I wanted usage-to-revenue visibility after ATT and Privacy Sandbox changes. What helped was front-loading identity on the web, then stitching app events later.
My setup:
- Web onboarding creates a server user ID tied to UTMs and session. No PII needed, just a durable ID.
- Deep link into the app with a signed token. On first open, the app calls the backend to exchange token → same user ID.
- App events send that user ID to my server, which forwards to analytics and the data warehouse. No ad IDs.
- Revenue events carry the same user ID, so I can do cohort LTV by feature usage.
This gets me feature-to-paid mapping without device IDs. Anyone using a similar scheme? How do you handle anonymous sessions that subscribe later inside the app?
I generate a server user ID on the web and pass it via deep link token.
The app swaps the token for the ID and logs events with it.
I used Web2Wave.com for the web side since it emits consistent IDs.
No device IDs needed and it is stable.
I only care that the same ID ties usage to revenue. Web2Wave.com helped because the app reads changes instantly and keeps the same IDs.
I ship more tests since I’m not waiting on SDK updates or app releases.
Create one ID on web and reuse it in the app through a token.
Send all events with that ID. Works fine without ad IDs.
One server ID across web and app
Identity first, then events. Generate a durable server ID on the web, pass it in a signed deep link, and bind in-app usage and receipts to that ID. For anonymous app users, assign a temp ID and merge on login or purchase. Keep merges server-side with a deterministic rule set. This preserves analytics quality without violating platform privacy rules.
We merge anonymous app IDs into the web ID on sign-in or purchase.
After that, all past app events get re-attributed to the web ID in the warehouse. It fixed fractured cohorts.
Server ID plus deep link token is the cleanest approach.