Seven: cost of customer acquisition by channel, conversion rate, ACoS against the product's stage target, organic share of revenue, return rate, Buy Box percentage, and inventory cover in weeks. Plus one sentence per change made and why. Spend and sales alone are not reporting.
The short version
- Cost of customer acquisition by channel is the number that says whether growth is worth having.
- Organic share of revenue reveals whether an asset is being built or sales are being rented.
- ACoS against the stage target, not against an industry average.
- Return rate, because it destroys economics silently.
- What changed and why, in writing. That becomes your documentation.
The seven weekly metrics
I run Flapen with 50 operators managing about 70 brands, and the majority are profitable within their first year. This is the set we report, and each one is here because its absence hides a specific failure.
| # | Metric | What its absence hides |
|---|---|---|
| 1 | Cost of customer acquisition by channel | Unprofitable growth that looks like growth |
| 2 | Conversion rate | A listing problem being treated as an ads problem |
| 3 | ACoS vs stage target | A launch target quietly persisting into maturity |
| 4 | Organic share of revenue | Total dependence on paid |
| 5 | Return rate | A sourcing or quality problem |
| 6 | Buy Box percentage | Slow erosion from pricing or account health |
| 7 | Inventory cover in weeks | A stockout that will cost you ranking |
1. Cost of customer acquisition, by channel
Not blended. By channel, across the five: organic, paid, promotions, influencer and creator, and off-channel.
Blended cost of customer acquisition hides which channel is carrying the account and which is wasting money. Most sellers run two of the five, so a blended number is usually just paid wearing a disguise.
3. ACoS against the stage target
ACoS in isolation is close to meaningless. A 40 percent ACoS is excellent at launch, when you are buying velocity and ranking, and alarming on a mature product where the job is protecting margin.
So the report should show ACoS next to the current target and state the product's stage. If the target has not changed in a year, that is the finding.
4. Organic share of revenue
Moves slowly, which is why it belongs in a weekly report rather than a quarterly one. You want to see the trend early.
Rising paid share month over month means revenue is being rented. Every other metric can look healthy while this one deteriorates, and it is the single best predictor of what happens if you ever pause spend.
7. Inventory cover
Not glamorous and it decides more outcomes than most ad decisions. A stockout costs you ranking that took months and real money to build, and recovering it costs more than the inventory would have.
Weeks of cover, per product, against current velocity. If the agency does not track it, they are not managing the account, they are managing the ads.
What the weekly report should not be
A slide deck. Length is not a proxy for work, and the correlation usually runs the other way: reporting gets longer and less specific when there is less to say.
What good looks like in practice is a short written update posted on a fixed day: the seven numbers, what changed this week and why, what is planned next week, and anything that needs a decision from you. Ours is a written Slack update every week, a live review with the client's team every two weeks, and 24/7 access in between.
How often should we actually meet
Weekly written, bi-weekly live. That combination works because it separates information from decisions.
The written update keeps the numbers flowing without consuming an hour. The bi-weekly call handles the things that need discussion. Weekly calls on a stable account mostly generate a call, and monthly-only reporting lets a problem run for four weeks before anyone speaks.
After onboarding, a client of ours spends about 2 hours a month on this, mostly the bi-weekly call.
What most agencies will not tell you
Spend and sales are chosen for reports because they are the two numbers most likely to be rising, and both respond directly to budget. A report built on them can show a business getting steadily worse as a success story for a year.
The two metrics that would expose it, cost of customer acquisition by channel and organic share of revenue, require slightly more work to produce and can look bad while everything else looks good. That is exactly why they should be mandatory.
One more thing worth requiring: the sentence explaining why each change was made. Agencies resist it as busywork. It is the documentation that makes you portable, and a year of it means a new agency inherits a machine with a manual instead of starting by rebuilding.
Related answers
- KPI benchmarks for Amazon PPC agencies
- What to include in an Amazon PPC agency SLA
- Red flags your Amazon agency is underperforming
- How to audit current Amazon agency performance
- Hiring an Amazon agency: the complete guide
Our reporting cadence is fixed and published, at Flapen.
