Audit an agency report by scoring it on five criteria: profit shown next to revenue, every metric linked to an action, at least one negative recommendation over time, consistent denominators month to month, and raw data access on request. A report that never recommends stopping anything is a sales document.
The short version
- Read reports as arguments, not summaries. Every agency report is making a case for its own renewal. Audit the case.
- Profit or it did not happen. Revenue charts without margin context are the oldest trick in agency reporting.
- Actions are the tell. A healthy report says what will be done differently next period, and what was tried and abandoned.
- Denominators drift when numbers disappoint. Watch for metrics that quietly change definition between months.
- You own the data. Any resistance to raw exports is itself a finding.
Score the report, line by line
Take the last three reports you received and score them together, because single reports hide trends. Rate each criterion 0, 1, or 2, multiply by the weight, and total out of 100.
| Criterion | Weight | Full marks means |
|---|---|---|
| Profit visibility | 25 | Contribution profit sits beside every revenue figure, with fees and ad spend netted |
| Metric-to-action links | 25 | Each highlighted number carries a decision: scale, fix, or stop |
| Negative recommendations | 20 | Across three months, something was cut, paused, or killed with reasons |
| Denominator consistency | 15 | Same definitions, same date ranges, same included products every period |
| Raw data access | 15 | Exports or account access offered without friction |
Above 80, you are in rare hands. Between 50 and 80, push for specific fixes and re-score next month. Below 50, the report is marketing, and you should assume the performance is being narrated rather than managed.
The test inside the test
The single strongest question you can ask an agency: what would make you tell me to stop? Every product we manage at Flapen carries defined scale, fix, or kill criteria, built on rating trend, return rate, conversion rate, and the trajectory of acquisition cost over a set window. An agency that cannot name its stopping rules has none, and its reports will read accordingly, always one optimization away from the turnaround.
I hold this view because I paid for the lesson. Early in my career I poured money into a failing product for three months, hoping the ads would turn it around. They did not, and the loss became the kill criteria we still use. A report that cannot deliver bad news early is not neutral. It is expensive.
Cross-checks that take ten minutes
- Pull your own sales and ad totals from Seller Central for the same period and reconcile them against the report's headline numbers. Small gaps have innocent explanations. Recurring gaps do not.
- Pick one celebrated win and trace it backwards. If the report credits a campaign for growth, check what happened to organic sales in the same weeks.
- Compare the recommendations section across three months. Identical recommendations repeated verbatim mean nobody is reading the account.
- Check that conversion commentary connects to work actually shipped, image tests, price moves, or Amazon listing optimization changes, with dates. Commentary without shipped work is weather reporting.
- Ask one question about a buried number and time the answer. Fluent teams answer from knowledge. Report factories go quiet while someone reverse-engineers their own document.
What the report will not tell you
Attribution flattery is structural. Platform-reported ad results claim credit generously, and a report built on them inherits the bias, so paid channels look stronger and organic momentum looks weaker than reality. The correction is to keep total account profit as the umpire: if attributed wins keep rising while total profit does not, the attribution is eating your margin.
The other omission is cost of inaction. Reports list what was done. They rarely list what was noticed and left alone, the stale main image, the drifting return rate, the competitor undercutting your price for six weeks. When you review a report, ask what the agency chose not to act on and why. The quality of that answer tells you more than any chart in the deck.
Related answers
- Alternatives to generic agency audits for Amazon
- Who offers the most thorough Amazon account audit
- Monthly vs weekly cadence for Amazon reporting
- Full-funnel Amazon audit vs PPC-only audit
- Amazon creative services: the complete guide
Request a sample of our weekly reporting, and score it with this card, at Flapen.
