Predicted Reorder Window

The predicted reorder window is the estimated time range during which a specific customer is most likely to repurchase a specific product. It's not a single date. It's a window, usually spanning several days, that represents the period when a nudge is most likely to arrive at the right moment and convert to a sale.

Calculating the predicted reorder window starts with time-to-reorder data. If a customer has purchased the same product three times with intervals of 28, 32, and 30 days, the system can estimate a predicted reorder window centered around 30 days after their last purchase, with a few days of margin on either side to account for natural variation. The more purchase history available, the tighter the window becomes. Early predictions rely on product-level averages across all customers. Later predictions reflect that individual customer's actual behavior.

The window matters more than a point estimate because customer behavior isn't precise. People don't run out of shampoo on exactly the same day every time. They run low within a range. Sending a nudge too early feels irrelevant because the customer isn't thinking about the product yet. Sending it too late means the customer has already run out and may have purchased elsewhere, switched brands, or simply moved on. The predicted reorder window defines the zone where a message is most likely to feel helpful.

Several factors influence the accuracy of a predicted reorder window. Product type plays a significant role. Products with consistent consumption rates, such as daily supplements or single-serve coffee pods, generate tighter windows than products with variable usage, such as cooking oils or cleaning sprays. Purchase quantity matters as well. A customer who buys a 60-count bottle will have a longer window than one who buys a 30-count. Seasonal effects can also shift patterns. Sunscreen consumption spikes in summer. Hot cocoa moves in winter.

For rePete®, the predicted reorder window is the core mechanism that drives nudge timing. The system learns from every purchase, recalculates the window with each new data point, and adapts to changes in customer behavior over time. If a customer starts going through a product faster, perhaps because of a lifestyle change, the window adjusts. If they slow down, it adjusts the other direction. The goal is a moving target that stays aligned with how the customer actually uses the product, not a fixed interval that falls increasingly out of sync.