Retail Media Planning
Part of Retail audience planning
Reviewing the recency of retailer shopping signals
Check event time, data arrival, segment refresh and coverage before treating shopping signals as current.
Recency indicates when a shopping signal was observed, not whether the shopper is ready to buy today. Review it alongside the product’s buying cycle, the retailer’s data coverage and the time required for a new event to enter the audience. A label such as “recent buyer” is incomplete without those details.
Separate event age from update delay
An event can be recent in the database but older in real life if store transactions arrive after a delay. Ask for the timestamp of each: when the purchase or interaction occurred, when it was processed and when the activated audience was last refreshed. Also ask whether returns, cancellations or identity changes remove an event later.
| Timestamp | Why it matters |
|---|---|
| Event date | Shows how old the observed action is |
| Data arrival | Reveals processing delay or missing channels |
| Segment refresh | Shows when targeting rules last reflected new data |
| Campaign exposure | Allows the team to compare message timing with the original event |
If a field is unavailable, note the gap. Do not substitute the day the report was downloaded.
Relate the window to the category
An appropriate lookback period depends on the purchase pattern. For a frequently bought item, a long window may mix current and stale interest. For a durable item, a very short window may find mainly people who have just completed their purchase.
Build two or three candidate windows as a planning exercise and ask how reach and overlap change. Do not adopt a universal number of days because another campaign used it.
Consider whether the event is a category purchase, a brand purchase or browsing. Browsing can indicate exploration but may be less decisive than a completed purchase. A purchase is a strong record of past behaviour, but can still be for another person or a one-off need. Match the confidence of the message to the evidence.
Planning Recency Windows by Product Category
- Identify purchase frequency for categorye.g., groceries (weekly), furniture (annual)
- Assess reach and overlap across windowsCompare audience size and duplication
- Align window to buyer intent and confidence levelShorter for high-intent actions; longer for exploratory behaviour
Check coverage before interpreting absence
Retailer data may cover only particular stores, loyalty identifiers, online transactions or eligible customers. Ask what share of relevant sales can be observed and whether that share changes by channel or period. “No recent purchase” means no qualifying event in the available data. It does not mean the individual did not buy elsewhere or without an identifiable record.
This matters when defining lapsed buyers or suppressing recent purchasers. An incomplete dataset can make a shopper appear inactive. Document the definition in the brief and avoid definitive language in the creative or report.
Put recency into the analysis plan
Store the final window and refresh rule with the campaign version. If the rule changes mid-flight, mark the change; otherwise later results blend different audiences. Ask for reach, frequency and sales reporting by the agreed window where privacy and sample size allow. Do not request individual-level purchase histories merely to satisfy curiosity.
IAB Europe’s measurement standards give common terms for commerce media reporting, but they do not replace the retailer’s explanation of its actual data pipeline. Use the standard to ask better questions, then state any difference between the agreed definition and what the final report could deliver.
