Server-side identity
Web subscribers identify via session and auth, not IDFA. No ATT decay, no SKAdNetwork aggregation.
Optimise
Test. Measure. Attribute.
Your measurement stack continues when subs move to web — per-subscriber experiments, analytics tied to billing, MMP integration, attribution that survives ATT.
Per-subscriberServer-sideMMP-ready
Web rails
What web rails unlock for measurement
When subscribers flow through Recurr's web surfaces, the measurement substrate underneath looks different from in-app. Three structural shifts that shape the rest of the page.
Billing-tied events
Every subscription event lands on your backend through Recurr's pipeline. No separate App Store Server API integration.
Web-surface variants
A/B test on Recurr's surfaces — checkout, sign-up, email. No app release needed.
Dashboard
What you'll see
A representative view from the Recurr dashboard — cohort math, experiment outcomes, and attribution events surfaced in one place. Sample data shown.
dashboard.recurr.dev/insights
Subscriber insights — Q3 2026
Last 90 daysActive web subs
47,328
+1,847 last 30d
MRR
$487K
+9.4% vs prior month
Net retention
94%
+2pp vs store cohort
LTV avg
$312
+$48 post-migration
Net retention lift — vs store baseline
+2pp avgWeb cohort holds consistent positive lift over store baseline. Recovery layer and annual nudge motion compound month-over-month gains.
Retention tables stream to BI
BigQuerySnowflakedbt
Experiments
Experimentation on your web surfaces
Per-subscriber experimentation on Recurr's web surfaces — checkout, sign-up, email. The same testing your team runs in-app, extended to web.
A/B testing
Variants assigned per subscriber at sign-up. Split across offers, copy, or sequencing on Recurr's web surfaces — checkout, sign-up, email. No SDK, no release cycle.
Pricing experiments
Test plan tiers, billing intervals, and promotional structures at the subscriber level. Revenue, churn, and payback per variant are measured directly.
Analytics
Subscription analytics, web-side
Standard subscription analytics on the web book — cohort retention, like-for-like comparisons against store cohorts, GAAP / IFRS revenue recognition. Numbers come from billing data directly.
Cohort analytics
Slice the web subscriber base by acquisition date, source, plan, region, or custom attribute. Retention, LTV, churn drivers — straight from billing data.
Like-for-like measurement
Migrated cohorts and store-billing cohorts compared on the same axes — retention, LTV, churn. Measure the migration lift directly against your pre-migration baseline.
Revenue recognition
GAAP and IFRS revenue recognition schedules on the web book. Deferred revenue, recognition cuts, refund adjustments flow into your accounting system. Period reports sync on close.
Attribution · the migration unlock
Attribution that survived ATT
On store rails, in-app attribution lost most of its fidelity to ATT and SKAdNetwork. The migration is the precondition: once subscribers are on web rails, identification is deterministic, server-side, and per-subscriber.
Default
In-app (ATT)
App installcampaign-only
ATT promptopt-in
IDFA assignedwhen allowed
In-app purchasestore-side
SKAdNetwork postbackbucketed
MMP receivesrandomised timer
Recurr
Server-side web
Web sign-upsession
Subscriber identifieddeterministic
Subscription eventserver-side
MMP webhookreal-time
Per-subscriber recordend-to-end
Migration moves subscribers off store rails — that's what makes the right column reachable. The measurement layer is downstream of the rails.
Server-side conversions survive the ATT gate
Only 46% of iOS non-gaming app users opt in to App Tracking Transparency. Channel-level attribution on store rails depends on the minority that do.
AppsFlyer
ATT data · Q1 2024 · non-gaming apps
Integration
A pipeline your data team can build on
Webhook events you can verify, retry safely, and backfill on demand. Recurr's pipeline behaves like a stream your warehouse and downstream tools can trust as a source of record.
Signed webhooks
HMAC signatures per delivery. Your endpoint verifies origin before processing — no spoofed events make it into your warehouse.
Idempotent retries
Stable event IDs, retries on 5xx with exponential backoff. Downstream dedup is straightforward — same event ID, same outcome, no double-counting.
Versioned schemas
Event payloads are versioned. New fields land additively; deprecations ship with a lead time — your warehouse parsers don't break when Recurr evolves the schema.
Roadmap
Historical replay
Replay any window through the API. Backfill a new warehouse, recover from a downstream incident, run an audit — without re-running production.
One signed webhook stream lands across your existing data stack — BI, product analytics, and MMPs.
Encrypted in transitSubscriber emails stay privateGDPR-aligned
Apply
Pilot on your cohort
See the visibility return on your numbers. The pilot runs against your existing cohort — measurement layer included by default.