How to Increase Sales on Amazon: The Twelve Levers, in the Order to Pull Them

Every answer to this question is a list, and the lists are all the same: better images, more keywords, run ads, get reviews, lower the price. They are not wrong. They are unordered, and the order is where the money is, because a lever pulled before the one it depends on costs the same and moves nothing. Advertising a listing that does not convert buys traffic to a page that loses it.

Sales on Amazon are three numbers multiplied: the sessions a listing gets, the share of them that buy, and whether there was stock to sell them. This is the twelve levers grouped by which number they move, in the order we pull them on an account we take over.

Increase sales on Amazon by fixing conversion before buying traffic. The first four levers are the main image, the title and bullets, the price against the page's other offers, and reviews: a listing that converts better turns every existing visit into more sales without paying for a single click. Only then come the five traffic levers, advertising among them at seventh of twelve, followed by the three that keep the sales already made, stock, the Buy Box and account health. The business report tells you which problem you have: few sessions with a good conversion rate is a traffic problem, many sessions with a poor rate is the listing or the price, and both low means the listing first. The sections below take each lever in that order, with a worked example.

The order matters because each lever multiplies the ones before it: money spent bringing visitors to a page that converts poorly buys the same poor conversion at the price of a click.

One, the main image and the image set. The main image decides the click on the results page and the second image decides the read on the product page, and both are fixed before anything else is touched, to the rules in the image requirements and beyond them to what makes a buyer stop. Two, the title and bullets, rewritten for the buyer's question rather than the keyword count: what it is, who it is for, what is in the box, in the first line. The listing optimisation article is the method.

Three, the price, set against the unit economics and the competing offers rather than against a wish; a product priced ten percent above the page's next offer with no visible reason converts at a fraction of its rate. Four, reviews: the count and the recent ones, earned through the Request a Review button and the Vine programme for a new product, and protected by fixing what the bad ones say. A listing with these four right converts, and only then is traffic worth paying for.

Five, indexing: the listing has to be found for the terms that matter, which means the terms in the title, bullets and backend search terms, checked by searching for them, because a product that is not indexed for its own category term is invisible whatever the ads do. Six, organic rank, which follows sales velocity and conversion on the term, so it is a result of the first four levers and this one, not a separate campaign. Seven, sponsored products on the product's own terms, structured so that the search terms report can be read and acted on, which is the routine in the weekly PPC article.

Eight, the brand's own surfaces: A+ content, the brand store and sponsored brands video, which bring traffic from the brand's other products and from the video placements on the results page. Nine, deals and coupons, used for velocity on a term the product is close to ranking for, rather than as a permanent discount. Advertising is lever seven of twelve, and it is the one most sellers pull first. Pulled first, it measures the conversion problem at the price of a click each.

Ten, stock: every day out of stock resets velocity and rank, and the forecast is the lever, covered in the inventory management article. A product that sells out in the third week of a campaign hands its rank to the next offer on the page and pays for the climb twice. Eleven, the buy box: for a listing with more than one seller, the offer that holds the buy box takes most of the sales, and holding it is price, fulfilment method and account metrics together; a brand that has let resellers onto its own listing is competing for its own sales. Twelve, account health: a suppressed listing or a deactivated account makes every other lever irrelevant, and the routine that keeps the account clean is a sales lever in the only sense that counts.

These three do not add sales; they keep the ones the first nine made. They are last in the order because they are checked rather than built, and first in the calendar because a check takes an hour and a rebuild takes a month.

The twelve levers are the same as they were, but several of them work differently after changes Amazon made this year. A plan written before them needs adjusting.

Conversion: the title is shorter. Since 27 July 2026 titles outside media are limited to 75 characters, and the attributes that used to fill the end of a long title now belong in Item Highlights, a separate 125 character field that is searchable and shown with the title. Amazon's 75 character title limit and Item Highlights covers what to put where.

Traffic and price: more offers compete for the button. From July 2026 Amazon stopped gating sellers out of the Featured Offer on account performance, so more offers on a shared page are in the running, and performance counts against an offer in the ranking instead. How the buy box is won covers what is weighed.

Stock: running thin costs more. Amazon's low inventory level fee has been judged for each variation since January 2026, so one size running short can be charged while the family looks well stocked. Stock levers now carry a fee as well as lost sales.

Account: more checks on who sells. Annual verification and third party testing in more categories mean the account levers include paperwork that has deadlines.

None of these changes the order of the levers. They change the details of pulling several of them.

A kitchen product with 3,000 sessions a month, a unit session percentage of 4%, and an advertising budget that is half the margin. The seller's plan was more budget. The report said conversion: 4% on a page with fifteen offers and a main image that showed the box. The main image was reshot, the second image became the product in use with the one line the reviews kept repeating, the price moved to the page's median, and the Request a Review button was pressed on every order for a month. Conversion went to 9% on the same 3,000 sessions, which is the same as doubling the traffic for nothing. The advertising, unchanged, then paid at twice the rate, and the budget increase the seller wanted in month one was made in month three against a listing that could use it.

A lever pulled without a target is a change made in hope. Each group of levers has a number it should move, and setting that number before the work starts is how a seller knows when to move to the next lever.

Conversion levers: unit session percentage. The share of sessions that become orders, read for each product before and after the change, over the same number of weeks with price and advertising unchanged. Compare with similar products in the category, not with an account average.

Traffic levers: sessions, and where they come from. Total sessions for the product, and the split between advertising and organic. Advertising that grows sessions and is followed by organic sessions growing is building rank; advertising that only adds its own sessions is renting them.

Stock levers: days in stock. The share of days each product was available to buy, in each region where it sells. A product in stock nine days in ten is losing a tenth of its potential before any other lever matters.

Account levers: the rating and the open items. The Account Health Rating and the number of open policy items, read monthly.

One target at a time. Work on one lever per product until its number moves, then the next. Changing three at once means nobody knows which worked, and the next product cannot learn from this one.

The business report gives sessions and unit session percentage by product, and those two numbers say which group of levers to pull. Sessions low and conversion high is a traffic problem: levers five to nine. Sessions high and conversion low is a listing or price problem: levers one to four. Both low is the first group first, always. Sales that spike and stop is stock or the buy box. The search term report says which terms the product is seen for and which it converts on, and the gap between the two is the next lever.

Growing sales is a plan; a sudden fall is an emergency, and it has its own order. Most sudden falls have one of a handful of causes, and checking them in this order finds the cause fastest.

Stock. Is the product in stock, in every region, and not stuck in transfer between fulfilment centres? An inventory page with plenty of units can hide stock that cannot be sold where the shoppers are.

Suppression. Has the listing been removed from search, for an image, a missing field or a price? Search suppressed listings shows where the reason is.

The Featured Offer. Has the seller's share of the button dropped, to a competitor or to nobody? A fall in share is a fall in sales on a shared page within days.

Price. Has a competitor cut, or has the seller's own price moved, through a repricer, a coupon ending or a deal finishing?

Reviews and badges. Has a run of poor reviews arrived, or a frequently returned label appeared? The frequently returned badge covers the second.

Advertising. Has a budget run out, a campaign paused or a bid changed? Advertising that stopped takes its sessions with it.

The season. Only after these, check whether the same weeks fell last year. A seasonal dip is the last explanation to accept, not the first.

For the brands we run under Amazon marketplace management, the first month on an account is levers one to four and ten, with advertising held at maintenance, and the second month is the traffic. It looks slow from outside. It is the order that makes the advertising in month two pay, and it is why the article on how SEO and PPC feed each other is about sequence rather than either one alone. The twelve levers are a list; the order is the strategy.