A directory is only as useful as the thing it sorts by.
GoodFirms sorts by more than the star rating, which is what makes its listings worth reading and also what makes them easy to misread: a company near the top of a category is not simply the one with the best reviews.
In its own words, from its announcement of the 2026 rankings on 16 June 2026, GoodFirms "applies a multi-factor evaluation methodology that combines qualitative and quantitative signals." Five criteria are named. Each is testing something a buyer actually cares about, and each can be checked on the listing in front of you.
GoodFirms ranks companies by several criteria rather than reviews alone. In its own words they are verified client reviews collected through the platform, the depth and breadth of services in the category, portfolio quality and evidence of relevant work, market presence and industry recognition, and business maturity and operational reliability.
So a higher position means a stronger profile across those five, not necessarily better work for your project. The useful reading is in the individual reviews: whether each describes a real engagement in specifics, with its scope, its length and what changed, and whether the projects resemble yours. A directory beats a search engine for comparing agencies side by side on the same terms; it cannot tell you how a team works day to day, how it handles a problem, or whether it will be a fit, which are the questions to take to a call.
Read that way, a listing narrows a long list to a short one, which is what it is good for, and leaves the final choice to the conversations.
A practical rule: shortlist from the reviews, not the rank, and ask each agency on the shortlist for a client you can speak to whose project looked like yours.
"Verified client reviews collected directly through the platform." Collected directly matters: it means the review went through the platform's own process rather than being copied from a website. When you read one, read it for the specifics of the engagement, not the adjectives.
"Depth and breadth of service offerings within the category." This is a two edged test and the listing does not tell you which edge you are on. Breadth is valuable when your problem crosses channels, because one team that runs the advertising, the listings and the fulfilment settings is one conversation rather than three. It is a liability when a company lists forty services and staffs six of them. The check is to ask who does the work you are buying, and what else they do that week.
"Portfolio quality and evidence of relevant project delivery." Relevant is the word to hold on to. Read the portfolio for your channel and your category, not for the logos. A team that has grown a supplements brand on Amazon has met the compliance problems a supplements brand has; a team that has grown a furniture brand has met the
size band and storage economics a furniture brand has. Both are experienced. Only one has your experience.
"Market presence and industry recognition." This is the softest of the five and the easiest to buy, so weigh it last. It is genuinely useful for one thing: continuity. A company with several years of presence is less likely to disappear halfway through your fourth quarter.
"Business maturity and operational reliability." This is the one that separates an agency from a group of freelancers with a website. It is also the one a listing cannot fully prove, so it is the one to test on a call: ask what happens when the person running your account is ill, and what happened the last time it occurred.
Business directories earn money in two ways that sit side by side on the same page: a ranking produced by their research, and positions a company pays for. Both are legitimate. Reading one as the other is the mistake.
What payment usually buys. Most directories sell featured or sponsored placements: a position at the top of a category, a highlighted card, a place in a newsletter. They are labelled as such, often with a small word near the company's name.
What payment should not buy. The research ranking and the reviews. A directory whose methodology is published, and whose reviews are collected through its own process, is telling the reader that its order reflects the criteria rather than the budget. The sponsored slots sit beside that order, not inside it.
How to read a category page. Look for the label on each of the first few companies. Skip past the sponsored ones to where the research ranking begins, and read both groups by the same criteria: the reviews, the portfolio, the breadth of service.
A sponsored listing is not a warning sign. A company paying for visibility is marketing itself, which most good companies do. It simply means its position on the page says nothing about how it ranked, so the rest of its profile has to make the case.
Search puts the companies best at being found at the top. A directory with a research methodology puts the companies best at being described at the top, which is a different bias and a more useful one when you are shortlisting rather than browsing.
The practical use is to build the shortlist from the directory and then leave it. Take four companies, then go and read what they have written: their case studies, their writing about the platforms, the specificity of it. An agency that cannot explain
how the Featured Offer is won in public is unlikely to explain it to you in a meeting.
Reviews on a directory are longer than the ones on a consumer platform, because the form asks for more. Use that. The first thing to read is not the rating but the engagement: what was bought, over what period, at what size, and what changed. A review that says the communication was great, and nothing else, is a real review of nothing in particular.
The second is whose problem it describes. A brand selling twelve products through one channel and a brand selling four hundred across five are different companies with different failure modes, and the agency that suits one can be wrong for the other. You are looking for a reviewer whose sentences sound like your business.
The third is the dates. An agency's capability is a team, and teams change. A set of glowing reviews that all stop two years ago is a story about the agency's past. Recent reviews from long engagements are the strongest thing on any profile, because they mean somebody stayed.
And read what the company writes back. A reply that answers the specific point tells you what a monthly account review will feel like; a reply that thanks everybody equally tells you the same thing about a template. It is the test that works on
a marketplace's own feedback too: the reply is the evidence, not the rating.
A directory profile is one view of a company, written partly by the company. The shortlist gets stronger when that view is checked against a few others that the company controls less.
Open review platforms. A profile on a consumer style review platform, where anybody can write and the company cannot remove what it dislikes, shows a different side. Read the most recent reviews and the negative ones with their replies.
Reading a TrustScore for a service business sets out how.
The people. The team named on the company's own site and on professional networks: whether the people who would run the account exist, how long they have been there, and what they did before. A team page of stock photographs is a question worth asking.
The work. Case studies that name the client, with the client's agreement, and show the numbers before and after, are evidence a reader can test. Case studies with no client, no dates and only percentages are claims.
The company itself. How long it has traded, where it is registered, where its people work. None of this proves quality, but each answer that matches the profile adds weight, and each that does not is worth a question on the call.
Where the directory, the open reviews, the people and the work all tell the same story, the shortlist is on firm ground.
Who runs my account, and what else do they run? Ask for the person, not the agency, and ask how many accounts that person holds.
What does month eight look like? Launches are pitched; the steady state is what you buy. On marketplaces month eight is the
weekly advertising routine, the inventory arithmetic and the listings that drifted, which is unglamorous and is where the money is.
What do you do when it breaks? Suppressions, suspensions and stranded stock are not exceptional; they are Tuesday. The answer you want has names, hours and a first step in it.
A reference call with a current client is the closest a buyer gets to the evidence a review cannot give. It is worth preparing, because a reference asked whether they are happy will say yes.
Ask about the work, not the feeling. What did the agency actually do each week? Who did it? How did the reporting look, and did it change anything the brand decided?
Ask about a problem. What went wrong at some point, and what did the agency do about it? Every engagement has a bad month. A reference who cannot name one either has not been a client long or was chosen for being uncritical; one who describes a problem handled well is the best evidence on offer.
Ask about the people. Is the person running the account the one they met in the sales process? Has it changed, and how often? Continuity of the people is one of the strongest predictors of how an engagement goes.
Ask about the numbers, roughly. Without asking for confidential figures, ask whether the results matched what was expected at the start, and how long it took to see them.
Ask what they would change. The answer often says more about the agency's weak points than anything on its profile. Weigh it against the needs of your own account, not theirs.
Our listing is at
goodfirms.co/company/xcelerate-brands. Read it the way this article describes: the reviews for the engagement details, the portfolio for your channel, the breadth for whether it is staffed, and then take the three questions above to a call. If
marketplace management is what you are weighing up, that call will settle more in twenty minutes than an afternoon of listings.
And if you are still assembling the shortlist,
the guide to choosing an agency covers the parts a directory leaves out: the contract, the access, the reporting, and the questions that separate an agency that manages from one that reports.