Give your data team clean subscription data, not data fires
Subscription revenue is easy to report on one platform and hard to reconcile across several.
A data layer that normalizes store events into one schema removes most of that reconciliation work before it reaches your warehouse. This guide explains what such a layer delivers, how its webhook event types differ, and how to choose between webhooks, scheduled exports, and pre-built integrations. Capability statements below describe the vendor product captured in the source material; confirm your own provider offers them before relying on them.
What a subscription data layer actually delivers
A cross-platform subscription data layer sits between the app stores and your own systems. Instead of stitching together App Store, Google Play, and web purchase records yourself, you receive one normalized stream of purchase and subscription events.
In the captured vendor product, the backend verifies, parses, and validates receipts against each customer, then keeps subscription status continuously updated through its API and SDK. Your own server can still run alongside it and stay synced to each user's status — the difference is that receipt verification and status upkeep are handled upstream rather than rebuilt in-house.
On the client side, the SDK fetches product information from the stores and customer subscription data from the backend, manages purchase logic, and syncs receipts and attribution data. That means the events your data team receives already reflect store-side truth rather than a client-side guess. Treat these as vendor-described capabilities to verify against your chosen provider, not as guarantees of any specific deployment.
Webhook event types and how to tell them apart
Webhooks notify you whenever an event happens in the app and let you monitor state changes for subscribers. The event names below are the exact identifiers you will see in payloads; treat them as the canonical field values when building warehouse models or alerting rules.
- Test — a test event issued through the dashboard, used to confirm your endpoint is reachable.
- Initial_purchase — a new subscription has been purchased.
- Non_renewing_purchase — a purchase that will not auto-renew.
- Renewal — an existing subscription renewed, either at the end of the current billing period or later when a lapsed user re-subscribes.
- Product_change — the subscriber switched to a different product.
- Cancellation — a subscription or non-renewing purchase was cancelled.
- Uncancellation — a cancelled but non-expired subscription was re-enabled.
- Billing_issue — a charge attempt failed. This does not mean the subscription has expired.
- Subscription_paused — the subscription was paused.
- Expiration — the subscription expired and access should be removed.
The distinction that most often trips up reporting is Billing_issue versus Expiration. A billing issue signals a failed charge while access may still be live; expiration is the point where access should be revoked. Model them as separate states, not as one "churn" flag, or your active-subscriber counts will drift.
Three delivery paths: webhooks, scheduled exports, and integrations
The same normalized data can reach your stack through more than one channel, and the right choice depends on latency and shape rather than on completeness.
- Webhooks push events server-side in near real time to your cloud or marketing stack. Best when downstream systems need to react to a state change quickly.
- Scheduled data exports automatically deliver all of your apps' transaction data to various cloud storage providers. Best when your warehouse needs a full historical table rather than a stream of deltas.
- No-code integrations connect customer lifecycle events and revenue data to the tools your teams already use, including analytics, attribution, messaging, and support platforms.
A practical pattern is to use exports for the durable warehouse table and webhooks for operational triggers, then let integrations handle the tools that do not need custom plumbing. These channel descriptions reflect the vendor product captured in the source; confirm availability with your own provider.
Checklist before you wire anything up
Use this list to decide which channel fits each downstream consumer:
- Does the consumer need to react within seconds? If yes, start with webhooks.
- Does the consumer need a complete, replayable history? If yes, start with scheduled exports.
- Is the destination a standard analytics, attribution, messaging, or support tool? If yes, check for a pre-built integration first.
- Will you model subscription state yourself? If yes, confirm you are consuming the normalized event names rather than raw store payloads.
- Do you need per-customer debugging? If yes, plan for a customer-level view of cross-platform usage and the purchase timeline.
- Have you separated billing_issue from expiration in your schema? If not, fix that before launch.
Worked example: reconciling a renewal and a product change
Suppose a subscriber on a monthly plan upgrades mid-cycle. Your endpoint receives a Product_change event, and at the end of the period a Renewal event fires for the new product. If you count every renewal as a new subscription, your new-subscriber metric will overstate growth.
A cleaner model keeps three facts separate: the customer identifier, the product identifier attached to each event, and the event type. Product_change updates the product dimension without incrementing new subscriptions; Renewal increments renewal counts against whichever product is current. If a charge later fails, a Billing_issue arrives and access stays live until an Expiration event confirms the end. That sequence lets you report active subscribers, at-risk subscribers, and expired subscribers as three distinct cohorts rather than one blended number.
Where each team plugs in
The same normalized source supports several functions at once. Marketing gets deep analytics and a cross-platform source of truth for customer data. Data teams are freed from ad-hoc requests, troubleshooting, and re-checking accuracy. Product teams get clean cross-platform data to guide the roadmap. Support can debug issues and grant exceptions using a customer-level view of usage and purchase history. Engineering spends less time on in-app purchase plumbing and more on features.
When connected data is trustworthy, the time spent troubleshooting and managing it drops, and more of the team's time goes to analysis that changes decisions.
Frequently asked questions
Does a billing issue mean the subscription has expired?
No. A billing issue means a charge attempt failed; the subscription has not necessarily expired. Access should be removed on an expiration event, which is a separate event type.
Can a renewal event fire after a lapse?
Yes. A renewal can occur at the end of the current billing period, or later if a lapsed user re-subscribes.
What is the difference between uncancellation and a new purchase?
Uncancellation re-enables a cancelled subscription that has not yet expired. It is not a new purchase, so it should not be counted as an initial purchase in acquisition reporting.
Should I use scheduled exports or webhooks?
They serve different needs. Scheduled exports deliver all transaction data to cloud storage on a schedule, which suits a durable warehouse table. Webhooks push events server-side in near real time, which suits operational triggers. Many teams use both.