Docs

AppsFlyer's PBA Retirement: What to Track

AppsFlyer is retiring People-Based Attribution for Web Attribution by end of 2026 — track these signals before your web-to-app funnel data breaks.

Acquisition

AppsFlyer is retiring People-Based Attribution (PBA), its long-standing web measurement product, and moving every account onto Web Attribution — a newer solution built on the same core attribution engine but with a different data model, a different SDK, and a different API surface. If your team measures web-to-app conversion, web-originated purchases, or cross-device journeys through PBA today, this isn’t a routine vendor update you can wait out. Accounts get migrated on AppsFlyer’s timeline, and teams that haven’t validated their event mapping before the cutover tend to discover gaps only after a reporting period already has a hole in it — a week or a month of web attribution data that doesn’t reconcile with what came before.

The risk isn’t that Web Attribution is worse than PBA — early accounts report it’s more accurate, with tighter identity resolution and a cleaner server-to-server path. The risk is silent parity loss: a custom event that fired under one name in the PBA SDK integration and a different name (or not at all) under the new one, a URL parameter your paid channels rely on that Web Attribution parses differently, or a conversion window default that shifts your last-click attribution without anyone changing a setting. None of that throws an error. It just quietly changes the numbers your growth team is making budget decisions on.

Data Points to Track

  • Event name parity between the legacy PBA SDK and the new Web Attribution SDK, for every custom event currently wired into dashboards or automated reports
  • Web-to-app deep link conversion rate, tracked separately before and after migration so a drop is attributable to the switch, not seasonality
  • Attribution window and lookback settings, confirmed to match your prior PBA configuration rather than defaulting to Web Attribution’s out-of-the-box values
  • Server-to-server (S2S) call success rate, since Web Attribution’s data path runs through the Web S2S API rather than PBA’s client-side model
  • Channel-level click-to-install rate, split by paid source, to catch a URL-parameter parsing difference before it distorts spend allocation
  • Duplicate or missing conversions during the overlap window if both PBA and Web Attribution are live simultaneously for any account

Setup Steps

  1. Inventory every custom event and URL parameter your current PBA integration passes into AppsFlyer, before touching the new SDK.
  2. Stand up Web Attribution in parallel using the Web S2S API, rather than switching over in place, so you get a real overlap window to compare output.
  3. Map each legacy event to its Web Attribution equivalent explicitly — don’t assume a same-named event behaves identically under the new data model.
  4. Re-set attribution window and lookback parameters to match your existing PBA configuration rather than accepting Web Attribution defaults.
  5. Run both systems for at least one full reporting cycle and reconcile totals — installs, conversions, and channel-level spend efficiency — before decommissioning PBA.
  6. Alert on a sudden drop in web-to-app conversion volume during and immediately after cutover, since that’s the fastest signal something didn’t map cleanly.

Actionable Insights

The overlap-window comparison is where this migration earns or loses trust. If Web Attribution and PBA report materially different totals for the same period, the gap tells you exactly where to look: a channel-level mismatch usually means a URL parameter isn’t being parsed the same way, while an even, sitewide drop usually means an attribution window setting reverted to default. Once parity is confirmed, Web Attribution’s tighter identity resolution should actually improve downstream numbers — fewer web sessions falling back to probabilistic modelling means cleaner channel-level ROAS and a paid budget allocation that reflects what’s actually converting, not what an older, looser attribution model guessed was converting.

Expert help

Need help tracking this in your app?

Our team sets up analytics pipelines for mobile and web teams every day. Talk to us and get your first events flowing in under an hour.

Talk to an expert