When might a customer be ready to buy again?
01 / THE APPROACH
Measure consecutive product purchases
For each identified customer and product, sort qualifying purchase events by date and measure consecutive intervals. Decide whether variations are analysed separately or grouped, and state that policy. Multiple units in one order do not establish multiple repeat purchases. Different orders on the same day may require explicit handling to avoid treating fulfilment quirks as a replenishment rhythm.
02 / THE APPROACH
Use distributions, not just an average
A median repurchase interval is less sensitive to unusually long gaps than an average. Show the 25th and 75th percentiles and the number of repeat customers to communicate how consistent the pattern is. A narrow interval supported by many customers is more useful for timing than a wide spread from a handful of purchases. Historical stock gaps or promotions can distort the pattern.
03 / THE APPROACH
Turn a pattern into an opportunity group
Compare time since the last purchase with a stated product-level or customer-level interval. Define due-soon and overdue bands explicitly; different definitions produce different counts. Exclude unreliable identities, check recent purchases, and avoid presenting overdue as confirmed churn. The initial analysis can identify pseudonymous customer groups without exporting names or email addresses.
Before you act on the analysis
- Use reliable customer identity and repeated product purchases.
- Show repeat-customer sample size and interval percentiles.
- State the rules for due-soon and overdue groups.
- Keep communication execution and direct customer identifiers outside the analytical report.
WHAT WE’RE BUILDING
From the question to a useful next step.
Retail Science Co is developing product-level reorder intervals and pseudonymous opportunity groups. Early access is designed to identify and explain the opportunity; it does not automate email campaigns or send customer messages.
Join the early-access waitlist