Prediction Window

The timeframe during which a customer is statistically most likely to make a repeat purchase or accept an upsell, identified through cohort analysis. For consumable physical goods, the window is based on the depletion rate implied by the purchase quantity and usage patterns. For subscription software, it is based on the point at which a usage threshold is likely to be crossed. Timing an offer to this window — rather than to a calendar date — increases relevance and conversion. The window is not a fixed guarantee but a probabilistic guide that can be refined as more data accumulates.

Example: A coffee roaster calculates that customers who buy a 12-ounce bag reorder after a median of 21 days and sends a refill reminder with a complementary product offer on day 19.