Ask for case studies, then read them backwards. Start with the starting position, the time elapsed, and the spend behind the growth. A revenue chart with no baseline, no calendar, and no cost line is a screenshot, not evidence. The useful case study names the constraint that was removed.
The short version
- The first number that matters is the starting number. Growth from $4,000 a month is a different discipline from growth from $400,000.
- A chart without a spend line is decoration. Revenue bought with unprofitable advertising is not scaling.
- The best case studies name a constraint. Rating gap, primary image, stock-outs, wrong keyword set, wrong price.
- Anonymised is fine. Vague is not. Category, timeframe, and mechanism can all be shared without naming the brand.
- Ask who did the work. A case study is a claim about a team, not about a logo.
The mistake that costs the most
The common failure is treating a case study as proof of ability rather than as a description of a situation. A seller reads "we grew this brand 340 percent" and hires on the strength of the percentage. What they have not asked is whether the brand was recovering from a suspension, whether it launched three new variations that quarter, whether the growth came from a category tailwind, or whether the agency inherited a listing that only needed a better main image.
Every one of those is a legitimate result. None of them tells you what will happen to your account, because your account has a different constraint.
So change the question. Instead of "show me your best results", ask this: show me a brand that looked like mine, and tell me what was holding it back. If the answer is a mechanism you can repeat, the case study is worth something. If the answer is a percentage, it is worth nothing.
Diagnostic: what a revenue chart is actually telling you
Use this table on any case study you are shown, including ours. The point is not to catch anybody out. It is to convert a marketing asset into information you can act on.
| What the case study shows | Most likely cause | Who actually fixes it |
|---|---|---|
| Revenue up, ACoS up faster | Growth bought with advertising, not earned | An operator willing to cut spend and rebuild conversion first |
| Revenue up, sessions flat | Conversion rate improved (images, price, reviews) | Listing and creative team |
| Sessions up, revenue flat | Traffic landing on a page that does not convert | Same team, in the opposite order |
| Step change on one date | A stock-out ended, or a suspension was lifted | Supply chain or account health, not marketing |
| Steady climb over months | Organic rank compounding on a keyword set | Keyword and content ownership over time |
| Spike then decay | Promotion or external burst with no retention | Nobody. That pattern is not scaling |
Three of the six rows have nothing to do with advertising. That is about the ratio I see in real accounts, and it is why an agency that only shows you ad dashboards is showing you a slice of the machine.
The question behind every case study
Whoever presents the case study did not do the work alone, and the number of accounts each person carries decides whether the work in that case study can be repeated on yours.
At Flapen, our operators carry about 1.4 brands each. That ratio is the single most useful thing I can hand a buyer, because it converts an unanswerable question, whether they are good, into an answerable one, whether they have the hours. A manager holding twelve accounts can run reports on all twelve. They cannot rewrite twelve catalogs, brief twelve photoshoots, and sit in twelve weekly reviews.
So after the case study, ask the person presenting it: how many brands were you personally responsible for during that period, and how many are you responsible for now. Compare the two answers. If the second number is much larger, the case study describes a version of the company that no longer exists.
Building your own before you hire anyone
You can construct the case study you want in advance, which makes the hiring conversation much shorter.
- Write down today's baseline: units, sessions, conversion rate, ACoS, return rate, average rating.
- Name the one constraint you believe is holding the account back.
- Define what "fixed" looks like as a number, with a date attached.
- Ask each candidate what they would do about that specific constraint in the first 30 days.
- Compare their answer to the case studies they showed you. Consistency between the two is the actual signal.
Candidates who fail this are not necessarily bad operators. They are operators whose experience does not match your constraint, which is a scheduling problem, not a character judgment.
What most agencies will not tell you
Case studies are selected, not sampled. Every agency shows the brands that worked. Nobody publishes the account that plateaued for eight months, or the launch that got killed at week ten because the return rate would not come down. That is not dishonesty, it is marketing, but it means the case study page has almost no predictive value on its own.
Here is the disclosure that matters, and most agencies will not tell you this unprompted: ask for the churn story, not the success story. Ask what happened with the last three clients who left, and why. The answers are far more informative than any chart, because they describe the situations where the agency's method does not work. We tell people ours: brands that want us to promise a revenue number in month one, and brands sitting on a product that our own kill criteria say should be discontinued, do not stay long.
The majority of the brands we take on are profitable inside their first year. That number is a real benchmark and you should hold any agency to something like it. It is also not a promise about your brand, because I have not seen your brand yet.
Related answers
- Who can audit my Amazon account and fix conversion
- Best agencies for Amazon brand management
- KPIs an Amazon agency should report weekly
- Questions to ask before hiring an Amazon agency
- Done-for-you Amazon management: the complete guide
Send us your ASINs and we will write the diagnostic version of your account for free at Flapen.
