Skip to content

How to choose an Amazon agency with case studies

Vet agencies through five gates, a launch case, a maturity case, different ACoS targets for each, a real failure story, and a reference call you choose.
·4 min read
Competitor AnalysisPPCOrganic RankingAmazon FBA
Joel Turcotte Gaucher

Joel Turcotte Gaucher

Founder

Flapen cover for How to choose an Amazon agency with case studies: a product family of three sizes lined up on the studio sweep

Judge case studies by stage context, not revenue headlines. Ask for one launch case and one maturity case, then ask what ACoS target each was run to. An agency that quotes a single blended ACoS for both is presenting marketing, not method. Follow with a reference call to a comparable brand.

The short version

  • Revenue headlines are the least useful number in a case study. Anyone can buy topline with enough ad spend.
  • Ask for a launch case and a maturity case. They should look nothing alike.
  • Demand the advertising target behind each. One blended number for both stages means there is no method underneath.
  • The failure case matters more than the win. Every real portfolio contains one.
  • Finish with a reference call you select. From a list of clients, not their single favorite.

The five-gate vetting sequence

The most common buying mistake I see is picking the deck with the biggest before-and-after chart. It costs sellers months, because a chart proves the agency once had a client who grew, not that the growth came from the work. Run this sequence instead, and stop at the first gate a candidate fails.

  1. Request one launch case and one maturity case. Gate: both exist, and the tactics differ. Launching a product and defending a category leader are different jobs. A portfolio showing only one of them tells you which job the agency actually knows.
  2. Ask what ACoS target each case was run to. Gate: two different numbers, each with a reason. A launch tolerates expensive traffic to buy rank and reviews. A mature product is managed for efficiency. At Flapen the target moves with product stage, aggressive early and tight later, and any agency running real strategy can name both figures for their own cases without checking.
  3. Ask for the case that went wrong. Gate: a specific product, the decision taken, and what it cost. A firm that claims a spotless record is either new or editing. What you are listening for is whether they noticed early and acted, or noticed late and hid it.
  4. Make them walk you from revenue to profit. Gate: the case includes fees, ad spend, and margin, not just sales. Growth that loses money is trivially easy to manufacture on Amazon.
  5. Take a reference call with a brand shaped like yours. Gate: you pick the reference from several options. Similar price point, similar catalog size, similar stage. One hand-selected superfan proves nothing.

Why stage context is the whole test

Amazon economics invert as a product ages. Early on, you pay a premium for visibility because rank, reviews, and conversion data compound later. At maturity, the same spending pattern quietly eats the margin the launch was meant to create. A case study that hides which stage it covers hides whether the tactic was appropriate. That is why gate two filters more bad agencies than any other question I know, and why I would run it even if you skip the rest of the sequence.

Reading the numbers like an operator

Three adjustments before you believe any chart. Check the attribution window, because a long window claims organic sales as ad results. Check the season, because a September-to-December curve is mostly the calendar at work. Check the fee model, because a firm paid on ad spend has every reason to feature its highest-spending accounts. A flat-fee structure, like the per-product tiers we list at pricing, removes that last distortion from the story.

What most agencies will not tell you

Every case study is a survivorship sample. The portfolio you are shown is the numerator, and nobody volunteers the denominator. So ask for the base rate directly: of the last ten accounts you onboarded, how many were profitable a year later. Our own answer is that the majority of brands we manage reach profitability within their first year, and I quote that figure because almost nobody else in this industry will state theirs. A candidate who refuses the question is answering it.

Run the five gates on us first if you like, starting with the free 48-hour audit at Flapen.

Keep learning

Frequently Asked Questions

Share this post
The Flapen Weekly Product Research report, an Amazon niche shortlist scored 0–100 with its score radar on the cover

The weekly niche report

Product research, in your inbox

Every niche that cleared the bar this week: what it sells for, what it costs to enter, and why it passed. When we get one wrong, we publish the correction.