Send in-app purchase events to Statsig
The Statsig integration forwards billing, subscription and revenue activity from iOS and Android into a Statsig project, so experiment and feature-gate analysis can use purchase outcomes instead of app activity alone.
This guide explains how to turn its event list into a working experiment plan.
What the integration is for
The integration serves three practical jobs. Revenue impact: judge a product change against business metrics rather than engagement alone. Feature control: support feature gates and A/B tests for new capabilities. Post-session visibility: cancellation and renewal signals can arrive even when the user does not reopen the app, which matters for cohorts whose last session was weeks ago. Whether these hold in your project depends on your own configuration and verification.
A practical setup checklist
Before enabling the integration, give each event you care about a clear role in your experiment design:
- Pick one primary outcome per experiment (for example trial conversion or trial cancellation) and list the secondary events you will watch.
- Decide in advance whether a billing issue should count as a warning signal or as churn. The source is explicit that a billing issue does not by itself mean the subscription expired when a grace period applies.
- Confirm how product changes are handled in reporting, since a product change does not mean the new subscription is in effect immediately.
- Verify that renewal and cancellation signals are visible for users who do not reopen the app.
- Write down which events are excluded from the analysis and why, so a later reader can reconstruct the decision.
Reading the lifecycle events
The event set separates acquisition, trial stages, ongoing renewal status and recovery. A first paid auto-renewing subscription without a free trial is distinct from a one-off product purchase. Trial started, trial converted and trial cancelled mark the three points where a free trial can change state. Renewal covers both a normal renewal and a repurchase after a lapse, while cancellation and uncancellation track auto-renew status during the paid period. Billing issues, expiration and product change cover payment failure, loss of access and plan switching. The table below pairs each event with a concrete check to run before you rely on it in an experiment.
| Event | Source definition | Check before you rely on it |
|---|---|---|
| Initial Purchase | The first purchase of an auto-renewing subscription product that does not contain a free trial. | Confirm trial users are not double-counted as initial purchases in your cohort. |
| Trial Started | The start of an auto-renewing subscription product free trial. | Decide whether trial entry is a funnel step or an experiment exposure. |
| Trial Converted | When an auto-renewing subscription product converts from a free trial to normal paid period. | Use as the trial-to-paid outcome; pair with Trial Cancelled to see both trial exits. |
| Trial Cancelled | When a user turns off renewals for an auto-renewing subscription product during a free trial period. | Distinguish from paid-period Cancellation when reporting trial churn. |
| Renewal | When an auto-renewing subscription product renews OR a user repurchases the auto-renewing subscription product after a lapse in their subscription. | Separate first renewal from post-lapse repurchase if recovery is a separate goal. |
| Cancellation | When a user turns off renewals for an auto-renewing subscription product during the normal paid period. | Treat as intent to churn, not immediate loss of access. |
| Uncancellation | When a user re-enables the auto-renew status for a subscription. | Pair with Cancellation to measure win-back within the same cohort. |
| Non Subscription Purchase | The purchase of any product that's not an auto-renewing subscription. | Keep one-off revenue out of subscription retention metrics. |
| Billing Issues | There has been a problem trying to charge the subscriber. This does not mean the subscription has expired (in the case of a grace period enabled). | Do not count as churn automatically; check grace-period handling first. |
| Expiration | A subscription has expired and access should be removed. | Use as the access-removal trigger; it is the clearer churn signal than Billing Issues. |
| Product Change | When a user has changed the product of their subscription. This does not mean the new subscription is in effect immediately. | Confirm effective timing in the subscription-management guidance before attributing revenue to the new product. |
Definitions follow the integration's event descriptions. Product-change timing is qualified; billing issues are explicitly not equivalent to expiration when a grace period applies.
Frequently asked questions
What does the Statsig integration send?
The integration automatically sends billing, subscription and revenue metrics from iOS and Android to a Statsig project, so experiment and feature-gate analysis can use purchase outcomes rather than app activity alone.
Can renewal and cancellation data arrive without an app open?
Yes. Cancellation and renewal data can be received even without app opens, which helps keep experiment cohorts current after a user's last session.
Does a billing issue mean the subscription expired?
No. A billing issue means a charge attempt failed. The source notes it does not mean the subscription has expired when a grace period is enabled, so treat it as a payment-risk signal rather than automatic churn.
Does a product change take effect immediately?
Not necessarily. A product change does not mean the new subscription is in effect immediately; see the subscription-management guidance for updates, downgrades and crossgrades.