Repeat-Purchase Metrics for DTC Brands
TL;DR: A useful retention metric needs a clear customer group and a fixed time window. Repeat-purchase rate tells you how many return; time to second order tells you how quickly. Compare cohorts with enough observation time, not all customers mixed together.
How do you calculate repeat-purchase rate?
For a first-order customer cohort, divide the number who place another eligible order within the observation window by the number in that cohort. Define canceled, refunded and subscription orders consistently.
Illustrative example, not a Mezzori client result: if 200 first-time customers have a full 90-day observation period and 40 order again within it, the cohort repeat-purchase rate is 40 / 200 = 20%.
Why does observation time matter?
A customer acquired yesterday has not had the same chance to reorder as someone acquired three months ago. Compare cohorts at the same age and choose a window that fits the product purchase cycle.
What does time to second order tell you?
It describes the delay between first and second purchases for customers who return. Read it alongside the share who return: a faster second order among a small group does not mean the whole cohort improved.
How should you read revenue per customer?
Use a defined period and consistent revenue treatment. State whether revenue is gross or net of refunds and discounts. Revenue is not contribution margin; fulfillment, product and channel costs still matter.
Can these metrics prove an email caused the order?
No. Cohort trends provide business context. Read Klaviyo attribution and incrementality for causal measurement, or return to the retention measurement guide.