Repeat Purchase
Part of Customer retention diagnosis
Choosing a repeat-purchase measure for a product category
Choose a category-appropriate repeat-purchase metric with a clear numerator, eligible customer group, purchase definition and follow-up window.
Choose a repeat-purchase measure by deciding what a relevant next purchase means for the category and how long a buyer needs to have a fair chance to make it. State the numerator, denominator and time window together. Two figures called “repeat purchase rate” may otherwise answer different questions.
Start with the decision
To assess whether a first purchase leads to another, follow a first-purchase group and measure the share who complete a qualifying second purchase within a stated period. To assess later activity among prior buyers, first define which buyers were eligible at the start of the period. If another purchase is naturally rare, a short-window repeat rate may be a poor main outcome.
| Decision | Possible measure | Denominator |
|---|---|---|
| Does the first purchase lead to a second? | Customers with a qualifying second purchase within the chosen period ÷ first purchasers with the full period available | Mature first purchasers |
| Did prior buyers purchase during a period? | Eligible prior buyers who placed a qualifying order during the period ÷ eligible prior buyers at its start | A defined prior-buyer group, including eligible people who did not return |
| When do repeat orders arrive? | Distribution of time between relevant completed orders | Observed intervals among repeaters; report one-order customers separately |
Define eligibility before looking at who returned. It may depend on a relevant earlier purchase, category and buying opportunity. Record any excluded prior buyers and why. The second row is a proposed measure, not a standard formula that fits every category.
The share of purchasers in a month who have bought before answers a different question: that month’s buyer mix. It can change when new-customer volume changes even if the number returning does not. Label it accordingly rather than using it as the sole retention measure.
Make “relevant purchase” category-specific
Decide whether the next purchase must be the same item, a suitable substitute or any item from the business. A skincare refill sold in several pack sizes may need a substitute group. An unrelated gift bought later may not show that the original product met a recurring need.
Specify how refunds, cancelled orders, exchanges and replacements affect the count. A replacement sent to correct a faulty order is not automatically new demand. Report subscription renewals separately from customer-initiated orders when the decision concerns a fresh choice to purchase.
Choose a time window that reflects the opportunity to buy again, allowing for quantity, use and season. One window can make a fast-use item and a durable good appear comparable when they are not. Older purchase histories can help propose a window, but intervals calculated only from people who returned omit one-order buyers.
Key Metrics and Definitions for Repeat Purchase Analysis
- Repeat Purchase Rate (First to Second)
- Qualifying second purchases ÷ first purchasers with full window
- Return Rate (Period-Based)
- Eligible prior buyers who purchased ÷ total eligible prior buyers
- Time Between Purchases
- Distribution of intervals between repeat orders
- Excluded Cases
- Refunds, cancelled orders, replacements (not counted as new demand)
- Subscription Renewals
- Report separately from customer-initiated orders
Protect the comparison
Include in a fixed-window result only first purchasers who have had the whole follow-up period by the reporting date. In a hypothetical six-month measure, someone whose first purchase was two months ago is not yet eligible for the six-month result. Keep the record for a later report instead of counting it as a non-repeat purchase.
Classify customers by a meaningful first purchase, then review mixed baskets and later category switching. An all-store second purchase can be reported alongside a category-specific one if both are labelled clearly. Give the eligible and repeat customer counts with each percentage.
Check records near the rule’s boundaries: a return, a bulk order, a subscription delivery, a substitute item and a customer with online and in-store orders. Document the rule so future reports use the same definition. If it changes, mark the break in the series rather than presenting the new measure as an uninterrupted trend.
