4.1/5 on Trustpilot The free audit is yours to keep — hire me or not
BUYER'S GUIDE

How to vet an Amazon agency

Five questions do most of the work: what number they optimise against, whether they can derive your break-even ACOS, whether they keep a change log, who actually runs the account, and what is explicitly out of scope. The specificity of the answers tells you more than the answers.

The five questions that do the work

Most vetting advice is a long list. These five separate operators from salespeople, and you can ask all of them on one call.

  1. "What number do you optimise against?" If the answer is ACOS rather than net profit, they are optimising a diagnostic. ACOS falls when you starve volume — that is not a win.
  2. "Can you work out my break-even ACOS?" It is arithmetic from your margins, and it is different per product. If they cannot, or they quote a target from memory, they are working from a number that has nothing to do with your business. You can check yours with our calculator.
  3. "Show me a change log." Dates, changes, reasons — client details removed. This is the single most revealing request you can make, because it cannot be faked in the room and most agencies do not keep one.
  4. "Who runs my account day to day?" Being sold by a senior and managed by a junior is the most common disappointment in this industry. Ask how many accounts that person holds.
  5. "What is explicitly out of scope?" A crisp answer signals a defined service. Vagueness here becomes a dispute in month three.

How do you read an agency case study?

Four things separate a real case study from decoration:

  • A starting point. "ACOS 45% to 22%" means something. "ACOS down 50%" without a baseline means nothing.
  • A timeframe. Over what period, and was it comparable — or is a Q4 lift being presented as their achievement?
  • Named caveats. What else changed in that window: a price change, a launch, a listing rewrite, seasonality. Their absence is what makes most competitor case studies read as fiction.
  • Honest attribution. Ad-attributed revenue is not incremental revenue. Some of those sales would have happened anyway, and a good agency says so unprompted.

If a case study has a large number and none of the four, treat the number as unverified.

What about certifications and badges?

Verify them at the source rather than on the agency's own page. A partner badge is an image file; anyone can put one on a website. Check the issuing directory lists them.

Also separate two different things that look alike: being listed in a partner directory, and holding a badge. They have different requirements, and the distinction is quietly blurred a lot.

What access should an agency ask for?

The minimum that lets them do the job — and never your password.

  • Scoped user invitations to Seller Central and the Ads console, not shared credentials.
  • Read-only is enough for an audit. Wider access should come with the engagement, not before it.
  • Ad spend stays on your own card. An agency routing your budget through their account is a real risk to you.
  • They should never need your banking or tax details.

An agency that asks for account-owner credentials is telling you how they treat every other client's account too.

What this guide cannot tell you

It cannot tell you whether a particular agency is good. Every question here can be answered well by someone who then executes poorly, and references are self-selected by definition.

What it can do is filter out the ones who cannot answer at all — which is most of the ones that would have wasted your quarter. After that, the only real test is a small, defined piece of work with a deadline you both agreed.

These questions apply to us too. If we cannot answer them on a call, do not hire us either.

Where to go next

Questions people actually ask

Ask what number they optimise against, whether they can derive your break-even ACOS from your margins, to see a real change log with client details removed, who runs the account day to day, and what is explicitly out of scope. The specificity of the answers matters more than the answers themselves.
Only when they show the starting numbers, the timeframe, what else changed in the window, and label ad-attributed revenue as attributed rather than as total business impact. A case study with a big percentage and no baseline is decoration.
No. Nobody controls Amazon, your competitors, or your stock position. A guarantee is either meaningless in the contract's small print or a sign they will chase the guaranteed metric at the expense of your profit.
Look for a named team with real profiles, a registered entity and address, and third-party reviews they did not write — Trustpilot, Clutch, Google. Certifications should be verifiable in the issuing directory, not just a badge image on a page.
Manufactured urgency, results promised before seeing the account, a fee tied to ad spend with no explanation of how they manage that incentive, reporting built around impressions and clicks, and any reluctance to put scope in writing.
Only after you understand what is missing at that price. A low fee usually means fewer optimisation passes, a more junior operator, or a template applied across many accounts. Ask who does the work and how often they touch the account.

Use these questions on us first.

Read-only access, three days, a real report with real numbers — no obligation. If we can't find enough to justify our fee, we'll be the ones to say so.

YOU WALK AWAY WITH
Your wasted ad spend, found and sized in dollars
Your break-even ACOS, worked out per ASIN
A prioritized fix list — yours to keep, no strings

No spam. No 12-field forms. Just a real reply from our team — usually the same day.
Prefer to talk first? Book a 30-minute call 

Please enter your name.
Please enter a valid email address.
Please enter your store or a product URL.
Please choose a range.

Got it — check your inbox.

We'll review your account and reply within 1 business day.