What Is an MMP? A Guide to Mobile Attribution
A mobile measurement partner (MMP) collects app events and available campaign signals, then applies attribution rules to report which marketing sources receive credit. It can bring installs, revenue and advertising costs into a common view. Privacy limits, network rules and implementation quality determine what it can observe. Attribution alone does not prove that an ad caused
An app team running several advertising channels soon faces a reconciliation problem: each network reports conversions using its own rules. A mobile measurement partner (MMP) applies a defined set of attribution rules across connected sources so the team can compare
Consider a hypothetical app with 100 first opens. Meta reports 80 conversions and Google reports 60. Adding them to claim 140 new users can double-count people, mix reporting dates or combine different conversion definitions. An MMP helps investigate that overlap. It cannot observe every journey or force every network’s dashboard
What an MMP actually does
| Step | What the platform does | What your team must establish |
|---|---|---|
| Collect app events | A software development kit (SDK) or server integration sends first opens, purchases and other defined events. | Event names, consent behavior, transaction identifiers, currency and duplicate handling. |
| Receive campaign signals | Partner integrations provide eligible ad interactions, referrers or attribution claims. | Which signals the operating system and each network permit. |
| Assign attribution | Rules select a credited source or produce aggregate campaign measurement. | Click and view windows, reinstall and re-engagement treatment, and report type. |
| Combine revenue and costs | Event revenue and connected cost feeds support cohort and ROAS reports. | Gross versus net revenue, refunds, account coverage, timezone and cohort age. |
| Share and inspect data | Reports, exports and permitted event postbacks support analysis and campaign optimization. | Export access, partner mappings, retention and fraud-rejection rules. |
An SDK reporting a purchase does not prove that the ad network received it as an optimization event. A connected network does not prove that all its advertising accounts supply costs. These are separate integrations
Fraud tools can flag or reject suspicious attribution according to configured rules. They do not remove an app from a user’s phone or guarantee that every invalid conversion
How to interpret attribution
Last-click attribution typically credits the last eligible click within a defined window; view-through rules may apply when there is no eligible click. The exact priority depends on the provider, network and report. A list of contributing touches does not necessarily divide conversion credit among them. For example, AppsFlyer’s attribution model distinguishes the credited source from
MMP reports also commonly count an install at first app launch. Store downloads and an advertising platform’s interaction-date reports can therefore differ without an
Attribution answers “which source receives credit under these rules?” A controlled incrementality study asks “what additional outcome did the advertising cause?” Neither high attributed ROAS nor agreement between dashboards answers that second question by itself. ROAS is also revenue divided by ad spend; profit after all costs is a
When comparing cohorts, use the same acquisition period, revenue definition and elapsed age. A seven-day cohort is not comparable to a mature 90-day cohort. Label forecasts separately from observed revenue and retain the data
What privacy rules change
On iOS, Apple’s App Tracking Transparency rules govern tracking across other companies’ apps and websites. An MMP integration does not grant permission to track. Apple prohibits fingerprinting and does not permit an alternative identifier to bypass a denied tracking request. The app developer remains responsible for third-party
AdAttributionKit and SKAdNetwork provide privacy-preserving campaign signals, with limits on timing and detail. They do not reconstruct an unrestricted user-level journey. Apple’s SKAdNetwork documentation now points developers to AdAttributionKit and explains interoperability; confirm support for your particular SDK, operating-system versions and
On Android, the Google Play Install Referrer API can supply referral information for eligible Play installs. Its coverage depends on the installation path. Google’s October 17, 2025 announcement said it would retire specific Privacy Sandbox technologies, including the Attribution Reporting API on Chrome and Android. That announcement did not end all Privacy Sandbox work or remove other
Which MMP belongs on your shortlist?
The table identifies areas described in each vendor’s documentation. We have not conducted a comparative accuracy test for this guide. Check the actual networks, SDKs and contract package
| Provider | Product area to examine | Useful demonstration request |
|---|---|---|
| AppsFlyer | Partner attribution, event sharing and cohort reporting | Trace one test purchase through the SDK, attribution report and partner mapping. |
| Adjust | Mobile attribution, reporting and fraud controls | Show how accepted and rejected attribution appears in an export. |
| Branch | Attribution alongside linking and web-to-app journeys | Demonstrate the destination and measurement when the app is installed and when it is absent. |
| Singular | Cost aggregation alongside mobile attribution | Reconcile a campaign’s normalized cost with its original network report. |
| Kochava | Measurement across app and other marketing channels | Identify which cross-channel results are observed, modeled or unavailable. |
| Tenjin | Mobile attribution and monetization reporting | Show the revenue sources and export access included in the proposed plan. |
Pricing needs a billing definition. On October 9, 2026, AppsFlyer advertised 12,000 free first-year conversions for Growth, followed by US$0.07 per billable conversion. That allowance belongs to Growth, not its Zero owned-media plan. Tenjin listed 2,000 conversions per month on its free plan, with US$0.04 per additional conversion. The allowances use different billing arrangements. Compare complete quotes for your expected usage. Ask about re-engagement charges, exports, cost integrations, support and
When the investment makes sense
An MMP becomes useful when decisions depend on comparing several paid networks, connecting acquisition to revenue, reconciling conflicting reports or sending consistent conversion events to partners. A small app still validating its core experience may first need reliable product events and subscription records. Buying attribution software will not repair a missing
Before committing, run a bounded pilot with your development team. Trace a first open, purchase, renewal, refund and reinstall where relevant. Document the expected attribution and revenue treatment for each. Reconcile one cohort with store or billing records and network costs; explain differences instead of forcing the totals to match. A short pilot can validate data flow, but cannot establish long-term retention or
Our fitness-app case study describes acquisition and app store optimization (ASO) work together. It is useful context for connecting measurement to business decisions. It does not show that one MMP caused the
If you need help deciding whether your reports support a budget decision, start with measurement and data analysis. Once events, costs and reporting limits are understood, paid acquisition management can use that evidence to plan and