MMP attribution
App attribution your team can actually trust
ADEX configures and validates AppsFlyer, Adjust, Singular or Branch — event mapping, SDKs, postbacks, SKAN and raw-data export to BigQuery — so paid channels optimize on verified events instead of self-reported network numbers.
The problem
Every ad network reports its own installs and conversions, and each one is incentivized to overcount. Optimize to self-reported network data and you overpay the channels that are best at claiming credit, not the ones that actually drive growth.
The technical layer makes it worse: unmapped events, duplicated SDK and server signals, misconfigured postbacks and a half-set-up SKAN schema all distort the picture quietly, until a payer number does not match the backend and nobody knows why.
How ADEX runs it
- 01
Event mapping
Define every meaningful event — install, activation, purchase, subscription — and map it consistently across app, SDK and backend so the taxonomy is unambiguous.
- 02
SDK validation
Verify the MMP SDK fires correctly by OS, event and edge case, and remove duplicate or conflicting signals.
- 03
Postbacks and networks
Configure postbacks and ad-network integrations so each channel receives the right signal to optimize on — nothing more, nothing missing.
- 04
SKAN and iOS
Set up a SKAN conversion-value schema that captures what matters on iOS without leaking privacy thresholds, and reconcile it with deterministic data where available.
- 05
Raw-data export
Export raw-level attribution data to BigQuery so the app data joins the warehouse and can be validated against CRM and backend truth.
Proof
- ADEX is certified across all four major MMPs — AppsFlyer, Adjust, Singular and Branch.
- Attribution flows straight into BigQuery dashboards, so app events reconcile with channel spend and backend revenue.
- The same validation discipline underpins ADEX MMP migrations — 12+ completed.
Setup complexity and timeline depend on your app count, event volume, ad networks and SKAN requirements. iOS attribution is probabilistic by design; we are explicit about what can and cannot be measured deterministically.
What you get
Configuration
- Event taxonomy and mapping
- SDK integration and validation
- Postback and network setup
iOS / privacy
- SKAN conversion-value schema
- Deterministic vs modeled reconciliation
- Privacy-threshold handling
Data
- Raw-data export to BigQuery
- CRM / backend reconciliation
- Validation report
When this is right
- You run an app and optimize paid channels on self-reported network data.
- Your MMP numbers do not reconcile with the backend or CRM.
- You are scaling app user acquisition and need measurement you can defend.
When it isn't
- You have no app — this is mobile measurement, not web analytics.
- You are pre-launch with almost no events — a lightweight setup is enough for now.
- You expect deterministic iOS attribution — post-ATT, some of it is modeled by design.