"Is this agency any good?" That is the question a seller types into a search bar or an AI after a sales call. It is the wrong question, because nobody on the internet can answer it for you, least of all a competitor.
The right question is what the agency publishes about itself, and what happens when you put the same eight questions to it and to me. A service list tells you what they sell. The eight questions tell you who will own your account, what it costs every month, and what you keep on the day you leave.
So the rule for every page in this cluster is one line. We quote them from their own site, we answer the same questions ourselves, and you run the test on both of us. If we lose on your version of that test, hire the agency that won it.
The numbers behind this guide
Six standing figures sit under this cluster, and every one of them is either published on our site or dated here.
| Claim | Figure | Captured |
|---|---|---|
| Agencies compared | 305, one page each, from a list of 477 Amazon agencies | 2026-09-05 |
| What each page says about the agency | Only what its own website states, quoted and dated | Standing rule, SPEC-Agencies §3 |
| Operators running our brands | 50, about 70 brands by hand, about 1.4 per operator | 2026-09-04 |
| Share of our work subcontracted | None. Sourcing in Guangzhou, creative in Dubai, software in-house | Standing term, SPEC §4 |
| Managed pricing, every service included | $800 a month for one product to $2,400 for five | Standing term, SPEC §4 |
| Brands built and launched through our FBA service, public on flapen.com/results | 9 | 2026-09-05 |
Every row is a question you can put to any agency on the list, and the comparison pages put it for you.
How to read a comparison page
Each page follows one shape. First, what the agency says it offers, taken from its own site on a stated date. Second, what Flapen offers against the same eight questions: who does the work and where, brands per account manager, whether they launch from zero, sourcing and creative, advertising, technology, the pricing model, and the contract and exit terms.
Third, a table with both columns side by side and the capture date underneath it. Where an agency does not publish an answer, the cell says so, because a verified absence is a fact and a guess is not.
Fourth, the test. Six weighted questions to send in writing to everyone on your shortlist, us included. The pages never grade the agency, they hand you the scorecard and let the written answers do the grading.
Our own eight answers never move from one page to the next, and they sit in full on Amazon brand management. Read them once before you open a comparison, then read every page in this cluster against them. A page that concludes "hire Flapen" would be worth nothing to you, so none of them does.
Full-service generalists
Most of the agencies on the list describe themselves as full service. The phrase covers two different businesses. One puts a named person in charge of your whole account, and one delivers a service list through specialists you have not met.
The page for each generalist quotes the scope its own site publishes and puts the ownership question next to it. Ask how many brands the assigned manager carries. The pauses before the answers are most of the evaluation.
Our number is published for that reason: 50 operators run about 70 brands by hand, which is about 1.4 brands each. Nothing is subcontracted, so the person who answers your message is the person doing the work. Ask every generalist on your shortlist for both figures in writing.
Read the scope list last rather than first. Two agencies can publish the same twelve services and deliver them through different org charts, and the org chart is the thing you are buying. A generalist that names its studios and its staffing ratio has answered the two questions that constrain everything else.
Agencies outside the United States
Nearly half the list sits outside the United States: the United Kingdom, Germany, Canada, Spain, India, and the Gulf. A local agency can operate a marketplace in its own language and under its own tax rules. The page for each one quotes the marketplaces its site names.
We run all 23 Amazon marketplaces from Abu Dhabi, and we produce content in English, German, Spanish, and French. Sourcing and quality control sit in Guangzhou, and the creative studio sits in Dubai. Neither one is a partner network, and both are on payroll.
Distance is not the risk here. The risk is the handoff you cannot see, where the manager sits in one country and the work happens in another. Ask where each function physically sits, then ask who employs the people doing it.
Currency, tax registration, and language support are three things a local agency holds that a remote one cannot fake. Weigh all three against caseload before you weigh any of them against price.
Advertising specialists
An advertising specialist runs one function, usually Sponsored Products and Sponsored Brands, and prices it against your ad spend. That is a fair hire when you already own the listing, the catalog, and the inventory decisions.
It is the wrong hire when conversion is the problem, because the campaign buys the click and the page decides what that click costs you. The pages for the specialists state what each one publishes about its advertising work. They leave the listing question open for you to ask.
Ask what the target advertising cost of sale should be at launch and what it should be at maturity. An operator gives you two different numbers and the reason each one moves. One target held across every product stage is maintenance dressed as management.
Percentage of ad spend is the pricing model to refuse outright. It pays a provider to grow a budget rather than to defend a margin. Ours is a flat fee with advertising included at every tier, so nobody here earns more when your spend climbs.
Vendor Central and retail brokers
Some agencies on the list grew out of retail brokerage and wholesale. Their sites name Vendor Central, consumer packaged goods, and the retailers beyond Amazon. That is a different business from running a private label brand in Seller Central.
If you sell to Amazon rather than through it, that is a real fit, and the page says so. If your brand sits on a supermarket shelf as well as in a search result, the same holds. The page then names the questions to ask about the Seller Central side of your revenue.
The two models pull in opposite directions on margin. A broker optimizes the terms of the trade, and an operator optimizes the unit economics of the listing you own. Decide which of those two problems is costing you more this quarter.
Purchase orders, chargebacks, and allowances have no equivalent inside a Seller Central account. If those three words describe your week, weigh a broker seriously, then ask who would run the private label half.
Software with a service attached
A few companies on the list sell a tool first and management second. The page for each one quotes what the software says it does and what the service adds on top of it. Read the pricing line twice, because a tool fee and a management fee are often quoted as one number.
Our own position is published. The platform is the data layer 15,000 sellers a month use to find products, and the operators run accounts on tools our tech team built. The agents that take actions ship next, and no client account is run by one today.
The question to ask any hybrid is who is accountable when the dashboard and the operator disagree. A named person with the authority to overrule the tool is the answer you want. A dashboard is not accountable for a lost quarter.
Ask also what happens to your data and your campaign history when the subscription ends. Ours leaves with you, along with the Seller Central account and every creative file. Get that answer in writing before the first invoice, never during the handover.
What most agencies will not tell you
Four things stay out of the pitch, and on a bad day that includes ours.
- A comparison page is not evidence. Every page in this cluster is written by one agency about another. That is why each one quotes only the agency's own words, carries a scorecard instead of a verdict, and invites a correction.
- Full service is a delivery model, not a service list. Two agencies can publish the same list, one delivered by employees and one by a network of partners you will never meet. The org chart tells you what the list cannot.
- Attention per account is the number that moves revenue. It is almost never in the proposal, and it is never in the service list. Ask what the ratio becomes after the next ten clients sign.
- Nobody wants to stop a product. A monthly fee earns the same whether a product thrives or limps. Ask what would make them tell you to stop, and hold us to our four signals: rating trend, return rate, conversion rate, and cost of customer acquisition trajectory.
Every page in this cluster carries that disclosure in its own words. It is there because the seller who reads it asks us the same four questions one call later. A written answer settles them faster than a spoken one.
If we cannot name the person who would run your account and what they would stop, keep looking.
Do this week
One thing to do this week, at no cost. Take the six weighted questions from any page here and send them, in writing, to every agency on your shortlist, including the one you already pay. Give them one working day to reply.
Score the written answers and not the calls. Full weight for a specific answer, half weight for a general one, and zero for a refusal. The order that falls out of that page is your decision, and it cost you nothing.
To get those six answered for your own account, request the free written audit, back inside 48 hours at no charge from Flapen.
