Seller playbook · pricing evidence
Read the story behind a price—not just the last number.
A useful repricing review answers four questions: what changed, what was requested, what the marketplace actually showed, and what happened to the economics.
You do not need to understand proprietary algorithms to review your catalogue confidently. You do need the correct listing, time window and evidence. This guide is a review framework, not a promise that every historical data point exists in your account.
1. Start with the exact listing and period
Use marketplace, store and SKU together. The same SKU text can exist in more than one store. A product title is not a reliable identifier. Choose the period before comparing totals, and check which timezone defines the date boundaries.
For a busy listing, a fixed number of recent events may cover hours rather than a month. Check whether the history is complete, paginated, sampled or limited by retention. “No records loaded” is different from “nothing happened.”
2. Separate the decision from the outcome
A recommended or requested price is not necessarily the price customers saw. A preview can calculate a decision without sending anything. A hold may leave price unchanged. A marketplace acknowledgement can arrive before a matching price is observed.
Requested
The system recorded a target price and reason. Check whether this was preview-only, blocked or actually submitted.
Acknowledged
The marketplace accepted or processed a request. That status alone is not a matching live-price readback.
Observed
A later readback provides a price and timestamp. Compare it with the requested value and keep mismatches visible.
Even a matching price does not, by itself, prove that repricing caused a later order or a Buy Box win. Those outcomes require their own records and appropriate attribution.
3. Understand raises, cuts and holds
Why might a price go down?
A strategy may respond to a relevant lower offer, a lost Buy Box observation or a price-related recovery state. Check the recorded reason. Do not assume every decrease means a new competitor entered; that requires comparable offer evidence before and after the change.
Why might a price go up?
A strategy may seek a better margin within its ceiling, respond to changed economics, or follow a configured no-competition behavior. An upward request is an experiment in positioning, not proof that demand or the Buy Box will remain unchanged.
Why might the price stay the same?
A floor, ceiling or another configured limit may bind. The current price may already match the target, or the system may lack enough evidence to decide safely. A hold deserves an explanation rather than being counted as a price move.
4. Read cost changes as a separate story
A changed product cost, shipping charge, fee estimate or target margin can change the allowable price range even when competitors do nothing. A trustworthy explanation identifies the relevant cost version and effective time.
For older records without a saved economics snapshot, do not reconstruct precise historical margins using today’s costs and present them as facts. Show the missing evidence. Also avoid adding goods, shipping or fulfillment costs twice when the authoritative unit cost already includes them.
5. Choose the right average margin
A simple average of SKU percentages gives a product with one sale the same influence as a product with a thousand sales. For period performance, revenue-weighted contribution margin is usually the more useful measure: total included contribution ÷ total included revenue.
For example, one line with $100 revenue and $20 contribution has a 20% margin; another with $900 revenue and $90 contribution has a 10% margin. Their simple average is 15%, but combined contribution margin is $110 ÷ $1,000 = 11%.
Keep three labels separate: current estimated listing margin, historical order-time estimated margin, and reconciled contribution after known costs and adjustments. No orders is not a measured 0% margin. Missing costs are not zero costs.
6. Know what a Buy Box percentage measures
An observation-based win rate is winning observations divided by eligible known observations. It is not automatically the share of shopper impressions or the percentage of the day spent winning. Irregularly arriving observations can overrepresent busy periods.
Show the denominator and missing coverage. Unknown ownership should not silently become a loss. A later win after a loss is an observed recovery, not proof of which action caused it.
A repeatable five-minute review
Choose the same store and period, inspect unresolved requests, review cost changes, compare the confirmed price with the relevant limits, then open the decisions behind any unusual trend. Record what remains unknown before changing a strategy.
Use a shorter period to investigate an incident and a longer one to judge a recurring pattern. A single favorable screenshot is not a sufficient basis for a profitability claim.