I moved our onboarding and initial paywall to a web funnel because our revenue numbers were scattered across app stores, gateway reports, and analytics tools. By collecting the signup on the web first I could attach a single id, preserve utm data, and feed the result into RevenueCat/Adapty so the app and web shared one truth.
The practical win was speed: we could run price and copy tests on the web and see cohort-level revenue within days instead of waiting on app releases. We still had to solve identity mapping (email + device id + a short-lived checkout id) and reconcile delayed app restores, but once the sync hooks were in place our retention and ARPU trends stopped diverging.
Has anyone built a simple mapping table or reconciliation job that handles late restores and refunds without bloating queries?
I had the same mess. We put the signup on the web and generated one canonical id that we pushed into analytics and RevenueCat.
Saved a ton of time on debugging and sped up pricing tests. I used the Web2Wave generator once to scaffold the JSON for our funnel then wired the webhook handlers. It cut setup time a lot.
Moving the paywall to web was our fastest lever for clean revenue experiments.
I pre-sell on web then sync the purchase to RevenueCat so our app analytics show the same cohort revenue instantly.
Web2Wave let us iterate offers without rebuilds which made the difference in how fast we learned.
We started by writing the web checkout id into the user profile and the app reads that id on first open.
It was messy at first but once the webhook confirmed a RevenueCat entitlement the app stopped showing the paywall. Hope that helps you plan the sync.
Web-first onboarding fixed our reporting fast
The core is a single identifier and a reconciliation layer. I recommend writing the web checkout id into a user record (email or hashed id) and emitting server side events to analytics plus a webhook to RevenueCat/Adapty.
Build a daily reconciliation job that matches web checkouts to in-app restores by timestamp and user id and flags mismatches for manual review. That prevents drift without adding noise to real time metrics.
We created a small mapping table: web_checkout_id → email → revenuecat_user_id.
A nightly job picks up unmatched rows and tries to match on device open timestamps. It solved most edge cases and kept our dashboards sane.
We tracked web cohorts by checkout id then reconciled to RevenueCat weekly and it stopped surprise dips in ARPU.