Mobile Measurement

SKAdNetwork & iOS ATT · Mobile Attribution After App Tracking Transparency

How Apple's App Tracking Transparency and SKAdNetwork reshaped mobile measurement starting in 2021 — and how mobile marketers actually attribute installs and post-install events in 2026. The framework, the workarounds, and the strategic implications.

Attribution. App Tracking Transparency (ATT) was introduced by Apple in iOS 14.5 (April 2021). SKAdNetwork is Apple's privacy-preserving alternative attribution framework, now in its fourth major version (SKAdNetwork 4 launched 2022). AdAttributionKit (2024) extended the approach. This article synthesizes the public documentation and operator experience.

What changed in 2021

Before ATT, mobile attribution worked the way web attribution did. The IDFA (Identifier for Advertisers) — a device-level identifier — was readable by apps and ad networks. Mobile Measurement Partners (MMPs) like AppsFlyer, Adjust, Branch, Kochava, and Singular used the IDFA to attribute installs to the ad networks that drove them.

In April 2021, Apple required apps to ask user permission via the ATT prompt before reading the IDFA. Opt-in rates settled around 21–25% globally (with significant variance by region and app category). For 75–80% of iOS users, the IDFA became unavailable.

This broke pre-existing measurement assumptions across mobile marketing.

What SKAdNetwork provides

SKAdNetwork is Apple's alternative attribution framework. It works without the IDFA but with significant constraints:

AdAttributionKit · the 2024 successor

Apple introduced AdAttributionKit in 2024 as an updated framework that extends SKAdNetwork concepts to support more flexibility:

The basic privacy-preserving architecture remains: aggregated, delayed, limited conversion values, crowd anonymity.

What mobile marketers actually do in 2026

The honest practitioner stack today:

  1. SKAdNetwork / AdAttributionKit as the primary iOS attribution source. Imperfect but compliant.
  2. MMPs as integration layer. AppsFlyer, Adjust, Branch, Kochava, Singular all handle SKAdNetwork postback parsing, conversion value modeling, and cross-network reporting.
  3. SSOT (single source of truth) probabilistic models. When deterministic attribution isn't available, MMPs blend SKAdNetwork data with their own probabilistic models for richer reporting.
  4. Media Mix Modeling (MMM). For incremental lift and channel-level decisions, MMM becomes more important when user-level attribution is degraded.
  5. Geo holdout testing. Incrementality measurement that doesn't depend on cookies or IDFA at all.
  6. Web-to-app attribution via deep linking. When the journey crosses web and app, deep link platforms (Branch, AppsFlyer OneLink) help preserve attribution.
The strategic implication. Pre-ATT mobile marketing was deterministic-attribution-heavy. Post-ATT mobile marketing is multi-method: aggregated platform attribution + MMP modeling + MMM + incrementality tests. Teams that pick one and refuse the others have incomplete picture. The competent operator triangulates.

Android is different (but converging)

Android's GAID (Google Advertising ID) is still readable for most users, though Google's Privacy Sandbox for Android (Attribution Reporting API) is the planned successor. As of 2026, the timeline for full deprecation remains uncertain, but the direction is clear: Android attribution will move toward privacy-preserving aggregation similar to iOS over the next several years.

Related on RGM

Sources & further reading
  1. Apple SKAdNetwork developer documentation — developer.apple.com/documentation/storekit/skadnetwork
  2. Apple AdAttributionKit documentation — developer.apple.com/documentation/adattributionkit
  3. Apple App Tracking Transparency framework documentation.
  4. AppsFlyer, Adjust, Branch, Kochava, Singular — MMP documentation on SKAdNetwork support.
  5. Mobile Marketing Association (MMA) measurement working group publications.
  6. RGM operator notes — mobile attribution engagements 2022–2026.