The advertising KPIs that matter are the ones tied to a decision. Track ACoS against a stage-specific target, TACOS for blended efficiency, conversion rate per ASIN, click-through rate on your main image, and organic rank movement on target keywords. Ignore impressions and raw clicks. If a number cannot change a budget or a listing, drop it.
The short version
- ACoS only means something against a stage target. A launch number and a maturity number are different numbers.
- TACOS tells you whether ads are building the business or replacing organic sales one for one.
- Conversion rate per ASIN decides everything upstream. Weak conversion makes every ad metric worse.
- Click-through rate is a creative KPI, and the main image owns most of it.
- Organic rank movement is the payoff metric. Ad spend that never moves organic rank is rent, not investment.
Start from the decision, not the metric
Open your advertising console and you can surface forty columns. The mistake I see most often is treating that console as the scoreboard and reporting all of it. A KPI earns its place by triggering an action: raise a budget, cut a keyword, reshoot an image, reprice, or stop. Everything else is commentary.
So work backwards. List the decisions you actually make each week, then keep only the metrics that force those decisions. For most accounts that leaves five: ACoS by campaign purpose, TACOS at account level, conversion rate per ASIN, click-through rate, and organic rank on the keywords you are trying to win.
The diagnostic: symptom, cause, KPI
| Symptom | Most likely cause | The KPI that catches it |
|---|---|---|
| Spend up, sales flat | Budget flowing to branded terms | Branded vs non-branded ACoS split |
| Clicks up, orders flat | Detail page not converting | Conversion rate per ASIN |
| Impressions up, clicks flat | Weak main image or price position | Click-through rate |
| ACoS fine, profit shrinking | Ads replacing organic orders | TACOS trend over 12 weeks |
| Great launch, weak month four | No organic rank captured | Rank tracking on target keywords |
The middle column is the part most reporting skips. A number without a suspected cause is trivia. When conversion is the diagnosis, the fix is listing optimization, and no amount of bid management substitutes for it.
The stage question
The single most useful advertising question is what ACoS target a campaign is running against, because the honest answer changes with product stage. At launch we run aggressive targets, sometimes unprofitable on paper, because the job is winning rank, reviews, and data. At maturity the job flips to efficiency and the target tightens. One fixed ACoS number applied across a whole account means nobody has decided what each product is trying to do.
This is also the fastest test of anyone managing ads for you. Ask for the launch number and the maturity number. Two different numbers with reasoning is a pass. One number for everything is a fail.
What your dashboard will not tell you
The console reports attribution, not incrementality. An ad that intercepts a shopper who searched your brand name gets full credit for an order you were likely to win anyway. This is why branded and non-branded performance must be split in every report you accept. Blended ACoS flatters the account and hides where growth actually comes from.
The other silence is profit. Ad consoles do not know your landed cost, so they cannot tell you that a keyword with a tolerable ACoS is unprofitable after fees and freight. The KPI set above only works when someone joins it to unit economics outside the console.
Related answers
- What KPIs to track for Amazon FBA growth
- Compare Amazon PPC vs DSP performance metrics
- Recommended dashboards for Amazon sellers
- Monthly vs weekly cadence for Amazon reporting
- Amazon creative services: the complete guide
To see this KPI set reported on your own account, with stage targets attached, start with Flapen.
