Upload a Sponsored Products search-term report to find profitable queries, wasted spend, negative-keyword candidates, and terms worth moving into exact-match campaigns.
Your files stay private: all analysis runs locally in your browser.
Download a Sponsored Products search-term report from the Amazon Ads console, choose a meaningful date range, and upload the CSV above. The analyzer recognizes common Amazon column names for customer search term, clicks, spend, sales, and orders.
Set target ACoS to the maximum advertising cost your economics can support. For a defensible ceiling, start with your pre-ad contribution margin rather than an arbitrary industry benchmark.
A term with orders and ACoS at or below your target is marked for exact-match consideration. A clicked term with meaningful spend and no order is marked as a negative candidate. Terms above target are flagged for bid reduction or a deeper listing-relevance review.
These are triage rules, not automatic campaign commands. Check attribution lag, conversion volume, match type, placement, and whether a query has strategic value before making changes.
Review search terms on a consistent schedule. Harvest proven customer language into controlled campaigns, isolate expensive queries, and compare changes over equivalent periods. Do not react to a single click or one unusually large order.
Use this analyzer with the Amazon ACoS calculator and break-even CPC calculator to connect keyword decisions to product economics.
The Sponsored Products search-term report has grown to dozens of columns. These are the ones that drive nearly every optimization decision:
| Column | Definition | What a bad number looks like |
|---|---|---|
| Impressions | How many times your ad rendered for this query. | High impressions with a very low CTR usually means the query is related to your product but your ad (image, price, title) is not winning the click — or the query is broader than your niche. |
| Click-Through Rate (CTR) | Clicks ÷ impressions. Amazon averages are commonly cited around 0.3–0.5% for Sponsored Products. | Under ~0.2% on meaningful impressions suggests weak relevance or an uncompetitive main image and price. |
| Cost Per Click (CPC) | Spend ÷ clicks for the query. | A CPC near or above your per-unit profit means even a perfect conversion rate cannot make the term profitable. |
| ACoS | Ad spend ÷ ad-attributed sales, as a percentage. The core efficiency metric. | Above your break-even ACoS, every attributed sale loses money before overhead. |
| RoAS | Sales ÷ spend — the inverse of ACoS (RoAS of 4 = 25% ACoS). | Below ~2.5 is unprofitable for most private-label margin structures. |
| 7/14 Day Total Orders & Sales | Conversions attributed to the click within the attribution window. | Recent clicks may not have converted yet — judging a term on 2 days of data negates keywords that were still working. |
The most expensive PPC mistake is negating too early; the second most expensive is never negating at all. A defensible middle ground uses your own conversion rate to set the threshold: if you convert 10% of clicks, then 20+ clicks without an order is roughly a 12% probability event — unlikely, but not damning. This matrix is a starting framework:
| Situation | Suggested action | Why |
|---|---|---|
| 0 orders, fewer than 10 clicks | Collect more data | Too little evidence. At a 10% conversion rate, 9 clicks without a sale happens 39% of the time by pure chance. |
| 0 orders, 10–20 clicks, spend under your target cost-per-order | Watch, or reduce bid 20–30% | The term is suspicious but hasn't cost a full conversion yet. Lower the bid to cut the cost of the remaining test. |
| 0 orders, clicks ≥ 2–3× your typical clicks-per-order | Negative-exact candidate | You have paid for two orders' worth of clicks with nothing back. Negate the exact query; keep the broader theme if other variants convert. |
| 0 orders and clearly irrelevant (wrong product, wrong use case) | Negate immediately | Don't wait for statistics when relevance is obviously wrong — every click is pure waste. |
| Orders present, ACoS at or below target | Harvest to exact match | Move it into an exact-match campaign with its own bid, then negate it in the source campaign so the two don't compete. |
| Orders present, ACoS above target | Bid down 20–30%, re-check in 2 weeks | The term converts — the price of the click is the problem, not the query. |
Use negative-exact for individual bad queries and negative-phrase only for whole themes you never want (e.g. "used", a competitor brand you cannot convert, an incompatible model number). A careless negative-phrase can silently block dozens of good queries.
Suppose a product sells at $24.99 with $6.50 landed cost and $8.70 total Amazon fees. Pre-ad contribution is $24.99 − $6.50 − $8.70 = $9.79 per unit.
Break-even ACoS = 9.79 ÷ 24.99 ≈ 39.2%. At exactly 39.2% ACoS, ad-attributed sales make zero contribution — every point above it loses money on those units.
Target ACoS should sit below break-even by your required profit. To keep $5 of every unit sold through ads: (9.79 − 5.00) ÷ 24.99 ≈ 19.2%. That, not a generic "30% is fine", is the number to enter above — and it's why two sellers can look at the same search term and rationally make opposite decisions.
One nuance: for a brand-new launch, running above break-even ACoS temporarily can be rational if the paid orders buy organic rank and reviews. Just make it a dated decision with a budget, not a default.
No. The browser reads and analyzes the CSV locally.
No. Consider clicks, spend, attribution lag, relevance, and your normal conversion rate first.
Use an ACoS below your break-even level, with room for overhead and your desired profit.
It depends entirely on your margin. Compute break-even ACoS as pre-ad profit divided by price — often 30–45% for private label — then set your target below it by the per-unit profit you want to keep. Cross-category averages are frequently cited around 25–35%, but your own break-even is the only number that matters.
Scale it to your conversion rate: roughly 2–3× your typical clicks-per-order without a sale is a statistically reasonable threshold. At a 10% conversion rate that's 20–30 clicks. Obviously irrelevant queries can be negated immediately.
ACoS measures ad spend against ad-attributed sales only. TACoS (total ACoS) divides ad spend by total revenue including organic. A falling TACoS with stable ACoS is the classic sign that advertising is building organic rank rather than just buying sales.
Last updated:
Powered by HumanCalculations — free online calculators