The key idea
A demand increase during a price promotion should not automatically become the new normal. Record the price change, promotion window and relevant order behavior together. A stock decision needs to consider what demand may look like when the promotion ends, not simply extend the most recent discounted period indefinitely.
Keepa documents price observations. The scenario analysis above is a proposed way to avoid overinterpreting them. [1]
Separate temporary conditions from a persistent change
Annotate discounts, bundles and seller changes in the timeline used by the planner. Check whether the price is the one customers actually encountered in the relevant offer and market. An observation of one offer does not establish that every customer bought at that price. When there is uncertainty, keep it explicit. Compare the campaign with a reasonable prepromotion period while noting other changes such as advertising or availability.
Use scenarios before changing the target
Consider what happens if demand returns to baseline, remains partly elevated or continues at the promotional level. Then compare the receipt date of a new run with the campaign end. If the replenishment arrives after the discount ends, ordering solely from the peak rate may create excess stock. A smaller staged response can preserve flexibility while additional orders reveal whether the increase persists.
Worked example · illustrative assumptions
Assume an edition rises from two to eight daily orders during a five day offer. A new batch takes six days to become available. Ordering 48 copies by multiplying eight by six treats the temporary promotion as persistent and ignores the missed selling window. A scenario using confirmed waiting orders plus a postpromotion forecast may recommend a different amount. The arithmetic is simple; choosing the relevant demand assumption is the real decision.
Put it into practice
- Record campaign start and end dates alongside price observations.
- Compare the expected receipt date with the remaining promotion window.
- Review the target again when the price changes back.
Watch out for
Do not claim that a promotion caused all additional sales without a suitable comparison. Coverage, seasonality and stock availability may change at the same time. A planning annotation is not proof of causation.
Should price be a feature in a forecast?
It can be, if the field is reliable and its contribution is tested. The model also needs to know which future price assumption is being used.
Sources and further reading
- Keepa: Product object. Documented market observations, missing values and field timestamps.
AI assisted editorial content with original planning examples. Source links support the stated source notes; operational suggestions are guidance rather than guarantees. Provider rules and product coverage can change. Editorial approach · BooksAhead.ai connection status.