Give search terms a fair test before adding negatives.

The question

A search term has 15 clicks and no orders. Should you block it? The click count alone cannot answer. You need to know how often it must convert to pay for those clicks, and whether the reported orders have had time to arrive.

Work through the numbers

In this teaching example, a click costs $1.20 and a converting purchase contributes $12 before advertising. Break-even conversion rate is $1.20 divided by $12, or 10%. On average, one purchase per ten clicks pays for the advertising.

Imagine the term really does convert at 10%, with independent clicks and a stable rate. The chance of no purchase on one click is 90%. The chance of no purchases on 15 clicks is 0.9 raised to the fifteenth power, about 20.6%. Zero orders is disappointing, but that result is still fairly common for a break-even term.

At 29 clicks, the chance falls below 5%. At the tool’s 95% setting, that is a flag to review a negative under this model. It is not a 95% probability that the term is unprofitable. The calculation asks how unusual this evidence would be if the term converted at break-even.

Now predict what happens if contribution before ads doubles to $24 while CPC stays $1.20. Break-even falls to 5%. Zero orders need 59 clicks before crossing the same threshold: 0.95 raised to the fifty-ninth power is about 4.8%. A blanket “negate after 15 clicks” rule treats these products as if their economics were identical.

Relevance is a separate check. If a query clearly asks for something your product cannot provide, you can review it for exclusion without buying enough clicks to complete a profitability test.

A published case to examine

My sample cosmetics audit identified $504 in spend on search terms with zero orders, 19.4% of the $2,591.91 total. The published review described gaps in negatives and broad or phrase traffic that was not converting.

The $504 is a review queue, not a proven savings amount. The public case does not show every term’s clicks, margin and attribution maturity. Those details determine which terms merit a negative, a lower bid or more observation. This was an audit snapshot, not a completed improvement result.

Read the cosmetics sample audit.

Try it on your numbers

1. Choose one relevant term from a settled Search Term Report. Use clicks, spend and click-attributed purchases from the same full attribution window.
2. Open When to Negate. Enter the three totals and the product’s contribution before ads, or use the product from Catalog Profit Check.
3. Start with the 95% setting. Read both the result and the break-even conversion rate.
4. For an early or watch result, decide how much further testing fits your budget. The displayed extra-click cost is conditional on no new order and unchanged CPC.
5. For a negative flag, check the query’s relevance and the intended exclusion scope before changing Amazon settings. Keep a note of the reason and date.

Open When to Negate

What still needs your judgment

The test assumes stable conversion, CPC and contribution, and at most one converting purchase per click. Recent report dates can miss later-attributed orders. Repeated checks and screening many terms increase the chance of false flags; the selected threshold does not remove that problem.

A negative can affect more traffic than the one row you inspected. Review the match type and campaign scope. The tool does not choose those settings, account for assisted sales, or diagnose a listing that stopped converting.

Explain it to your team

“I checked this query against the conversion rate its margin requires. The evidence supports a review, and I confirmed the exclusion scope.”