Ecommerce repeat-purchase and retention statistics

Define the event, numerator and window

Use a definition that answers the business question. The table gives working definitions for your own reporting; published benchmarks retain their publishers' definitions. Express each ratio as a percentage by multiplying by 100. Shopify's cohort documentation supports following first-purchase groups, but its returning-customer reports can use order history beyond the selected period.

For follow-up imagery of the same product: build a consistent clothing catalogue.

[7][1][2]

Keep customer identity and purchase events consistent

Count each customer once in a customer-based numerator, using confirmed account or identity links. Treat exchanges, replacement shipments and split fulfillment consistently so operational events do not become extra purchases. Set cancellation and refund rules before comparing cohorts.

For subscription churn, state how pauses, failed payments and cancellations change active status. A missing status is not evidence of retention. Keep loyalty enrollment separate from a qualifying purchase, and leave a metric unavailable when its denominator or follow-up cannot be established.

Prepare a consistent collection for returning buyers: plan useful lifestyle product images.

Match the follow-up to the product

Allow a normal reorder opportunity for replenishable goods. Durable-goods customers may need an accessory or another category long before replacing the original item. A monthly subscription churn definition does not describe those purchases.

Review product accuracy, delivery and the relevance of the next recommendation alongside repeat orders. Report unique returning customers, net repeat-order revenue and contribution separately. A few large orders can raise revenue without increasing the share of customers who return.

The campaign set generator can help plan a coordinated product story; it does not measure customer retention.

What a completed retention experiment specifies

The published coupon trial has a defined first-purchase cohort, a business-as-usual control and a follow-up that extends beyond the incentive. It provides direct evidence about repeat purchases in a non-subscription setting. Product, channel and purchase frequency still limit transfer to an apparel store.

[6]

Build an acquisition cohort with equal follow-up

Group customers by first qualifying order month and keep the customer identifier stable. Choose the follow-up window before examining results, then include only cohorts old enough to complete it. Record a second qualifying order and contribution per acquired customer, including customers who never buy again.

For apparel, separate repeat buying from replacement after a return and exclude cancelled or fully refunded orders consistently. A customer may buy seasonally without declaring a cancellation. Use purchase history and category cadence to decide when a reminder is useful rather than treating silence as confirmed churn.

[7]