Defining purchase, refund and repeat-customer events

Define purchases, refunds and repeat customers from the store’s transaction history before using those labels in an analytics report. Specify when an order qualifies, when an adjustment takes effect and how orders are linked to a customer.

Define a qualifying purchase

Choose the milestone that fits the business question: order placed, payment confirmed or another documented status. Align the order ID, occurrence time, currency and amount definition with that choice. If counting retained orders, specify how later cancellations change the result.

Google Analytics describes ecommerce events such as purchase or refund, and notes that currency is set at event level when value is sent, with items presented as an array. This event format does not make its value equivalent to every store sales total. Shopify sales reports include pending, cancelled and unpaid orders in gross sales and exclude test and deleted orders; gross sales are before taxes, shipping, discounts and sales reversals.

Purchase, Refund and Repeat Customer Definitions Across Platforms

Define an actual refund

Choose the transaction-system milestone at which a refund is processed or posted. Keep a request or approval that has not changed the transaction separate from a completed refund. Distinguish a full refund from a partial item refund, and retain the original order ID and affected item details where available.

Google Analytics describes ecommerce events such as purchase or refund, with currency set at event level for value and items presented as an array. Shopify defines sales reversals as all order adjustments that result in negative monetary value. Do not treat every sales reversal as a return or subtract it again from an amount that already includes it.

Define repeat status from order history

Choose a stable customer key and a qualifying first purchase. Count an order as a repeat purchase when the same resolved customer has an earlier qualifying order. State whether the sequence uses order time, payment confirmation or another milestone. If records have the same timestamp, use a documented tie-break rule.

Keep orders with uncertain identity in an unknown group. An unlinked guest order does not prove the shopper is new. Report repeat orders separately from distinct customers who have repeated: one customer can place several repeat orders. If a customer-type field is sent with a purchase event, derive it from the documented order history.

A current customer segment is not necessarily a historical snapshot. Shopify automatically adds customers when they match segment criteria and automatically removes them when they no longer match. Preserve the classification used for a historical analysis instead of applying today’s membership to an earlier purchase.

Check example records before reporting

Apply the rules to a normal order, a cancelled order, a partial refund, a linked second order and a guest order with uncertain identity. For each, write the expected purchase count, refund amount and repeat classification. The outcome depends on the store’s records and chosen definitions.

Keep the rules beside the report: qualifying statuses, amount calculation, time zone, customer key, unknown-identity treatment and adjustment window. This lets readers explain later changes without changing what the labels mean.