or distributor invoices from the last 180 days showing ten or more units, or a brand's own documents. Some sub categories and products carry their own approval on top.
More than 90 days remaining when the product arrives at the fulfilment centre, for lot controlled food and drink and anything shipped with them. Units within 50 days of expiry on arrival are marked for disposal and cannot be returned.
Not through FBA. Fulfilment centres refuse meltable inventory from 15 April to 15 October every year, defined as products that melt with prolonged exposure to 155 degrees Fahrenheit, and stock left after 15 April can be disposed of for a fee from May. Summer chocolate is merchant fulfilled or held elsewhere.
The federal label: statement of identity, net quantity, ingredients in descending order, the nutrition facts panel, the allergen declaration and the responsible business's name and address, plus an expiry date Amazon's centres can read on the unit and, for FBA, on the carton.
Usually both: FBA for shelf stable products through the months it is allowed, merchant fulfilment for meltables and short dated stock, with the listing's fulfilment switched by season. Merchant fulfilment removes the shelf life and meltable rules and gives up the Prime badge unless the seller qualifies for Seller Fulfilled Prime.
As reported, a $50 fee per package for hazmat that was shipped undeclared, mislabelled or otherwise out of compliance, charged from 12 July 2026 to the account that bought the label. Declared hazmat carries no handling fee on Ground Advantage and Parcel Select and $7.50 on Priority Mail and Priority Mail Express.
Yes. Perfumes and colognes are flammable and are on USPS's own list of hazmat, along with nail polish, aerosols, hand sanitiser, mouthwash and lithium batteries. The contents decide it, not the size of the bottle.
Within the limits of Publication 52 and with the required markings. New devices with batteries installed ship under quantity and cell size limits. Used, damaged or defective devices containing or packed with lithium batteries ship by ground only with the specified marking.
Most often because the box carried old hazmat markings from a previous life, which USPS's guidance warns can draw the fee on its own. Cover or remove markings on reused cartons, and dispute the fee through the marketplace or tool that sold the label with a photograph of the parcel and the product's documentation.
Both label tools ask whether the parcel contains hazmat when the label is bought, restrict the services to the ones that carry it, and print the marking. Set the hazmat flag on the SKU so every label for it is bought with the declaration rather than from memory.
A supplier sells to Walmart: a merchant decides to stock the product, Walmart buys it on purchase orders and sets the retail price, and it appears on shelves and as a Walmart sold item online. A marketplace seller sells on Walmart.com under its own name, holds its own stock, sets its own price and pays a referral fee per sale.
Walmart's annual sourcing event for products made, grown or assembled in the United States. In 2026 it runs on 6 and 7 October in Bentonville; applications closed on 3 August; submitting is free; selected applicants get a thirty minute meeting with a merchant, with outcomes from a store test to Walmart.com to an invitation onto the marketplace.
A verifiable business with a tax identity, a history of eCommerce sales, products with GTINs, and a fulfilment and returns plan that meets Walmart's delivery promise, through Walmart Fulfillment Services or the seller's own operation. Approval is on the business rather than on a merchant's view of the product.
Marketplace first, unless the brand already supplies national retail. The marketplace is open all year, approved on the business, and produces the sales figures a merchant asks for at Open Call or a category review. A year of Walmart.com sales is the evidence that opens the supplier door.
Yes, and it is the usual route. A merchant who sees a product selling on Walmart.com with reviews and a share of search has a number to test on the shelf, and Open Call meetings often end with a marketplace invitation as the first step toward a supplier relationship.
Search Amazon for the keyword followed by the product's ASIN in the same search box. If the listing appears it is indexed for that term; if there are no results it is not. Indexed is not ranked: an indexed listing can sit on any page.
Usually because it is new and has no sales history on its other terms, so Amazon has indexed the title's brand and little else yet. It can also be an over limit backend search terms field, a wrong browse node, or a child of a variation family whose terms sit on the parent.
Title terms within hours, the rest over days, and competitive terms once there are sales. Sellers commonly report thirty to sixty days for a new ASIN to settle, which matches our experience; exact match advertising on the terms shortens it.
No. Deleting resets the sales history, the indexing progress and possibly the reviews, and risks a duplicate listing violation. A listing that appears for its brand name is live and healthy; a listing that is suppressed is fixed in place through the attribute or image that suppressed it.
Indexing is whether a listing appears anywhere in the results for a term; it is a yes or no. Ranking is where it appears, which sales, conversion and advertising move over months. Fix indexing first, because ranking cannot begin on a term the listing is not indexed for.
Amazon's programme for new branded FBA ASINs launched from 30 July 2026: fee credits equivalent to a 10% referral fee cap on the first 100 units and 5% (or the existing rate if lower) on the next 100, free storage, returns and liquidation on the first 200 units for 120 days with no low inventory or storage utilisation surcharges, and $50 in coupon and $75 in Vine credits usable within 60 days.
Confirm your enrolment on the programme's enrolment page in Seller Central. The introductory period ends on 31 October, and only sellers who have confirmed keep the benefits after that date.
New branded FBA ASINs launched from 30 July 2026: the brand enrolled in Brand Registry, a genuinely new ASIN, and units fulfilled by Amazon. Merchant fulfilled launches do not qualify.
No. Amazon states the two do not stack, so a seller in their first year chooses the set that is larger for the products they are launching.
It ended on 30 July 2026, the day the 2026 programme launched. ASINs launched before that date were under the old terms; the new benefits apply to launches from 30 July.
Because another contribution won. Every seller on an ASIN can submit a title, and Amazon shows the one its trust rules prefer: a brand owner's over a reseller's, one matching manufacturer data over one that does not, and its own systems' where a field is contested. Your title is still stored; it is not the one shown.
The catalogue matched your product identifier, the GTIN or UPC, to a page that already exists. One product has one page, and a matching identifier means Amazon treats yours as the same product. A wrong identifier attaches your offer to somebody else's product.
Amazon does, through its trust rules, and the brand owner enrolled in Brand Registry has the highest trust for its own products' pages. A brand that is not enrolled is an equal contributor with every reseller on its pages.
No. The page belongs to the catalogue, not the offer; deleting the offer changes nothing about the page and resets the seller's history on it. Fix the contribution, confirm the Brand Registry role, or open a case with evidence.
A value a seller submits for an attribute of an ASIN: the title, a bullet, an image, a dimension. Amazon stores every seller's contributions and shows the one its rules trust most for each attribute.
One page of numbers: the unit economics for every product (price, referral fee, fulfilment fee, storage, advertising per unit, landed cost, returns, contribution), the launch cost, the cash tied up in stock by month with reorders, and the break even month. And one column of risks with what each costs to carry.
In six specific ways: the account, the supplier, fee changes, competitors copying the product, stock, and dependence on one platform. Each has a price a plan can carry, from a compliance routine to a second supplier to a margin buffer to a second channel. The risk is in not pricing them.
Usually between the sixth and twelfth month for a single product, once the launch cost and the advertising that buys rank are counted. A plan that breaks even in month two has a mistake in the advertising line; one beyond month eighteen needs a different price, supplier or category.
Enough to hold the stock between paying the supplier and receiving the payout, often four months for sea freight, plus the second order placed before the first sells through, plus the launch's advertising at a loss. Working capital ends more first year FBA businesses than margin does.
Because a reserve was placed after a risk assessment found one of the factors Shopify names: long billing cycles such as annual subscriptions, elevated chargebacks, a rising refund rate, a long delivery timeline, or a sudden surge in volume. The reserve email states the terms, and the payouts page shows the reserved amount separately.
Typically 120 days, as a fixed amount for the period or a percentage of each transaction held for the 120 days after it. Shopify reviews the account before expiry and may keep, reduce or lift the reserve; funds no longer needed are returned in full a few business days after it ends.
By replying to the reserve email within the timeframe it states, with evidence that answers the risk it names. Some reserves cannot be appealed and instead state the conditions for automatic removal; meeting them is the only route.
No. It means the processor is covering itself against refunds and disputes it might have to fund. A pre-order launch or a viral product can trigger one on an honest store. Chargeback and refund rates, shipping on time with tracking, and fulfilment that keeps up with growth are what keep a store off the list.
In Shopify's terms, Delivered Duty Paid means the seller takes responsibility for the import costs and can collect them at checkout; Delivered at Place means the seller is responsible only for shipping and the customer pays the import costs on delivery, usually to the carrier before the parcel is handed over.
An HS code and a country of origin on each product. Without an HS code the estimate falls back to the product's description and category, and without those there is no estimate at all and the customer pays at the door.
Through Shopify's own carrier accounts, DDP labels are available for Canada Post to the United States, DHL Express, DHL Express Canada and DHL eCommerce. With any other carrier on Shopify's accounts a label cannot be bought for an order where duties were collected; a store's own carrier account or 3PL needs its own DDP service.
No. Shopify's help records that as of 29 August 2025 de minimis does not apply to shipments to the United States, so every parcel entering the US is assessed for duty regardless of value.
As eBay reported to sellers: an additional 50% US tariff on certain Canadian goods from 22 August 2026, and Canadian tariffs of 15%, 25% and 50% from 8 September 2026 on listed American products including beauty, apparel, household goods, textiles and lighting. Duty estimates for those categories need checking against a real shipment.
Amazon's name for the percentage of your seller-fulfilled shipments delivered to Amazon Business customers within their operating hours, measured over a rolling fourteen day period. From 30 September 2026 it has to be 90% or higher in the US store, and it is read in Account Health.
No. Amazon states that FBA offers and retail offers are unaffected. It measures only the orders you ship yourself to Amazon Business customers. Moving the products that carry most of your Business volume to FBA takes those deliveries, and the metric, out of your hands.
From 30 September 2026 Amazon sends a notification with recommendations. If the rate has not improved by 30 October, Amazon says your seller-fulfilled offers may be deactivated for Amazon Business customers. Consumers can still buy the offer; the Business channel is what is switched off.
Amazon says shipments that use Automated Handling Time, Shipping Settings Automation and Amazon Buy Shipping together are guaranteed to meet it. Switch all three on, buy Business labels through Buy Shipping, choose weekday morning services for customers with narrow hours, and read the failure report in Account Health for the deliveries that still miss.
Amazon's announcement on the US Seller Forums applies to the US store. The seller press reports the same requirement announced for the United Kingdom and Germany with their own dates; a seller in those stores should read the announcement in their own Seller Central for the effective date that applies to them.
Register your interest on eBay's Live seller page; eBay says it will notify you as soon as you are eligible. Live is in beta with select sellers and eBay has not published the criteria. The seller press reports that registration is now self-service, asks for a business account with business policies and your categories, and can hold an account on a waitlist for review.
eBay's help page lists Collectibles, Luxury, Sneakers, Apparel, Sporting Goods, Home & Garden, Health & Beauty and Electronics, and says more categories are opening. The Live policy bans baby gear, gift cards, sexual wellness items, real estate, firearm parts, alcohol and micromobility devices, restricts vitamins, medicines, blades, pesticides and knives to complete listings shipped domestically, and does not allow chance-based sales.
Not from 15 October 2026. eBay told Live sellers, as the seller press quotes it, that transactions completed through eBay Live will no longer be included in Seller Level, Top Rated Plus eligibility or Service Metrics, and that Live performance will be evaluated separately. Live orders still appear in your order history, reports and payouts.
eBay's policy says cancellations are not allowed for most eBay Live purchases: a bid or a Buy It Now is a binding agreement, and eBay charges the buyer's payment method automatically as soon as the auction ends. Video Receipts of each sale are kept for up to 180 days as the record of what was shown.
Not to host an event, but flash deals, the timed percentage or amount discounts that run for 2, 5, 10 or 15 minutes or the whole event, require a store subscription. You also need a desktop or laptop to create and manage events, a stable connection, and a phone camera or a camera through a broadcast tool such as OBS.
From four numbers: the all in cost to ship an order, the gross margin on the goods, where your order values sit, and how far customers stretch to reach a line. The threshold pays when the margin on the extra items customers add to reach it is greater than the shipping given away to orders that were already above it. A little above the average order value is the usual starting point; a month with the numbers written down is the test.
Two ways. In the shipping zone, add a rate with the price at zero, click Add conditions, choose based on order price and enter the minimum. Or create a free shipping discount, a code or an automatic one, with a minimum purchase amount, an eligibility of countries or customer segments, and the option to exclude shipping rates over a certain amount so express stays paid.
Merchants on the Shopify Community report that a price-based free shipping rate is evaluated on the order after a discount code has been applied, so a code can drop an order under the threshold. Test it on your own store with a test order and your code. If it happens, set the threshold as a discount that combines with the code, or set the line knowing the two cuts can stack.
Usually not. A free shipping discount in Shopify can exclude shipping rates over a certain amount, which Shopify notes applies to shipping rates only and not to order amounts, so the standard rate goes free and express stays paid. A conditional free rate simply sits beside the paid rates, and the customer chooses.
Only if you set it there. A conditional rate belongs to a shipping zone, and a discount can be limited to countries or markets, so a domestic threshold can stay domestic while international orders carry a higher line or none. An international parcel can cost three times a domestic one, and one threshold for both gives that difference away.
Until the tracking moves, eBay has no evidence the item was shipped: its help page says tracking must be from an integrated carrier and must show an acceptance scan to prove shipping. Upload the tracking anyway, reply to the buyer with the posting date, and watch the record; the first scan on the route or the delivery scan closes an item not received request by itself. If the parcel was on an eBay label and is delivered after a refund, eBay reimburses you, removes the defects and excludes the case from your service metrics.
If you have not uploaded tracking, eBay gives you three business days to upload it and respond to the buyer. After three business days without a resolution the buyer can ask eBay to step in, for up to 21 business days from opening the request. A request closes automatically within 24 hours once tracking shows delivery to the buyer's ZIP code.
Yes, when the tracking shows you shipped within your handling time. eBay's seller protections policy says the carrier scan showing shipment within handling time counts even if the item arrives late, and a scan showing arrival by the latest estimated delivery date counts even if it was shipped late. A parcel with no scan at all meets neither test until it is scanned.
Use the Postal Service's SCAN form, PS Form 5630, which lets a clerk record every parcel in a batch against one barcode in a single scan, and keep the receipt. A booked pickup with a scan does the same at your door. Dropping parcels in a slot often gets no acceptance scan, and sellers report the tracking then first moves at the destination facility.
When the estimated delivery date has passed and the tracking is still silent, yes. A refund you issue carries no defect; a case eBay closes in the buyer's favour counts against your seller standards. Open a carrier trace and, on an insured label, the carrier's claim, because the loss is the carrier's to pay, not eBay's and not the buyer's.
Walmart Marketplace's incentive for sellers who go live after 1 February 2026: 20% off base referral fees on the first $50,000 in GMV, 30% between $50,000 and $500,000, and 40% above $500,000, up to $72,000 in total, plus up to $2,000 in WFS credits as 10% off fulfilled items, up to $1,000 in search engine marketing reimbursement and up to $500 in Walmart Connect credits. It runs from 2 February 2026 to 31 January 2027, subject to eligibility.
No. Walmart's page says to go to the payments page after you go live to opt in. A seller who is live and has not opted in pays the full referral fee, and the page does not say the discount is applied to earlier sales.
It is a percentage of the base referral fee, which runs from 6% to 15% by category. In a 15% category, a $50,000 first year saves $1,500 in referral fees, a $200,000 year about $8,250, and a $1,000,000 year about $51,750 across the three tiers, before the fulfilment and advertising credits. In a 6% category the same sales save less than half of that.
Sellers new to Walmart Marketplace who go live after 1 February 2026 and before the window closes on 31 January 2027, from the countries Walmart lists, including the United States, the United Kingdom, Canada, Mexico, Germany, India, China, Japan, South Korea, Singapore, Thailand, Vietnam, Turkey and Chile. All offers are subject to Walmart's eligibility criteria and the savings vary by seller.
Better not. The discount ends with the programme, and a price set on a temporary fee has to rise when it does, while Walmart's pricing rules run regardless. Hold the price where it will still stand after the window, and treat the saving as margin, or as budget for the listing work and the first campaigns, which keep paying after the discount stops.
Yes, and separately from the savings. Walmart's Marketplace Learn describes a New Seller Payment Hold, a rolling delay of part of each payout so a new account builds sales history before it is paid in full. As the policy is summarised there, the delay is up to 14 days for a US-incorporated seller and up to 21 for one incorporated elsewhere, and it ends once 90 days have passed since the first shipped order and 7,500 dollars have been paid out; after that, payouts follow the usual two-week cycle. Treat those figures as Walmart's to confirm in Seller Center, and plan the first quarter's cash around the hold rather than the commission saving.
It runs the whole account rather than one part of it: the listings and the catalogue, the advertising, inventory and FBA, account health and compliance, and the reporting that shows what the channel returned after fees. If a proposal is missing any of those five, it is a specialist service, whatever it is called.
Hire when the account is already selling, the margin can carry the work, and the goal is to take share from sellers ahead of you. Run it in house when you want the capability to stay with you, or when the account only needs its advertising tuned, which a smaller service or a tool can do. An agency can also do the audit and the handover for a team that then runs it.
Contribution by product: what each product returned after Amazon's fees, advertising and fulfilment, compared with a baseline agreed at the start. Revenue and advertising cost of sales on their own can both be improved by spending more, so neither says whether the work is paying.
The brand, always. The seller account, brand registry, advertising console, listings and campaigns stay in your name, with the agency added as a user you can remove. If you leave, you keep everything. An agency that wants the account under its own credentials is asking for a hold over you that no good agency needs.
A score from 0 to 1,000 that TikTok Shop gives every US seller since 29 June 2026, replacing the old Violation Points. It starts at 200, rises with completed orders and passed policy quizzes, falls with violations and performance failures, and counts activity over the last 180 days. TikTok takes enforcement actions when it drops to 150, 100, 50 and 0.
A 7-day restriction on creating new listings and on enrolling in mega campaigns. At 100 the same restrictions run for 14 days, at 50 for 28 days, and at 0 the shop may be permanently deactivated. Passing the milestone quiz TikTok offers shortens the restriction; it does not remove it or restore the points.
Two ways: completed orders, at 4 points per 200 orders in the last 180 days and capped at 20 points a week, and policy quizzes, which earn a variable number of points and are offered after a violation. A successful appeal restores what a violation took. There is no faster route, so the score is protected by avoiding deductions more than by earning them back.
Yes. On the US policy both earned and lost points reset 180 days after the date they were earned or lost, so the rating is always a picture of the last six months. The 90-day figure that circulates is TikTok's rule for Indonesia, the Philippines, Singapore, Thailand and Vietnam, not for a US shop.
At 0 points TikTok says the shop may be permanently deactivated and that reactivation is at its discretion. The route is an appeal on the violations that took the score down, with evidence for each; the realistic position is that a shop should never reach zero, because every milestone before it was a warning with a fixed cost and a quiz to shorten it.
By performance over a rolling window, which Walmart's 9 September 2026 release note shortened to 14 days. As reported from that note, the item needs an on-time delivery rate above 90% and a cancellation rate below 2.5% over the window; Walmart's badge guide in Seller Center is the place to confirm the thresholds. Items fulfilled by Walmart Fulfillment Services carry the badge by the nature of the service.
Since 4 September 2026, eligible Pro, Rising and Advanced sellers get up to 35% off 2-day labels and up to 15% off 3-day labels bought through Ship with Walmart, applied automatically, provided the shipping speed matches the delivery promise the listing made. A 2-day label on an order promised in five days earns nothing.
Because the last fourteen days fell below the thresholds. The window that lets a seller recover in two weeks also drops an item in two weeks, so a bad fortnight of late deliveries or cancellations removes it. Open Speed Insights in Seller Center to see whether the promise failed at handling, at the carrier or at the promise itself, fix that one, and the badge returns within the next window.
Only if the operation keeps it. A shorter promise earns the badge and the label discount, and it also raises the standard the on-time rate is measured against, so a promise the warehouse cannot keep on its worst day loses both within fourteen days. Read Speed Insights first, fix handling if that is where the gap is, and tighten the promise on the listings the operation can honour.
Yes. Items fulfilled by Walmart Fulfillment Services carry the badge as part of the service, which is one of the things the WFS fee buys. The badge rules and the label discounts in this article are for seller-fulfilled items; the choice between the two is a fulfilment decision, covered in the WFS article.
Mostly not. eBay's rule is that a coded coupon does not stack with an order discount or volume pricing, that a buyer gets the best single discount when several apply, and that a coupon does stack with a sale event. A shipping discount can sit on top of any other. Bundle discounts, as reported from eBay's help page, combine with eBay's own promotions and with sale events but not with seller coupons, order discounts, volume pricing or promotional codes.
Volume pricing is a tiered discount on multiples of one item or a group you choose, set from the Marketing tab and shown on the listing. A bundle discount is a tiered discount for buying two, three or four or more items from your whole inventory, switched on for everything and trimmed by an exclusion list. A buyer gets one or the other, never both, so they are alternatives rather than layers.
For the six Seller Hub tools, yes: eBay's own page says sale events, order size discounts, volume pricing, coded coupons, offers and shipping discounts are free for Store subscribers, and a seller without a subscription does not see them. Bundle discounts are the exception; eBay opened them to all sellers in September 2026.
Usually one of three rules. An item with a Best Offer or a seller-initiated offer on it takes the offer and refuses the coupon. A coupon does not stack with an order discount or volume pricing, and the buyer gets whichever is best. And a coupon campaign pauses itself from display once its discount spend reaches 90% of the budget you set, which looks like a coupon that stopped working.
No, but once on they apply to everything. As reported from eBay's help page, bundle discounts are opt-in and off until you switch them on; after that every eligible listing is included, new listings join automatically, and you can exclude up to 100 individual listings. Exclude the low-margin lines before switching on, and check the discounts dashboard after the first weekend.
A sale event on the lines you want to move, because it shows the strikethrough in search and stacks with the most, with a public coupon over it for shoppers comparing sellers. Then one of volume pricing or a bundle discount, not both, for bigger baskets. Set every percentage from the margin first, and read the dashboard's lift figure per campaign before extending any of them.
The enrolment fee is tiered by the units enrolled per parent product: up to 2 units is $0, 3 to 10 units is $75, and 11 to 30 units is $200. Amazon does not charge it until 30 days after enrolment and after the first review is published, and if no Vine review arrives within 90 days of enrolment the fee is not charged at all. The units themselves and the FBA fees on them are the larger cost.
A seller with a Professional selling account and a Brand Representative or Reseller role on a brand enrolled in Brand Registry, or with generic products in the catalogue. The product must be fulfilled by Amazon and have fewer than 30 reviews. Adult, digital and bundled products are not eligible.
Up to 30, which is the enrolment ceiling per parent product. Amazon's own wording is up to 30 high quality reviews. Sellers commonly report that some enrolled units are ordered without a review following, so thirty units enrolled is not thirty reviews guaranteed.
Only if it has fewer than 30. Vine is built for a listing at the start, not for adding reviews to an established one. A product at 30 or more reviews is not eligible however new the listing feels.
Amazon does not say so, and the reviewer is under no instruction either way. Sellers commonly report that Vine reviews run more critical than organic ones, which is consistent with a reviewer who paid nothing and has no reason to round up. The practical answer is to enrol a product that is ready, because whatever is wrong with it will be written down.
Under Orders, on the Returns page, through the View returns insights button, as Walmart's Marketplace Learn guides describe it. Walmart's release note of 11 September 2026 says the dashboard now shows return metrics, trends and reasons, item level details including ratings, and personalised recommendations by category.
Walmart does not publish a target in the release note, and the right number depends on the category: apparel and footwear return at multiples of what hard goods do. The useful comparison is the dashboard's category recommendation and your own items against each other. A product well above the account's median is the one to fix first, whatever the absolute number.
It depends on the return policy the seller set for the item within Walmart's rules and on the reason for the return. The rules are in the Returns and refunds guides in Marketplace Learn and in Walmart's customer return policy, which are the pages to read in Seller Center; this article does not quote figures from them because Walmart's help pages are not readable from outside a signed in browser.
Directly it affects margin, which is the arithmetic in this article. Indirectly a product that is returned often is usually rated lower and converts less, and both feed the offer's position. Walmart's dashboard now puts the return data beside the ratings for exactly that reason.
The listing, for fit and expectation returns: a size guide the customer can measure against, a first image that shows scale, and a description that says what the product is not for. It is the cheapest change and it moves the largest group of reasons. Packaging is second, delivery speed third and the product itself last, because it is the slowest to change.
The 30 most recent interested buyers, which means people watching the item or holding it in a cart for at least 2 days. Each buyer can receive up to 3 offers on an item, and each offer is valid for 96 hours or until the listing sells.
At least 5% below the Buy It Now price, on the item rather than on shipping alone. Five percent is also the sensible default, because the notification and the four day clock do most of the work and a deeper first cut leaves no room for the second or third offer.
No fee beyond the standard fees for fixed price listings. The one cost that can follow is an advertising fee: if the buyer clicked a Promoted Listing for the item within the previous 30 days, the ad fee is charged on the final sale amount whether or not the sale came from an offer.
Yes. After the first offer eBay asks whether to send the same offer to future interested buyers automatically, at up to 30 offers every 48 hours. Store subscribers can set rule based offers by category, price range and condition for up to 150 days. Automated offers can be found and switched off from Active listings by filtering for offers sending automatically.
Not on auctions, and not in the motor vehicle category. On multi variation listings eBay's help page says the feature is not available, while its Seller Center page says such items require parent SKU IDs; the two pages differ, so treat a variation listing as eligible only when the tool offers it on that listing.
For a regular order, 1 to 2 business days by default, and the order must be In Transit within 2 business days of Awaiting Shipment. Longer windows apply to order types declared as such on the product: made to order 3 to 14 business days, backorder 3 to 15, custom handling 3 to 7, and pre-orders 3 to 60 calendar days.
The order is auto-cancelled if tracking has not been added within 5 business days, and that cancellation counts as seller fault against the 2.5% Seller Fault Cancellation Rate the policy requires.
A regular order must be marked as Delivered within 6 business days of Awaiting Shipment, which is six business days from the order rather than from the ship date. The On Time Delivery Rate the policy requires is 80% or above.
95% or above. Valid means accurate and verifiable tracking numbers entered with the correct carrier name and service, which is why labels bought through TikTok Shipping, where the tracking is recorded automatically, are the safer route.
The policy recommends 4% or below and states that above 10% triggers enforcement, which can mean point deductions on the Account Health Rating, product delisting, order volume limits and, for serious violations, account deactivation.
From 16 July 2026 a collection can combine automated conditions, hand picked products, exclusions, other collections and app sources in one place; it can hold specific variants such as a colour or size; and it can be used as a source inside another collection and stay in sync. The same collection applies to discounts and taxes. Shopify describes it as a new model replacing the legacy manual and smart collections.
They migrated automatically to the new model with no breaking changes, according to Shopify's changelog. A former manual or smart collection is now a collection with one source, and nothing had to be deleted or recreated.
Yes. A collection can be built around specific variants, and those variants show in the collection page, in filters and in sales channels including the Online Store and POS. Shopify's own example is a collection of only red products for Valentine's Day.
Shopify's changelog says developers must update to API version 2026-07 for compatibility, so an app on an older version may not read or write a multi-source or variant level collection correctly until it is updated. Check every app that touches collections before rebuilding.
It removes the reasons stores ended up with several collections answering one search term. Nesting lets one page own a term with sub collections feeding it, and variant level collections let a colour or size query be answered by a real page rather than a filtered URL. The page still needs its own copy, title and structured data, and filtered URLs still need the same canonical and robots handling as before.
They are collected under a verification process rather than submitted anonymously: a reviewer needs a LinkedIn account, a Google account or a company email address, and Clutch says it verifies reviews using reviewer-provided information and online research. It deliberately does not publish the specifics of those checks. Where it cannot confirm the reviewer or the project it still publishes the review, marked Not Verified, so the status on each review is the thing to read rather than the platform's reputation in general.
In Clutch's words, that it could not verify the reviewer's identity or project details but judged the information useful to buyers, so it published the review with that label. A reviewer or the agency can complete the missing details afterwards and the review can be verified after publication. Treat one as unfinished rather than false, and weigh it below the verified reviews beside it.
The project size, the duration, the reviewer's role and the industry on each review. Almost every agency's average sits between 4.6 and 5.0, so the rating separates nobody. The details tell you whether the agency has run work like yours, at your budget, for long enough to have hit a difficult quarter.
Ratings come from client reviews, and the review process is the verification described above rather than something an agency writes for itself. Directories in general do sell sponsorship and advertising placement, which is why the sensible habit is to read the reviews and their details rather than the position in a list, on any platform.
Fewer than you would think, and more recent than you would accept. Five detailed reviews from the last eighteen months, in your channel and your budget band, are worth more than forty across six years and nine industries. What you are testing is whether the agency has done your job recently, not whether it has been busy.
No. Google's policy removes content posted because of an incentive such as payment, discounts, free goods or services, and prohibits a business from offering those incentives in exchange for a review. Asking a happy customer to leave a review is fine; paying for one in any form is not.
No. Google treats them as a conflict of interest, which covers current or former employment, a contractual or consultory relationship, and other professional or personal affiliations. A business also may not ask staff to solicit a set number of reviews or to solicit reviews containing particular content.
Look at the shape rather than the score. A cluster of reviews in one week from accounts with no other review history is the oldest pattern on the platform. Reviews spread over time, from people with review histories, describing specific work, are the ordinary look of a business that asks its clients properly.
It is not disqualifying for a business-to-business service, where most clients never think to leave one and the substantive reviews live on the platforms built for buyers. What matters more is whether detailed, recent, verifiable reviews exist somewhere, and whether they describe work like yours.
It removes content that is not based on a real experience or that misrepresents the location, and its enforcement is not limited to deleting the review: Google states that behaviour harming its users, community, employees or ecosystem can lead to actions ranging from suspending account privileges to terminating the account.
Trustpilot publishes the method in its own help centre, and that is where to read it. It is widely reported to be a weighted calculation rather than a simple average of the stars: recent reviews count for more than old ones, and a profile with very few reviews is pulled towards the middle of the scale until it has enough of them to be meaningful. We could not load Trustpilot's own article on the day this was written, so treat the mechanics here as reported.
Because a profile with a handful of reviews is reported to be weighted towards the scale's middle until more arrive, which stops three reviews carrying the same authority as three hundred. It is the platform limiting what a small sample is allowed to claim, not a penalty against the company.
Yes, with a change of expectation. An agency has tens of clients where a retailer has thousands of orders, so read the content rather than the volume: five reviews that name the channel, the problem and the outcome are stronger evidence than two hundred one-line ratings.
A business can report a review it believes breaches the platform's guidelines, and the platform decides. It cannot simply remove a review it dislikes, which is the reason a profile with visible negatives and specific replies is more informative than a page of curated testimonials.
The most recent five reviews, then every negative one with the company's reply, then the dates. Recency tells you what the company is now, the replies tell you how it behaves when something goes wrong, and the dates tell you whether the reviews arrived naturally or in one campaign.
By a multi-factor evaluation rather than reviews alone. In its own words the criteria are verified client reviews collected directly through the platform, the depth and breadth of service offerings within the category, portfolio quality and evidence of relevant project delivery, market presence and industry recognition, and business maturity and operational reliability.
GoodFirms names verified client reviews collected directly through the platform as one of its ranking criteria, which means the review goes through its own process rather than being reproduced from a company's website. As on any platform, read the individual review for whether it describes a real engagement in specifics.
It means a stronger profile across the five criteria, only one of which is client reviews. A company can rank well on breadth, portfolio and presence while being the wrong fit for your channel, which is why the listing is a way to build a shortlist rather than a way to choose from it.
Use it for the shortlist, not the decision. A directory with a research methodology surfaces companies that can be described and checked, which is a better starting point than a search result. The decision comes from reading what they have published about the work and asking who runs your account, what month eight looks like, and what happens when something breaks.
Both are business services directories with a review process, and they weight things differently: Clutch's profiles put the project details of each review in front of you, while GoodFirms names a broader set of ranking criteria including portfolio, service breadth and business maturity. Read both for the same thing, which is evidence of work like yours, recently, at your size.
DesignRush says its ranking is built on a Base Score made up of many factors, and names five as the most important: reviews, portfolio, awards and press, team bios, and top services. Four of the five are things an agency supplies on its own profile, so a high organic position reflects a complete, evidenced profile as much as a strong one. Sponsored placements sit above the organic list and are paid for.
It can pay for a placement above the organic results. DesignRush states that sponsored listings pay for a position above organic ones, and that sponsorship does not affect star ratings or review content. So a paid slot buys the top of the page and nothing about the rating beside it.
Because the rating is deliberately smoothed. DesignRush uses a Bayesian method that adds a 3.5 star rating to every profile as a baseline, which stops a handful of perfect reviews from producing a perfect score. A 4.7 across many reviews is a stronger signal than a 4.9 across three, and reading the review count first is the way to use it.
It means the DesignRush team has checked that the agency is what it says it is. Reviews are verified separately: reviewers are authorised through LinkedIn, may be checked by email or a call, and claims in a review are fact-checked. Neither badge says the agency has done your particular job, which is the thing to establish yourself.
Not usually. A category such as digital marketing holds brand agencies, performance agencies, content studios and marketplace operators, ranked against each other although they do not compete for the same brief. Compare the services each profile declares before comparing positions.
Yes. Amazon's wording is that titles need to be 75 characters or less including spaces. A title of roughly twelve words is already at the ceiling, which is why counting words rather than characters leads sellers to underestimate how many of their listings are affected.
Media only. Amazon's announcement applies the limit to titles in all categories except for media, which covers books, music and video. Every other category is included, whatever the norm in it used to be.
Amazon said that after 27 July 2026, titles still over the limit are updated to its AI recommendation gradually, and that listings stay active while this happens. Brand owners get 14 days before implementation to review, modify and approve the recommendation in Review Listing Changes. Sellers without Brand Registry are not offered that window.
In Manage All Inventory, select Edit on the listing and click View enhancements on the left of the page. The recommended title and Item Highlights appear there, and you can accept them, change them or write your own instead.
Both. Amazon states that the content is searchable and visible with titles in search results and on product detail pages. That makes it the right home for comparison attributes such as material, count or compatibility, and the wrong home for a list of keywords a shopper would notice.
More than 100 orders in a rolling 90 day window, and the account must have been active for at least 90 days since onboarding or reinstatement. That gate applies to all three tiers, so a healthy but low volume account can be blocked by order count alone.
The same six metrics at rising thresholds. On time delivery moves from 90% to 92.5% to 95%, cancellation rate from 2.5% to 2.0% to 1.5%, response rate from 90% to 92.5% to 95%, shipping speed score from 5% to 25% to 50%, and both content quality and price competitiveness from 65% to 70% to 75%. Pro also requires a seller rating of 4.0 or higher.
In our experience it is shipping speed score. The other five respond to operational discipline, but 50% of the catalogue available at the faster speeds is a fulfilment decision about where stock sits, not a process fix.
On the fifth and the twentieth of every month. When the criteria are met you receive the highest tier you qualify for at the next refresh, so a fix made early in a month can be reflected within about two weeks.
The badge itself is a modest conversion signal. The referral fee discount is the real number: 5% at Advanced and 10% at Pro, taken off the largest variable cost in a marketplace account. Work out what that is worth on your own volume before committing to the fulfilment change that shipping speed score usually requires.
eBay does not publish the list. It says withholding it protects the integrity of the marketplace and prevents misuse of the policy, and that you will be notified when you attempt to create a listing in a category that is not currently available to your account. The practical test is to attempt a listing before you commit to stock.
The policy applies to sellers whose registered account address is in one of the covered countries. That list is long and includes markets such as India, Brazil, Mexico, Nigeria and Vietnam. Check eBay's policy page for the current list rather than a summary of it, because countries can be added.
Yes. eBay states that existing listings which do not meet policy requirements may be ended. That is the main reason to treat it as a stock risk rather than as a listing inconvenience.
There is no application described. eBay says it reviews relevant accounts periodically and that an account meeting the qualification criteria may be enabled in future. The four things named are account standing, selling history, buyer experience performance and support level, so the work is clearing policy issues, building history and improving dispatch, tracking and case outcomes.
No. A selling limit caps how much you can list and sell, is visible in Seller Hub, and has a route to request an increase. Category readiness removes access to a category entirely, is not published, and reopens only on a periodic account review.
Yes, if you are authorized to use it. TikTok's requirement from 5 August 2026 is that return addresses are ones the seller owns or is authorized to use and that are affiliated with business operations. A supplier address without the supplier's agreement does not meet that, and neither does a warehouse you have no relationship with.
TikTok names residential addresses among those that are now an enforcement issue, alongside unauthorized, unaffiliated and non verified addresses. If your shop is run from home, a business address or an authorized 3PL is the safer setup.
No. From July 2026 forward shipping fees are no longer refunded for buyer responsible return reasons. On a low margin item sold with free delivery this can remove the contribution from several sales, so it is worth reviewing pricing on the products that return most.
For certain refunds, only once the returned package is marked as delivered. Review used to be possible while the parcel was in transit. If your process triggers on the tracking number rather than the delivery scan, it needs changing.
In stagnant reverse logistics cases with no update for 10 or more days, the logistics partner becomes liable. Eligible package not received refund only cases on 4PL shipments also moved from seller review to platform review. Both mean a stalled return is a case to escalate rather than a loss to absorb.
Nothing visible. From that date apps can no longer create or update script tags: the scriptTagCreate and scriptTagUpdate mutations return a user error and the REST ScriptTag resource rejects POST and PUT. Existing script tags keep running until 1 March 2027.
Shopify stops injecting script tags into storefronts, so anything still delivered that way stops appearing. Typically that means analytics and pixels, review widgets, popups, cookie banners, currency selectors and upsell tools. Nothing errors and nothing warns you, which is why the usual first sign is a reporting number that looks wrong.
No. Shopify states that the deprecation applies to all API versions, including older ones, so pinning will not defer it. An app pinned to an old version is on the same timetable as everything else.
An app embed block shipped in a theme app extension for anything that renders or behaves on the storefront, which the merchant activates in the theme editor. A web pixel for anything whose only job is analytics, which Shopify notes needs no action from the app user.
Ask each app developer or read their changelog for whether they have moved to a theme app extension or a web pixel and whether you have to activate anything. Then check App embeds in the theme customiser, since a migrated app with its embed switched off is the commonest failure. Finally view source on a product page and account for everything loading from outside your theme, including leftovers from apps you have uninstalled.
It is Amazon's connection between a selling account and an outside AI tool, announced on 23 September 2026. It launched with Amazon Quick and is in beta with Claude. Once connected, the tool can reach listing contributions, real-time performance metrics, inventory levels and sales analytics, and Amazon says it can act on the business, not only read from it.
Not yet. Amazon says the plugin is for sellers in its US stores, with international expansion to follow. A seller selling only in other Amazon stores cannot connect it for now.
Amazon says every plugin interaction is protected by clear boundaries on what it can access, human approval on actions, and complete audit trails. An action needs a person to approve it. The practical protection is deciding in advance who approves and which kinds of change they approve, and reading the audit trail regularly.
Every primary account holder can sign up for a free 12 month Amazon Quick Plus subscription, available until 31 December 2026, and can name two co-workers to receive a free year as well. The primary account holder is often not the person who runs the account day to day, so the co-worker places are worth giving to the people who will use it.
No. Amazon says connecting takes roughly 60 seconds and needs no coding. That is also the reason to decide beforehand who on the team connects it, because it is quick enough for anybody to do on impulse.
Often not. Walmart's guide says most sellers have an enrolled status 60 days after becoming a Walmart Marketplace seller, and you must be an active seller in good standing. Another Walmart guide describes the program as invitation only with applications reviewed first come, first served. Check the program status in Seller Center and apply if it is not showing.
Walmart says there are no additional costs to joining the program, but standard Walmart Marketplace referral fees apply to business orders, as they do to Walmart.com orders.
Updates can take from 15 minutes to four hours to appear, only logged-in business customers see the business price, some items are excluded by product type, and if your Walmart.com price is lower, Walmart Business displays the lower price instead. The B2B item report shows which of your items are eligible.
Walmart may automatically pause offers whose volume pricing goes too low, to protect you from unintended loss. Adjust the discount tiers or the item price within your business strategy so the price does not fall too far as quantities rise. Once the pricing rules are met, the offer can take up to 48 hours to return.
Not currently. Walmart says Sponsored Product ads are not available on Walmart Business, though there are display opportunities you can ask a Walmart Connect manager about.
From 23 October 2026, for orders shipped to US addresses on stores that use Shopify Tax or Tax Platform. Shopify announced the change in its changelog on 25 September 2026.
No. Shopify's return rules page says tax isn't calculated on restocking fees. The change on 23 October applies to return shipping fees only.
No. Shopify says tax is not calculated on return fees if you use another tax service, such as Manual Tax or Basic Tax, if the order was shipped outside the United States, or if the order is tax exempt.
No. Your existing shipping tax settings still apply, so if shipping is set to untaxed in a state, return shipping fees in that state remain untaxed. Shopify also only calculates the tax where the fee is taxable in the jurisdiction the order was shipped to.
No. Shopify says return fees aren't automatically deducted from refunds, and you need to deduct any applicable return fees manually when you create the refund. That does not change on 23 October.
eBay sends payment reminders 24 and 48 hours after the listing ended. If the buyer still hasn't paid four calendar days after the listing ended, you can cancel the order and choose Buyer hasn't paid as the reason.
eBay says it is important to cancel any unpaid items within 30 days from the day the buyer commits to buy, so the item can be relisted.
When you cancel an unpaid item, eBay removes any feedback left by the buyer of that item.
Yes. Buyer requirements let you block buyers with a history of cancelled orders because they didn't pay, and limit buyers who are winning or have bought a set number of your items in the last 10 days. You can also add up to 5,000 usernames to your blocked buyers list. eBay has also said that buyers with repeat unpaid items across multiple accounts will be blocked from bidding on auctions.
eBay's September 2026 seller update announced that it may verify additional buyer details before auctions and will block buyers with repeat unpaid items across multiple accounts, but it did not give dates. The existing settings work in the meantime.
The importer of record is the party responsible for a customs entry into the United States: that the goods are declared, classified and valued correctly, and that duties are paid. Its identity is registered with CBP on Form 5106.
Yes. Under a Federal Register notice published on 19 August 2026, taking effect 30 days later, CBP may immediately void an importer of record number and take other enforcement action if it determines that the Form 5106 information is inaccurate or incomplete.
For now, yes, but CBP says foreign importers will face restrictions on their ability to file entry, and CTPAT-validated customs brokers must do additional vetting on their foreign importer clients. Ask your broker what that means for your shipments.
It is the executive order that, in CBP's words, directs comprehensive reform of US customs and trade laws to strengthen enforcement, combat duty evasion, modernize processes, and protect national security and the domestic economy. CBP's page on strengthening customs enforcement lists the changes that have followed from it.
The marketplace and its fulfilment centres receive the stock, but the importer of record is whoever is named on the customs entry, usually the seller, the supplier or an importer arranged through a forwarder. Check the entry summaries for your shipments to see whose name is on them.
It is Meta's subscription for businesses, rolling out from 15 September 2026. Four plans, Essential, Advanced, Expert and Max, bundle a verified badge and impersonation protection, Meta Business Agent responses on WhatsApp, publishing tools, analytics and, on the higher plans, human account support, across Facebook, Instagram and WhatsApp.
Meta lists Essential at $14.99 a month, Advanced at $49.99, Expert at $149.00 and Max at $499.00, priced per profile. Meta also says plans, benefits, pricing and availability may vary by region, by app and by account, so check the prices shown on your own account.
Meta says it is bringing the best of Meta Verified into Meta One, that nothing changes right now for current Meta Verified subscribers, and that it will notify them once they are able to make the switch. There is nothing to do until that notice arrives.
Nothing Meta has published says so. The announcement does not mention ads at all. Treat a plan as a subscription for the account's presence, messaging and publishing, and judge the advertising on its own results.
Per profile. Meta's wording is that each plan covers one account, so a business that wants the same plan on several accounts pays for each of them.
An electronic base application is $350 per class of goods or services. Surcharges can add $100 per class for missing information, $200 per class for describing the goods in the free-form text box, and $200 per class for each further 1,000 characters of free-form description beyond the first 1,000.
It is a $100 per class fee on an application that leaves out any of the first nineteen items the USPTO lists as required for a base application, such as the applicant's domicile address, entity type, citizenship, basis for filing, verified statement, or a translation of non-English wording in the mark.
Choose the goods and services from the ID Manual inside the electronic application instead of typing them. Typing or pasting ID Manual wording into the free-form box still counts as free-form, because the examiner cannot tell where it came from.
Yes. Amazon's Brand Registry page accepts an active registered trademark or a pending application from the designated trademark office of a country with an Amazon store, and the trademark text must match the brand name on the Brand Registry application.
Yes. The USPTO's rule requires an applicant whose domicile is not in the United States or its territories to designate an attorney as its representative and include the attorney's name, postal address, email address and bar information in the application.
It is a search type Google added to the Performance reports on 24 September 2026. It reports searches made with an image rather than words: Google Lens, Circle to Search on Android, image uploads to Google Search and Chrome's Search this image feature.
Yes, since September 2026. Lens searches are counted under the multimodal search type, in the Performance report for Search results and in the report for Generative AI features.
In the Performance report, choose the multimodal search type in the filter at the top. The same filter is in the report for Generative AI features, and the data can be exported like any other performance data.
Google says the metrics appear if a site is receiving traffic from these searches. If none appear after the rollout reaches your account, camera and image searches are not yet reaching your pages, and the image work on your product pages is the place to start.
Follow Google's image guidance: product photos as standard image elements, descriptive file names, alt text that describes the product in context, high quality images, image sitemaps where images are hard to find, and product structured data that includes the image. Then read the multimodal filter monthly to see which pages it reaches.
Apply through Amazon's Renewed programme with your seller type, itemised invoices showing at least $50,000 of qualifying refurbished purchases in the last 90 days (180 days in Home, Tools, Kitchen, or Lawn and Garden), trademark or Brand Registry evidence if you are the brand, and photographs of a sample product from all six sides with its packaging and accessories.
Amazon asks for itemised invoices totalling at least $50,000 in qualifying refurbished purchases within 90 days, or 180 days for Home, Tools, Kitchen, or Lawn and Garden products. For Apple branded items the minimum is $2.5 million over the same periods.
Amazon lists six seller types: distributors or resellers, manufacturers or brand owners, brand-authorised refurbishers, refurbishers, repairers, and IT asset disposition providers. A Professional selling account is required.
Under Amazon's fair pricing policy, Renewed products must be listed at a discount of at least 5% against the same product new. Because the new price changes, the Renewed price needs checking against it regularly.
Amazon says approval decisions are typically made within 10 days, and that processing times may vary. A complete application with clear invoices and the right photographs avoids the back and forth that makes it longer.
On 31 August 2026 the FTC and 22 states sued Amazon, alleging that it described its Sponsored Products ads as a second price auction while a hidden surcharge meant advertisers paid their own winning bid close to 80% of the time. The case is in the US District Court for the Western District of Washington, and the allegations have not been decided.
According to the FTC, it is the name Amazon gave to the surcharge the complaint says was hidden in the ad auction, raising the price advertisers paid above what the next highest bid would have set. The FTC quotes internal documents describing it as an invented auction participant and shill bids.
Not as things stand. The FTC's announcement does not describe the remedies being sought, no court has ruled, and no refund process exists. Be wary of anybody offering to recover money from it now.
Compare the average cost per click with the bid for each keyword or target in your Sponsored Products reports. If the cost per click sits at or just under the bid consistently, you are paying close to your bid on those clicks.
Set bids from what a click is worth to you, the order value times the conversion rate times your advertising cost of sale target, and treat that bid as the price you may pay. That holds whatever the case decides.
On 24 September 2026 the FTC voted to publish an advance notice of proposed rulemaking asking whether platforms' ad optimisation helps impersonation scams, and whether platforms should vet advertisers, monitor ads, and investigate and remove impersonation scam ads. Comments are due 60 days after it is published in the Federal Register.
Through Meta's Intellectual Property Help Center, one report at a time, or through Brand Rights Protection, a programme brands apply to that searches ads, Pages and Instagram accounts for impersonation, counterfeits and trademark or copyright misuse.
File a trademark complaint through Google's legal reporting route. Google only accepts complaints against specific advertisers identified by their URLs, so record the address the ad leads to. Google's policy does not restrict using a trademark as a keyword.
Amazon Brand Registry requires an active registered trademark or a pending trademark application for the brand name or logo. Enrolment gives access to Report a Violation, Transparency, Project Zero and the Counterfeit Crimes Unit.
No. An advance notice of proposed rulemaking starts the process and places no new duty on platforms. Until a rule is made, if one is, removal depends on each platform's own policies and on reports from the brand.
Shopify Checkout in 2026: The Changes That Reward Signed-In Customers, and Which Apply to Your Store
Legacy customer accounts were deprecated on 26 February, default addresses arrived in checkout on 13 April, the sign-in page became editable on 20 May, and on 21 August checkout began pre-selecting a signed-in customer's last used payment method where saved payment methods are enabled.
Shopify's help centre says saved payment methods are available only to stores on Shopify's Enterprise offering, including commerce components, with customer accounts set up, and that they are turned off by default. They work alongside Shop Pay.
They are deprecated. Since 26 February 2026 they are not available to new stores, existing stores can keep using them for now, and Shopify says it will announce a final sunset date later in 2026.
A signed-in customer's default address is set automatically the first time they enter an address or place an order, replaces the old last used address behaviour, and can be changed at checkout or in their account.
Ask after the first purchase, on the order status page and in the confirmation email, rather than before it; use the editable sign-in page to say what signing in gets them; turn on passwordless sign-in through the newer customer accounts; and give signed-in customers a benefit guests do not get.
It is Amazon's AI shopping assistant for US customers on the Amazon Shopping app, the website and Echo Show devices. Amazon folded Rufus into it this year, and it compares products, makes personalised recommendations and tracks prices.
On a product page, Amazon says it answers from the listing details, customer reviews and community Q&As. Rufus was trained on Amazon's product catalogue, customer reviews, community Q&As and information from across the web.
They are AI-generated prompts added to existing Sponsored Products and Sponsored Brands campaigns, shown in shopping results and on product pages, that can open a dialog in Rufus. They became generally available in the US on 25 March 2026, campaigns are enrolled automatically, they are charged within the campaign's cost per click, and they can be paused in the Ads Console.
Collect the questions buyers ask in reviews, Q&A and customer messages, group them, and make sure the bullets, description and attributes state the answers in words: measurements, materials, compatibility, what is included and who it suits. Amazon has not published how the assistant ranks products.
Yes. Ordinary search still runs on phrases, and Brand Analytics and the search terms report still measure that demand. The questions are an addition to keyword research, not a replacement for it.
Work that is done the same way most of the time, can be checked within a week, and runs on a platform's clock. Customer messages usually come first, then catalogue upkeep, then order and inventory administration.
The decisions about what to sell, what to charge, what to spend on advertising and what the brand says, and ownership of the accounts themselves: the primary login, the payment details and the power to add and remove users.
Sellers describe a 24 hour expectation that includes weekends and holidays. Amazon's help page on responding to messages sits behind the Seller Central sign in, so check the current rule there.
No. Give each person their own login through the marketplace's user permissions, with only the access their tasks need. A shared password cannot be withdrawn from one person without changing it for everyone.
Agree the measures before the handover: response time and resolution for messages, errors and suppressions for listings, backlog and deadlines for administration. Read them weekly for the first month, and run the old and new ways side by side for the first week or two.
From September 2026 for Search campaigns using automatically created assets or the campaign-level broad match setting. Google moved the Dynamic Search Ads sunset and auto-upgrade to begin in February 2027.
For campaigns using automatically created assets or campaign-level broad match, search term matching and text customisation are enabled by default. For Dynamic Search Ads, all three features, search term matching, text customisation and final URL expansion, will be enabled.
It is the AI Max feature that sends a searcher to the page on your site Google judges most relevant to their search, instead of always to the ad's final URL. URL controls let you exclude pages that should never be used.
In the search terms report, filter the match type column for AI Max or add the Source column. The Search terms and ad combinations view shows the headlines and landing pages used. Google notes the AI Max filter may show lower numbers because it excludes Other search terms.
Google recommends it, so the move happens with your controls set rather than the defaults. Set brand, location and URL controls first, agree the cost or return the campaign must hold, and review it after two to four weeks.
Yes. Since 29 September 2026 Shopify lets you set return window overrides by market for specific collections, products or variants, under Settings, Policies, Return and cancellation rules.
Shopify applies the shortest matching return window. A product in both a 30 day and a 14 day collection gets 14 days.
No. Shopify says items marked as final sale can't be returned or cancelled, even if they also match a return window override. In the EU, check the legal right to withdraw before marking goods final sale.
Yes. Shopify's return rules apply whether a return happens in store or through self-serve returns online, and in point of sale the return window and final sale rules are evaluated on each item in an order.
EU consumers have the right to withdraw from an online purchase within 14 days of delivery without giving a reason, with exceptions such as perishable, personalised and unsealed media goods. This is not legal advice; check each market's law.
Testing, inspection and certification. For categories including dietary supplements, children's toys, e-mobility products and pressure washers, Amazon takes compliance results only from an approved third-party provider. Sellers can no longer upload the documents themselves; the provider tests the product or verifies an existing report and reports to Amazon.
No. Since 3 September 2025 children's toys in the US and Canada stores need annual testing or document verification from an approved TIC organisation, and Amazon says compliance results must come directly from the provider.
Amazon says testing typically takes 30 days once the provider receives samples, and that the provider contacts the seller within two business days of the request.
The address issue date in the request email. Amazon says to submit the request for lab service by that date to avoid any impact on the listing. The testing itself can finish later; the request cannot.
In Account Health, open Manage Your Compliance, select Add or Appeal Compliance, choose Appeal Request, and submit the reason with supporting details. Keep the lab service request ready in case the appeal is refused.
A tool in Seller Central, under Reports and then Selling Economics and Fees, that analyses fees, costs and profitability across the catalogue, including sales, advertising spend, refunds and returns, from the past two years through the next 90 days, per ASIN or SKU, with a price simulator and a place to enter off-Amazon costs.
No. It lets you factor in off-Amazon costs, which means you enter the cost of goods, freight, duty, packaging and prep yourself. The profit it shows is only as accurate as those entries.
Selling price minus the referral fee, the fulfilment fee, storage per unit, advertising per unit sold, returns cost per unit, and the landed cost of the goods including freight, duty and prep. Do it per SKU and per channel.
Amazon said FBA fees would rise by an average of $0.08 per unit sold, effective 15 January 2026, with no new FBA fee types in 2026. Averages hide the products that moved more, which Profit Analytics can show per SKU.
As a percentage of the total price or a minimum amount, whichever is greater, with the percentage set by category. Amazon's pricing page carries the current rates; the Professional plan is $39.99 a month and the Individual plan $0.99 per item sold.
Listing Deactivated for a Potential Pricing Error: What Amazon Checks, and How to Get the Offer Back
Amazon says it detects potential pricing errors based on factors including the Marketplace Fair Pricing Policy and the Featured Offer price, notifies the seller, and in more serious cases deactivates the offer. The deactivated offers and Amazon's reference prices are shown in Pricing Health.
Open Pricing Health, read the reference price for the offer, correct the price and the price inputs (sellers report the minimum and maximum price are what clear most cases), check every variation and any connected repricer, save, and let the check run again. If the reference is missing or wrong, open a case.
In Pricing Health, beside the deactivated offer. Amazon provides reference prices there to help identify potential pricing errors; some sellers report the column is empty for their product, which is a case to raise rather than a price to guess.
A price far above Amazon's reference, read as a mistake: a typo, a currency confusion, or a repricer with no ceiling. A low pricing error is the reverse, a price far below the reference, often a decimal error or a repricer with no floor.
A deactivated offer cannot be the Featured Offer at all. Since July 2026 Amazon has been removing the separate seller eligibility step, and says it chooses the Featured Offer on competitive pricing, delivery speed and performance, so the offer's price is what competes.
An Amazon Ads product announced on 29 September 2026 that shows an ad to customers who recently bought a product, inviting them to leave a star rating or review while they shop on Amazon. A rating takes one click; a written review, video or images are optional. It opens as a beta in the United States in late October 2026.
Registered sellers and vendors with an Amazon Ads account in good standing and no active policy violations, on products with fewer than 1,000 total ratings and reviews. Campaigns pause automatically when a product reaches 1,000. At launch it is available in the United States only, with more marketplaces planned.
Vine gives free units to reviewers Amazon has invited, for products with fewer than 30 reviews, and is a launch tool. Review Requests is a paid ad that reaches people who bought the product themselves, for products under 1,000 ratings, so it works after launch on listings that already sell.
No. The button in Seller Central stays free and sends one automated email per order after delivery. Review Requests reaches recent buyers with an ad on Amazon instead of in their inbox. The two work together, and the button is worth keeping on every order whether or not the ad runs.
Amazon has not published how it is billed. Its announcement says campaigns use the existing bidding and budget controls in Amazon Ads Agent, and that questions go to the Amazon Ads account team. Until the console shows the pricing, any cost per rating figure is an estimate.
On 29 September 2026 Amazon Ads said its generative creative tools now live inside the conversational experience of Amazon Ads Agent, so advertisers can make TV-quality video and display images from existing product pages. Its unBoxed page quotes up to 71% more product coverage and up to 15% sales growth for advertisers using the suite.
Yes. Amazon Ads describes Video Generator as a free tool with unlimited video generation, available in the Sponsored Brands campaign creation workflow in the United States, Canada, Mexico, the United Kingdom, Germany, France, Italy, Spain, India and Japan. The videos are made for Sponsored Brands and can also be used in Brand Stores.
Amazon says Creative Agent examines product pages, Brand Stores, websites and customer shopping signals to understand a brand, then researches, storyboards and produces ads across display, Sponsored Brands video, online video, audio and streaming TV. The better the listing and Store it reads, the better the ad it makes.
No. The creative tools make advertising assets. The main image is governed by Amazon's product image rules, which require a true photograph of the product on a pure white background filling most of the frame, and a listing that breaks them can be suppressed from search.
They replace production time on the long tail of a catalogue, not the creative decisions. The tools read what the brand has already published, so the photography, the visual rules and the claims still have to exist first, and the hero products still earn a human production path.
An Amazon Ads format announced on 29 September 2026 that lets a brand give Alexa for Shopping its own expertise: what sets its products apart, which product suits which customer, and which products go together. The assistant uses it, with Amazon's own signals, when a shopper keeps a conversation going. It launches in the United States in January 2027 and is marked as Sponsored.
Sponsored Brands prompts, generally available in the US since 25 March 2026, open a conversation that draws only on what Amazon can already read about the brand. Branded Conversations adds brand-provided expertise to that conversation, so the assistant can speak to the brand's own differentiators and recommendations rather than inferring them.
A written standard: who the brand is, who each product is for and not for, how the products go together, the claims it can substantiate and the ones it must not make, its vocabulary, and the facts shoppers ask about that the listing does not state. The same statements then need to be carried into the listings, A+ content and Brand Store so every source agrees.
Amazon's announcement does not list eligibility, but the format builds on Sponsored Brands prompts and draws on the Brand Store, both of which require Brand Registry. A brand that is not enrolled has neither a Store for the assistant to read nor a Sponsored Brands campaign for a prompt to appear under.
They should not. The format is advertising, marked as Sponsored, and the brand is responsible for what is said in its name. A claim the listing cannot support is no safer in a conversation than in a bullet point, and a brand guidelines document that lists prohibited claims is the protection.
Yes, by Amazon's own description. A Sponsored Brands prompt conversation with Alexa for Shopping draws on signals from the brand's detail pages, Brand Store and campaign data, and Creative Agent examines product pages, Brand Stores and websites to understand a brand before producing ads. Branded Conversations, launching in January 2027, adds brand-supplied expertise on top of those sources.
By need rather than by SKU: top-level pages that read like the questions shoppers ask, a comparison page for the range, a routine or system page where products go together, one plain sentence above every product grid saying who it is for, and product names that match the listings exactly.
A beta feature in the advertising console, announced by Amazon Ads on 16 January 2026, that reports renders, viewable impressions, clicks and click-through rate for each section of a Store, broken out by traffic source, in a Sectional performance tab on the Brand Store insights dashboard. Metrics run from 10 January 2026.
Amazon says the videos it produces are made for Sponsored Brands video ads and can also be used in other locations such as Brand Stores. A video earns its place on a Store page when it shows the product doing what the listing claims, not when it only moves.
Yes. Brand Registry is the eligibility rule for a Brand Store, as it is for Sponsored Brands and A+ content, and a reseller cannot build a Store for a brand it does not own on Amazon.
Yes. Amazon requires the keyword contains-synthetic-performer in the dc:subject (XMP) metadata of any image or video that contains a photorealistic AI-generated person, before it is uploaded to a listing or to A+ content, in every Amazon store. The A+ Content Manager also carries checkboxes for AI-generated content and AI-generated people.
It is the keyword Amazon asks sellers to add to an image or video file's dc:subject field in its XMP metadata, using an IPTC-compatible metadata editor, when the file shows a photorealistic AI-generated person. When Amazon detects it, it shows customers an indicator that the image includes AI-generated people.
No. Amazon's post excludes images that contain only real people even if they were altered with AI, and its forum staff said AI retouching such as background removal, colour correction and lighting is treated like conventional editing, provided the result is accurate and meets the image policies.
Amazon's forum staff said the requirement applies prospectively and existing published listings do not need retroactive disclosure, while recommending older images are updated when the listing is next edited. Keeping a register of which assets contain generated people makes that edit a short job.
No. The main image has its own rules: a true photograph of the product on a pure white background, filling most of the frame, with no text and in most categories no people. A listing whose main image breaks them can be suppressed from search, and the AI-people tag does not change that.
Partly. TikTok Shop's Product Listing Policy, updated on 1 September 2026, prohibits placeholders and digital renderings of a product, so the product itself must be photographed. AI may be used for editing, backgrounds and styling, but content that is fully or significantly created or altered by AI must be labelled, and AI may not change the product's size, colour, material, quantity, bundle contents, features, compatibility, safety attributes or performance.
Yes. TikTok's rule, effective 2 September 2026 in the US, says that if content is fully or significantly created or altered by AI you should label it, either with on-screen text or with the platform's AI-generated content toggle.
Converting live-action footage into cartoon, anime or illustration styles; editing and quality improvement; translation, dubbing and subtitles; and copywriting and script support, provided the content remains truthful and accurate.
Content can have its visibility restricted, its product links removed and Creator Health Rating points deducted. Accounts receive warnings for minor violations and posting restrictions, commission suspension or permanent bans for repeated or severe ones. Under the listing policy a seller can also lose Account Health Rating points, have listings removed, lose selling privileges and be required to refund customers.
Under the Product Listing Policy effective 1 September 2026: up to nine square images of at least 600 by 600 pixels, in colour, with no added logos, text, borders, watermarks or graphics, plus one video of up to 5MB. The main image must show the front physical view of the product on a pure white background.
Under the Product Listing Policy effective 1 September 2026, the main image must show the front physical view of the product on a pure white background and be an objective and straightforward representation of it, with no added logos, text, borders, watermarks or graphics, and no placeholders or digital renderings. Images are square, at least 600 by 600 pixels, in colour.
Five or more high-resolution images, under TikTok Shop's listing quality guidelines dated 8 July 2026. A listing can hold up to nine. Without the category-specific information TikTok asks for, a listing is capped at the Fair tier regardless of its images.
Often, with two changes. TikTok requires the front physical view, so an angled Amazon hero shot belongs in a later slot, and TikTok images are square, so the file needs a square crop rather than a stretch. The white background, no text and accuracy rules are the same on both.
Not for the product itself. The Product Listing Policy prohibits placeholders and digital renderings of a product, and the listing quality guidelines list digital renderings among the things a main image must not be. The product in every image has to be a photograph of the real item.
The Product Listing Policy names four possible actions: Account Health Rating point deductions, removal of the listing, revocation of selling privileges and refunds to customers. Unlike an Amazon suppression, points on the Account Health Rating persist after the image is fixed.
A campaign type in Amazon Ads Agent, available to all US advertisers since 29 September 2026, that runs sponsored ads, display, video and streaming TV together in one campaign. The advertiser provides products, a budget and a goal; Amazon's AI handles the channel mix, the creative, the audiences and the optimisation, and allocates spend where it drives the most growth.
Amazon says beta testers saw 67% higher long-term return on ad spend and 29% lower cost to acquire new-to-brand customers, and quotes one early tester for whom 41% of the campaign's sales were new to the brand. These are Amazon's figures from its own beta advertisers over a window it chose, not a forecast for a new account.
No. Amazon lists Sponsored Ads as a separate campaign type alongside Full-Funnel Campaigns and DVA+. A brand still runs its own cost-per-click campaigns and decides how much of its budget to delegate to the Full-Funnel campaign.
DVA+ is a display, video and audio buying experience that combines Sponsored Display, Sponsored TV and programmatic campaigns, rolling out from late October 2026. Full-Funnel Campaigns go further and include sponsored search ads in the same campaign, with the AI choosing the mix across all of them.
On a small set of products with proven conversion and a margin that can pay for a new customer, with the existing Sponsored Products campaigns on those products left running as a control, a bounded budget, and a monthly read of new-to-brand share, total advertising cost of sales and organic rank.
Display, Video, Audio+: an Amazon Ads campaign type announced on 29 September 2026 that combines Sponsored Display, Sponsored TV and the programmatic campaigns formerly run through Amazon DSP into one buying experience, with one objective, one budget and one campaign across streaming TV, online video, display and audio on Amazon properties and the open internet. It rolls out to advertisers from late October 2026.
Amazon says Amazon DSP has evolved into Amazon Ads Agent, and that existing Amazon DSP capabilities and new ones are accessible through DVA+ with advanced settings. The levers remain; they sit inside the DVA+ campaign type rather than in a separate product.
Amazon's words: for teams that want speed it is streamlined, and for those that want full control the advanced settings put every programmatic lever at their fingertips. The streamlined path is select an objective, set a budget and launch; the advanced path is the former DSP's controls.
Amazon's pages do not say. They state that sponsored ads remain cost-per-click with no minimum spend, and describe self-service and managed-service options for programmatic, without stating a floor for DVA+. Until the console shows one, assume nothing and plan a bounded test.
Only if it passes the same readiness test as DSP: a retargeting pool worth buying back, a listing that converts strangers, creative for every format in the campaign, and a way to read the result in branded search and organic sales rather than in the campaign's own return. The change is the cost of finding out, which DVA+ lowers.
A platform Meta announced on 15 September 2026 that merges creator discovery and partnership ad creation in one place: discovery filters for products and affiliate content, content-level permissions with expiry dates, one-click ad creation, recommendations by ad objective, ad-readiness editing and partnership messaging. It is rolling out globally through the rest of 2026.
Ads that run a creator's post as the brand's advertisement, with the creator named as a paid partner, so the brand can put paid reach behind content its audience trusts. The Creator Marketing Hub is where a brand now finds the creator, secures permission for the post and turns it into the ad.
Meta said live video ads, previously available on Facebook, roll out on Instagram from 29 September 2026 to general availability globally, so a creator's livestream can be amplified as a partnership ad while it runs.
Meta's announcement does not state a price for the hub; the ads themselves are bought as normal campaigns. It is separate from the paid Meta One plans Meta announced the same day, which are business subscriptions rather than advertising products.
Yes. Meta's permissions govern the use of the post, not the truth of its claims, and a partnership ad is the brand's advertisement. The brief to the creator, naming the claims it may and may not make, is the brand's protection.
Shopify's built-in tool for scheduling and testing storefront changes. In Shopify's words, a rollout is a scheduled set of changes to the online store's main theme, checkout and accounts pages, discounts, product catalogues, or a combination, found under Markets in the admin. It comes in three types: a launch, an event and an experiment.
Shopify redesigned the setup so you choose the timing and traffic percentage first and the settings reflect whether the rollout is a launch, a temporary event or an experiment. One rollout can combine changes to the theme, checkout and customer account pages, the setup shows the traffic share and its split between versions, and conflicting settings are flagged in the flow. Discounts joined Rollouts on 1 October 2026.
Yes. A Rollouts experiment compares a treatment against a control on a chosen share of traffic so you can measure the impact of a change before launching it permanently. Test one change at a time, fix the split, and set a stopping rule in orders per version before you start.
On 21 September 2026 Shopify changed how sessions are measured to give a cleaner view of real shopper activity, filtering bot and system traffic more consistently. Session-based metrics, including conversion rate, have a new baseline from that date, so figures from before and after it are not directly comparable.
The first screen of the best-selling product's page, the shipping promise and where it appears, the free shipping threshold and the checkout configuration: the things that decide a purchase and that the store controls. Page speed and signed-in checkout are repairs and settings, not experiments.
Shopify Agentic Storefronts: Your Products Are Already in the AI Channels, and Here Is What to Check
Shopify's feature for selling in AI channels. In Shopify's words, it lets customers discover and purchase your products in AI channels such as ChatGPT, Google AI Mode and Gemini, Microsoft Copilot and Meta, with products shared through Shopify Catalog (or Google Merchant Center for Google's channels). It is active by default for eligible stores and managed under Sales channels, then Agentic, in the admin.
For eligible stores, yes. Shopify says agentic storefronts are active by default, and its changelog of 8 September 2026 says products are shared with Meta by default through Shopify Catalog. Each channel can be switched off in the agentic settings, and direct checkout is a separate control.
By default: title, description, options, images, price, availability and other key attributes, for direct-to-consumer products only. Metafields, metaobjects and custom categorisation reach the channels through Shopify Catalog Mapping, so they have to be mapped to be used.
Shopify's route is to set the product's status to Unlisted, which also hides it from sitemaps, search engines and the store's own search. Blocking AI crawlers in robots.txt does not stop Shopify Catalog sending product data to activated channels; the two are separate mechanisms.
It depends on where the store's margin comes from. With direct checkout on, the sale closes inside the assistant on a Shopify-powered checkout and the shopper never visits the store; off, the assistant refers the shopper to your checkout. ChatGPT refers to the store either way. A store that lives on repeat orders and sign-ins has more reason to keep the shopper on its own pages.
Five parts, in Walmart's words: content quality (the efficacy of item name, description, key features and images), price competitiveness (your price against similar products on external marketplaces), shipping (promised delivery speed across ZIP codes, excluding items over 50 pounds or priced at ten dollars or less), published and in stock (availability over the last seven days) and ratings and reviews (the number of reviews and the average rating). Walmart does not publish the weights.
From 14 September 2026 the dashboard gives personalised recommendations to improve multiple listings at once, supporting actions Walmart names as automating pricing, restocking inventory and enhancing content with GenAI.
Content quality, because it is the only part the seller controls completely and immediately: the item name, description, key features and images as attributes in Walmart's item spec. Then availability, then shipping speed, then price where the margin allows, and reviews last because they move slowest.
Yes. Walmart's guide says quality listings can unlock growth opportunities such as the Pro Seller badge, and the Pro Seller tiers each carry a listing quality threshold. The score is one of the gates to the tier.
Through Walmart's Recognized Reviewer programme, which from 25 September 2026 accepts items with fewer than 15 reviews. It is the one direct route to the review count the score measures; everything else about that part of the score follows from sales and product quality over time.
Google's Search Status Dashboard lists the September 2026 spam update as starting on 24 September 2026. The previous one, the August 2026 spam update, started on 18 August and ran for two days and sixteen hours. Google has published the name and start date and nothing about which techniques this update targets.
In Google's words, an improvement to SpamBrain, its AI-based spam-prevention system, to make it better at spotting spam and catching new types of spam. Sites that violate Google's spam policies may rank lower or not appear at all; a site that complies should see nothing.
Scaled content abuse (pages generated at scale for rankings), thin or scraped content (supplier copy and imported reviews), doorway pages (location, size or colour pages that funnel to one page), site reputation abuse (hosting a partner's content for the traffic), expired domain abuse, keyword stuffing, user-generated spam in reviews and Q&A, and link spam from a past campaign.
Google says changes to comply may help a site improve if its automated systems learn over a period of months that the site complies with the spam policies. For link spam, Google says any ranking benefit the links generated is lost and cannot be regained; the position has to be earned again.
Google's Search Status Dashboard, read on 2 October 2026, listed the September 2026 spam update and the August 2026 spam update for the period and no core update for September. Claims of a September core update in the trade press were not on Google's dashboard.
TikTok Shop's listing quality guidelines, dated 8 July 2026, give the Good tier a title of between 25 and 255 characters that is concise, accurate and free of unnecessary content, with no word or phrase repeated four or more times, no references to variants, colours, sizes or stock counts, and no clickbait terms.
The primary phrase goes in the title once, with the qualifiers that decide the purchase. Question-type phrases go in the description as sentences. Material, size, scope and benefits go in the category-specific attribute fields, which also decide whether a listing can reach the Good tier. Colour and size go in the variants, not the title.
As a starting list, with care. Amazon-style titles that repeat the head term and list variants break TikTok's repetition and variant rules and grade as Fair. The phrases TikTok shoppers use are found on TikTok: in the search suggestions, the Shop tab, the captions of selling videos and the comments under them.
For the Good tier the guidelines ask for a description of more than 80 characters or at least one image, relevant to the product and meaningful, with no broken hyperlinks, clickbait or duplicate content.
Often because the category-specific information is missing. TikTok's guidelines say a listing without the required category details (how to use, materials, specifications, sizing, model display images and so on, varying by category) is capped at Fair regardless of its title, description or images.
Six items, in Amazon's words: identity, bank account information, a working business phone number, a business address, a tax identification number and a working email address. The Act applies to sellers who have sold 200 or more new or unused consumer products and had $5,000 or more in gross revenues in any twelve-month period during the previous twenty-four months.
No. Amazon says that every year it will ask sellers to electronically certify that the six pieces of information are still accurate, and verification is also triggered when any of the items changes. An expired identity document or a changed address fails the next certification.
Usually one of three things an Amazon moderator named on the forums: the documents do not match the Seller Central entries exactly (an abbreviation such as St. against Street is enough), the form was saved but not submitted for verification, or one item (often identity) is still being processed after another (often the bank account) has cleared. Sellers report the status can take up to 24 hours to update, and identity 48 to 72 hours after a bank check.
Amazon's statement is that it may temporarily deactivate the selling account and hold disbursements if the requested information is not provided by the stated deadlines. The warning in Account Health gives the date.
Amazon points sellers to the Account Health page or the Account Information page. The six items are edited in their own settings (identity, deposit and tax, business information) and each change has to be submitted for verification, not only saved.
A video call with an Amazon associate in which the registered owner shows their identity document, confirms who else uses the account and explains where the inventory comes from. Amazon's public page says identity can be verified by a photograph with the ID or by video call, usually within three business days; after a deactivation, sellers report the notice gives fourteen days to complete the call.
Amazon publishes no rule on this outside Seller Central. Sellers who failed a call in 2026 report being told that no further appointments were available and that the reason could not be shared. The only route with any record of being read is an appeal that supplies what the call could not, with a request for a new interview and, where language was the problem, an interpreter.
From sellers' reports, three ways: a connection that drops or a call that cannot be completed cleanly; documents that are not to hand, expired, or not in the registered business's name; and answers about users, sourcing or addresses that contradict what the account already records.
The identity document on the account, in date; the company registration and proof of address matching the account exactly; the bank statement for the account on file; supplier invoices for recent months in the business's registered name with complete supplier details; and a note of everybody with access to the account.
Not necessarily. One seller reported an Account Health Specialist describing it as generic template language included in certain notifications, and sellers with long clean records receive the same wording. The practical reading is that the notice is a request for a verification conversation, and the conversation decides the account.
Under Amazon's windows effective 23 October 2024: 60 days after an item is reported lost or damaged in a fulfilment centre; between 60 and 120 days after the refund or replacement date for a customer return that never arrived, with a 60-day minimum wait; between 15 and 75 days from shipment creation for a removal lost in transit; and 60 days after delivery back for other removal claims.
Since 1 November 2024 Amazon says it proactively reimburses FBA items lost in its fulfilment centres, and most warehouse loss, damage and customer return claims are processed automatically. Anything not reimbursed automatically still needs a manual claim inside the window, which is why the ledger has to be reconciled against the reimbursement report.
Not since 1 January 2026 in the US store. Amazon ended prep and item labelling services on that date and, in its staff's words, shipments created after it that arrive without proper prep and item labelling are not eligible for reimbursement if damaged or untraceable. Press reports put the same change in Canada on 1 July 2026.
Because the 60-day windows from October and November losses close in December and January, when the team is shipping and counting rather than reconciling, and a quarterly reconciliation finds the losses after the windows have shut. The fix is a monthly reconciliation that runs through peak and a fixed audit date in the first fortnight of January.
Enter a sourcing cost for every ASIN, because lost and damaged stock is reimbursed at manufacturing cost and Amazon applies its own estimate where none is entered. And keep a complete file for every shipment: packing list by SKU, box contents, carrier proof of delivery, photographs of labelled units and the invoice.
Under TikTok Shop's Violation and Appeal Guide (updated 29 May 2026), a first appeal must be filed within 30 days of the notification, a second appeal within 15 days of a rejection, with a maximum of two appeals per case. Where enforcement is not yet immediate, the window is 12 to 72 hours.
In Seller Center: Account Health, then Violation Records, then View Details, then Appeal. TikTok's earlier guide names the same route as Shop Health, Violation Records, View and Appeal. The status shows as Under Review once filed.
TikTok's list: the business licence and government ID photographed front and back, proof of warehouse ownership (lease, property documents or utility bills), purchase contracts and invoices matching the business details, and a video of the account holder holding the ID and licence with their face clearly visible; shipping records and warehouse or qualification videos are optional. Documents must be complete, original and legible, and an appeal without a letter is not processed.
Under the Prohibited Products Policy of 29 September 2026, for using disclaimers, altered spelling, incomplete images, misleading categories or other methods that conceal a prohibited product; under the Product Listing Policy of 1 September 2026, for images or claims that do not match what the customer receives, including renderings, placeholders, text on images or a main image that is not a front view on white.
The enforcement stands. TikTok's guide allows a maximum of two appeals per case and says its decisions are at its sole discretion and final once the appeal is concluded. The listing can be fixed and the product relisted only where the policy allows it; relisting a removed product to get around a record is treated as concealment.
A set of tools in Seller Central, announced at Amazon Accelerate 2026, that connect a seller's eBay, Shopify, TikTok and Walmart accounts and bring their orders, listings and profitability into one place, with more channels to come. Amazon says they are available at no additional cost and are rolling out gradually to sellers in the US store over the coming months.
At launch, according to Amazon: connect external channel accounts to Seller Central, import and link off-Amazon listings, view multichannel sales and orders, and fulfil orders across channels in a few steps. Listing synchronisation, AI-generated channel-ready listings, unified fulfilment and cross-channel profitability analytics are described as coming later.
Amazon says the tools are available at no additional cost to sellers. Fulfilling other channels' orders through Amazon's fulfilment network carries its own fees, which are separate from the tools.
They will format a description for each channel's requirements and, later, generate channel-ready titles and details from Amazon's product data. They do not apply Walmart's listing quality scoring, TikTok Shop's tier guidelines or eBay's item specifics, which decide how a listing performs on each channel.
Consolidate in the tools when the brand already runs each channel competently and the problem is reporting and order handling. Partner when nobody in the business knows a channel's performance thresholds or can act on its notices. The tools make the tooling free; they do not supply the channel knowledge.
An expansion dashboard across the US, Europe and Japan with more countries by year end; one listing evaluated against every destination with translation, currency and unit conversion and compliance flags; a single compliance test covering several countries, starting with toys; Global Warehousing and Distribution expanding to seven more countries by year end; and supply chain visibility with a single reconciled unit count, plus Seller Assistant agents for inventory and inbound planning.
Amazon's figure is that sellers active in more than one country earn 70% more revenue on average, and that only 30% of its sellers currently list in more than one country. Both are Amazon's numbers from its announcement.
Not yet. Amazon says it is rolling out by category: toys now, electronics and baby products by the end of 2026, and additional categories in 2027. Pilot participants reported savings of up to 60% on compliance costs.
Amazon says GWD currently serves the US and expands by the end of 2026 to the United Kingdom, Japan, Germany, France, Italy, Spain and Canada, with storage costs up to 45% lower than US Amazon Warehousing and Distribution and replenishment five days faster than current options.
They do not decide whether the product's economics work in that market, register the brand for VAT or act as its importer, protect the brand under that country's trademark system, write a listing that sells to local shoppers rather than a translated one, or run the market's advertising, customer service and returns. Those are the brand's or a partner's.
Agency Spotter says its team verifies the identity and credibility of each reviewer once a review is submitted, and keeps published only the reviews that meet its criteria. Only current or former clients may review an agency; former employees and contractors may not. Agencies also ask clients to post reviews, so verification confirms the reviewer is a real client without showing how the agency chose whom to ask. Read each review's role, company and date rather than the average.
That the agency has claimed its listing. Agency Spotter's pricing page lists the badge on every plan, the free one included, and describes it as claimed by the agency on the free plan and claimed and verified by the agency on the paid plans. It tells you the profile is maintained by the agency itself. It says nothing about the quality of the work, and it is separate from the verification of individual reviews.
Its FAQ says it ranks agencies using a combination of objective scoring and expert curation, and each report names its inputs: verified client reviews, agency credentials and case studies for eCommerce development, for example. Because case studies and curation count, a report is not a sort by star rating. Each report also carries a published date and a service category, and both are worth checking against the work you are hiring for.
It can pay for placement. Agency Spotter's FAQ says agencies can pay to gain more exposure to a point, and its pricing page shows the paid plans buying a place in promoted search results, each tier above the one below, with more promoted services and more visible projects. Read the order of a search as partly a plan, and judge each agency on its projects and its reviews.
If you are signed in, it can. Agency Spotter tells agencies they can see who has been checking them out, and its paid plans include data on leads who interacted with a portfolio and an allowance of emails to contact them. It says it ranks every registered user to help agencies identify a verified lead, so if you want to research before being contacted, read profiles before you register.