Seller field guide
Amazon search term report analysis: decide what to keep, cap or investigate.
A review workflow for settled Sponsored Products data. Check the product margin, separate queries from ASINs and understand what zero orders can—and cannot—tell you.
By CEILR · Updated 25 October 2026 · Worked examples are illustrative.
Free spend-threshold calculator
Set a spending review threshold before the next click.
Enter contribution per order, the profit you want to keep and your average CPC. The calculator shows how much a search term can spend, and how many clicks that buys, before it consumes the ad allowance for one converted order. Reaching this threshold prompts a review; it does not prove that the term cannot convert.
Formula
Allowable ad spend per order = contribution per order − protected profit. Clicks at review threshold = allowable ad spend ÷ average CPC. Expected clicks per order = 100 ÷ conversion rate %. Seller-entered values only; nothing is fetched from Amazon.
Start with a report you can compare
Keep the marketplace, currency, advertised ASIN and attribution columns with the file. A US-dollar campaign and an Australian-dollar product profile cannot share a spending threshold. Nor should two products with different margins share one campaign-average CPC limit.
Use one report for a defined, completed date range. Record when you exported it and which purchase and sales columns you used. Recent clicks can still acquire attributed orders. Re-export the same range after the applicable attribution window before treating a zero-order row as settled. Amazon explains the reporting delay and attribution rules in its performance metrics and attribution documentation.
The customer-search-term column can contain a shopper query or an ASIN. An ASIN from product targeting is not text to paste into your listing. Keep it with the product-targeting review. Amazon's targeting guide describes ASINs appearing in that column.
Check these fields before making a decision
| Field | Check | Why it matters |
|---|---|---|
| Marketplace and currency | AU/AUD or US/USD, consistently | A dollar of allowable spend must use the same currency as the report. |
| Advertised product and target | Retain ASIN, campaign, ad group and match type | The same query can reach different products and economics. |
| Clicks, spend, orders and sales | Use one compatible attribution basis | Orders per click and spend per order must describe the same cohort. |
| Missing values | Keep blank separate from zero | An absent purchase column does not establish zero purchases. |
| Report grain | Use a summary or non-overlapping daily rows | Appending overlapping exports counts the same clicks more than once. |
When combining daily rows, add clicks, spend and orders, then recalculate the ratios. Do not average row-level CPCs or conversion percentages: a one-click row should not get the weight of a hundred-click row.
Four rows, one product, four different actions
This illustrative US product contributes US$12 per order before ads. The seller wants to retain US$4, leaving an ad allowance of US$8 per converted order. The rows below use the same completed reporting period. They are not CEILR customer results.
| Search term / target | Clicks | Spend | Orders | Working | Review |
|---|---|---|---|---|---|
| insulated hiking bottle | 40 | US$24 | 4 | CPC $0.60; CVR 10%; ceiling $0.80 | Hold the current economics; check stock and funding before more spend. |
| bottle with straw | 20 | US$18 | 2 | CPC $0.90; CVR 10%; ceiling $0.80 | Review bids and relevance: $9 of ads per order exceeds the $8 allowance. |
| water bottle | 10 | US$8 | 0 | One ad allowance spent; CVR estimate is thin | Review the settled data. Zero orders alone does not establish irrelevance. |
| A product-target ASIN | 5 | US$3 | 0 | Product placement; limited evidence | Review the target product and fit. Do not copy the ASIN into listing text. |
For the first row, the full calculation is (US$12 − US$4) × (4 orders ÷ 40 clicks) = US$0.80 per click. Use the PPC bid calculator for your inputs, and the bid-setting example before translating the ceiling into a campaign setting.
Ten clicks without an order is a budget signal, not a statistical verdict
Suppose the true conversion probability were 10% and each click were independent. The chance of no orders after ten clicks would be 0.9010 = 34.9%. That is too common to prove a relevant term cannot work. After 30 clicks the same model gives 4.2%. Your true rate is unknown, and conversion can change with price, placement and listing conditions.
| Clicks with no order | Modeled probability at 5% CVR | Modeled probability at 10% CVR |
|---|---|---|
| 10 | 59.9% | 34.9% |
| 20 | 35.8% | 12.2% |
| 30 | 21.5% | 4.2% |
This is a sensitivity check, not a prescription to buy another 30 clicks. Your cash limit may require a pause before statistical confidence improves. Separate “I cannot fund this test” from “I have evidence that the traffic is irrelevant.”
Choose the action that the evidence supports
- Irrelevant query: review the exact wording and the product before choosing the negative match type. A query asking for a feature you do not sell needs a different response from a relevant term with a thin sample.
- Relevant, above the economic ceiling: inspect the effective bid, placement adjustments and conversion assumptions. A lower bid or a capped test can be more appropriate than a negative.
- Converting, within the ceiling: consider a controlled exact-match test. Check cash and inventory first. Moving a term does not guarantee more orders.
- Missing or immature evidence: keep the decision unresolved. Name the missing column, settlement period or product cost rather than filling it with a benchmark.
Adding a negative exact in the source campaign can deliberately route traffic to the new exact campaign. It is a routing choice, not a required step. Make sure the destination is eligible and funded; blocking the source first can remove traffic you meant to retain.
Converting query language may help a listing when it accurately describes the product. Add it naturally and check Amazon's listing rules. The backend search-terms checker is for text queries, not targeted ASINs.
Keep the before-and-after evidence
For each change, keep the ASIN, marketplace, source campaign/target, report range, contribution, protected profit, observed CPC, selected action and date. Save the original export. Evaluate a later, non-overlapping cohort after its attribution window, with price, coupon and stock changes alongside it. Otherwise a changed result cannot reliably be assigned to the bid change.
CEILR's PPC workflow keeps the economics and evidence with a seller-reviewed decision. Amazon access is read-only. You make any bid, budget or negative-target change in Amazon. Missing cash inputs leave Fundable CPC unavailable.
Sources and limits
Amazon sources linked above were checked on 25 October 2026. Interface labels and attribution columns can vary by report, ad type and account. The worked rows and probability table are explanatory arithmetic, not observed account performance or a performance forecast.
Use your own numbers
Review your next report with the working attached.
Public calculators are free. The CEILR workspace is a paid subscription; Amazon changes stay under your control.