Mobile attribution and MMPs: the operator's guide

Mobile attribution is a separate discipline from web attribution. App installs, in-app events, and post-install LTV all require purpose-built tooling — Mobile Measurement Partners (MMPs) like AppsFlyer, Adjust, Singular, and Branch. Since iOS 14 App Tracking Transparency, mobile attribution has become substantially harder; the MMPs are the layer that makes it functional. This is the operator's guide.

By David Schaefer · LinkedIn · Updated May 2026

Why mobile attribution is different

  • App installs happen through app stores (App Store, Google Play), breaking the click-to-conversion chain.
  • iOS App Tracking Transparency (ATT) requires user opt-in for cross-app tracking; 60-80% opt out.
  • Android Privacy Sandbox is gradually tightening Android attribution.
  • Multiple ad networks each report installs differently; you need a neutral source of truth.
  • Deep links require careful implementation for in-app event attribution.

The major MMPs

APPSFLYER LARGEST BROAD NTWK ADJUST ENTERPRISE FRAUD GRD SINGULAR MARKETING IQ ROI FOCUS BRANCH DEEP LINK UNIVERSAL FIG. 01 RGM® · BLUEPRINT

FIG. 01 — MMP landscape

MMPStrengths
AppsFlyerLargest MMP, broadest network integration, mature analytics
AdjustEnterprise-friendly, strong fraud prevention
SingularMarketing intelligence + MMP combined, ROI focus
BranchDeep linking + MMP, strong universal links
KochavaFree tier available, MMP for SMB

SKAdNetwork (SKAN) — Apple's framework

SKAN is Apple's privacy-preserving attribution framework for iOS users who don't opt into ATT. Apple aggregates conversion data at the campaign level and returns postbacks with limited granularity. MMPs handle the SKAN integration, aggregate it with their own probabilistic models, and present unified reporting. SKAN 4.0 (the current version) added more flexibility but remains less granular than the pre-ATT era.

Deep links and universal links

Deep links connect off-app touchpoints (ads, email, SMS, social) to specific in-app destinations. Two implementations:

  • Universal Links (iOS) / App Links (Android). Native OS-level deep links. Strong for ad attribution.
  • Deferred deep links. Handle the case where a user clicks an ad, installs the app, then lands at the deep-linked content on first launch. Critical for paid-attribution accuracy.

Branch is the dominant deep linking specialist; other MMPs also offer deep linking.

How mobile attribution fits the broader stack

  • Pair with GA4 for Apps (Firebase-based) for in-app analytics.
  • Activate audiences via CAPI to Meta, TikTok, and other paid platforms.
  • Connect to lifecycle via push provider — see Flutter for the app-first stack.
  • Cross-reference with marketing mix modeling for full-portfolio attribution.
Do I need an MMP?

If you run paid acquisition for a mobile app, yes. Without an MMP you can't reliably attribute installs across networks. Free tiers exist (Kochava) for small apps.

What does an MMP cost?

Free tiers for small volumes. $5K-$50K/year for mid-market. $100K+ for enterprise. Pricing tied to attributed events and feature tier.

How does SKAN work?

Apple aggregates iOS install conversion data at the campaign level for users opted out of ATT. Returns postbacks with limited granularity. MMPs handle integration and combine SKAN with probabilistic models.

AppsFlyer vs Adjust?

AppsFlyer is the largest and most-integrated. Adjust is enterprise-friendly with strong fraud prevention. Either works for most apps; the choice often comes down to specific ad-network integrations and team preference.

Should I use Branch?

For deep linking + attribution combined, Branch is the specialist. Other MMPs offer deep linking; Branch is the deepest. Choice depends on whether deep linking complexity is a primary need.

What about Android?

Android attribution has been more permissive historically. Google's Privacy Sandbox for Android is gradually tightening. MMPs handle Android attribution similarly to iOS, though without the SKAN restrictions.

Operating checklist

  1. Define the business outcome before opening tools.
  2. Configure measurement and audit baseline.
  3. Onboard data, verify quality and coverage.
  4. Build foundational programs before advanced layers.
  5. Launch controlled; monitor daily.
  6. Refresh quarterly; document for the next operator.