adlibrary.com Logoadlibrary.com
← Back to Glossary

App Tracking Transparency (ATT)

App Tracking Transparency (ATT) is Apple's iOS framework, introduced in iOS 14.5 (April 2021), that requires apps to ask users for permission to track their activity across other apps and websites.

adlibrary ad library ads library meta ads library ai marketing tool 1

Definition

App Tracking Transparency (ATT) is Apple's consent framework, shipped in iOS 14.5 in April 2021, that gates every app's ability to track users across third-party apps and websites behind an explicit opt-in prompt. Before a developer can access the device's IDFA — the identifier Meta, attribution providers, and measurement SDKs used for cross-app targeting — the user must tap "Allow." The majority tap "Ask App Not to Track" instead.

How ATT works at the system level

When an iOS app calls requestTrackingAuthorization(), the OS surfaces a standardised dialog. If the user declines, the IDFA returns a zeroed string and any cross-app identifier matching is blocked. Safari's Intelligent Tracking Prevention (ITP) already limited web-to-app signal; ATT extended that curtain to in-app behavior. Meta's pixel relies on the browser IDFA chain — once that's severed on iOS, ViewContent and Purchase events fired from iPhone users either go dark or degrade to aggregated, modeled output via Apple's Privacy Click Measurement framework.

The downstream impact on Meta delivery

ATT didn't break Meta ads. It broke the deterministic pixel signal that powered Meta's optimisation engine for a decade — specifically the 1:1 user-level match between an iOS browse session and a conversion event. I've seen accounts in DTC verticals where attributed iOS conversions dropped 40–60% in the 60 days after ATT rollout, not because performance fell, but because measurement did. The rebuild required standing up Conversions API (CAPI) alongside the pixel to recover server-side signal that bypasses ATT entirely.

Meta's Advantage+ and Andromeda systems lean harder on aggregated and modeled conversions — attribution windows shrank, modeled fill stepped in for gaps, and SKAdNetwork postbacks became a parallel reporting stream. Rebuilding with CAPI + offline conversions typically recovers 50–70% of pre-ATT measurement fidelity.

Read why ad attribution is hard to track post-iOS and how AI reshapes social media advertising. The first-party data strategy is the natural complement.

The practitioner principle: treat iOS CPA benchmarks before and after April 2021 as different data series — they measure different things.

Why It Matters

ATT broke the deterministic pixel signal that powered Meta optimisation for a decade. Most modern attribution rebuilds — CAPI, modeled conversions, EMQ scoring — exist because of ATT. If you're running Meta campaigns against an iOS-heavy audience and haven't rebuilt with server-side signal, you're flying on a fraction of the data the algorithm needs to optimise effectively.

Examples

  • Industry opt-in rates for ATT have hovered around 20–30% globally, meaning Meta loses deterministic visibility on roughly 70% of iOS users.
  • A DTC brand running iOS-only campaigns rebuilt attribution with CAPI + offline conversions + modeled conversions to recover ~60% of pre-ATT signal accuracy.
  • Pre-ATT CPA benchmarks should not be compared to post-ATT CPA — the signal sources are different.

Common Mistakes

  • Trying to 'fix' ATT-era CPA without acknowledging the signal floor has changed permanently.
  • Skipping CAPI under the assumption pixel still carries the load; on iOS, server-side signal often carries more weight than pixel.
  • Reading short attribution windows post-ATT and concluding ads do not work — under-attribution is the default state.