Refund Rate: Share of Paid Transactions That Get Refunded

Each paid transaction, meaning a purchase, a renewal or a one-time purchase, is placed in the period of its purchase date. A transaction counts as refunded when the ledger holds a refund for it that was not reversed. The refund can have happened on any later date, up to now.

That is different from the Refunds chart, which places a refund on the day it happened. Here the refund goes back to the period of the purchase it belongs to. So the rate for a recent period can still rise as refunds arrive, and old periods settle. Free trial starts are not paid transactions and are excluded, and so are sandbox transactions, granted access and Family Sharing.

Filters and segments: app, store, product, product duration, offering, country, platform and app version. You can open it in the dashboard under Charts or fetch it with GET /v2/projects/{project_id}/charts/refund_rate. The response has the same shape as RevenueCat's Charts API, so scripts written for that API keep working. The secret key needs the charts_metrics:charts:read permission. The Charts guide lists every parameter.

The paid transactions of June, checked in September.

June's refund rate is 40 / 2,000 = 2.0%. If 6 more June purchases are refunded in October, the same period will read 2.3%.

A rising refund rate means revenue you already counted is going back to customers. It is also a quality signal. Refunds cluster around a bad release, a misleading paywall, a confusing price after a trial, or a plan sold in a market where the price surprised people.

Segment by product and offering to find which plan or paywall produces refunds, and by store to compare Apple and Google. A refund rate that is high for annual plans and low for monthly plans usually means customers noticed the charge too late.

How to read it

• Judge a period only after refunds have had time to arrive. Refunds can come weeks after the purchase.
• Compare the same age: a rate for last week is not comparable to a rate for last quarter.
• Use Refunds for the money and the count by refund date, and use this chart for the rate by purchase date.
• A reversed refund is removed, so a customer who wins a dispute no longer counts.
• Pair with Refund Request Outcomes to see how Apple requests turn into refunds.