MOBADLY
EnglishEN ▾
English Русский 中文
// Advertisers

SKAN attribution for iOS app campaigns

Since Apple's ATT changes, SKAdNetwork (SKAN) is the only sanctioned way to measure iOS app install campaigns without the IDFA. This page explains how SKAN attribution actually works, what it can and cannot tell you, and how to plan app acquisition spend in a privacy-first measurement world.

Start advertising Advertisers
▸Operating since 2019▸180+ GEOs, every tier▸MMP-verified reporting▸CPI, CPA, CPL, hybrid & rev share

What SKAN is and why it exists

SKAdNetwork is Apple's privacy-preserving attribution framework for iOS. When a user installs your app after seeing an ad, Apple — not the advertiser — validates the install and sends a postback to the ad network. No device identifier changes hands; Apple confirms the conversion without revealing which user converted. This design is the point: measurement without tracking. For advertisers the practical consequence is a fundamental shift — from user-level data to campaign-level signals, from real-time optimization to delayed, aggregated, thresholded reporting. SKAN is not a degraded version of the old attribution; it is a different system that rewards different planning.

Conversion values: the signal you control

SKAN's core mechanism is the conversion value: a number (0-63 in SKAN 3, or coarse/medium/fine in SKAN 4) that your app reports in the first days after install to encode what the user did. The mapping is your design decision — you might encode registration, tutorial completion, first purchase, or a revenue bucket. This is the only behavioral signal that flows back to the ad network, so its design matters enormously. A conversion value schema that encodes only 'installed' tells the network nothing to optimize toward. A schema that encodes early revenue proxies — first purchase, subscription start — lets network algorithms find users who pay, not just users who install.

Postbacks, timers and delays

SKAN reporting is deliberately delayed and noisy. Postbacks arrive after a timer window (24-48 hours minimum, longer in SKAN 4's multiple windows), with random delay added, and only when a privacy threshold of conversions is met — campaigns with too few conversions get no postback data at all. The practical impact: day-one ROAS is unmeasurable by design, small test budgets produce no signal, and creative-level granularity disappears below threshold volumes. Budget planning must account for this — SKAN campaigns need enough daily conversions per ad set to clear thresholds, which means consolidating rather than fragmenting spend.

What SKAN cannot tell you

Be clear about the limits, because planning around them is the job. No user-level journeys — you cannot see that this user saw three ads before installing. No precise real-time optimization — delays and thresholds prevent it. Limited re-engagement — SKAN is built for install attribution, not retargeting measurement. And no cross-device view. What fills the gaps: on-device measurement via your MMP's SKAN tooling, incrementality testing for true lift, and blending SKAN data with your own first-party cohort analysis. Advertisers who accept SKAN's frame and design for it outperform those waiting for the old world to return.

How we run SKAN-compatible campaigns

Our iOS campaigns are built for SKAN from the start: SKAN-compatible inventory and ad networks, conversion value schemas designed with you to encode your revenue proxies, budget structures that clear privacy thresholds, and reporting that blends SKAN postbacks with your first-party cohort data. We do not promise user-level iOS attribution — nobody can deliver it honestly. What we deliver is measurable, privacy-compliant acquisition that your MMP can verify, and honest counsel about what the numbers do and do not mean.

FAQ

Q.What is SKAN in simple terms?

Apple's framework that confirms your app installs to ad networks without revealing which user installed. It enables campaign measurement without device tracking — the only sanctioned method on iOS since ATT.

Q.What is a conversion value?

A signal your app reports after install encoding what the user did — registration, purchase, subscription. It is the only behavioral data that flows back, so its design determines whether networks can optimize toward revenue or just installs.

Q.Why is SKAN data delayed and sometimes missing?

By design: postbacks have timer windows and random delays, and only fire when enough conversions clear a privacy threshold. Small campaigns may get no data. Consolidate spend to clear thresholds.

Q.Can I get user-level attribution on iOS?

No — not honestly, not since ATT. Anyone promising user-level iOS tracking is describing a workaround that violates platform policy. Plan for campaign-level signals plus your own first-party cohort data.

Q.What is the difference between SKAN 3 and SKAN 4?

SKAN 4 adds coarser conversion values, multiple measurement windows (up to 35 days), and more flexible postbacks — better long-term signal, still privacy-thresholded. Your MMP handles the mechanics.

Q.How do I measure ROAS with SKAN?

Through conversion values encoding revenue proxies, plus blended measurement: SKAN postbacks for network optimization, incrementality tests for true lift, first-party cohorts for revenue truth. Day-one ROAS per user is gone by design.

Q.Does Mobadly run SKAN campaigns?

Yes — our iOS inventory and tracking are SKAN-compatible, and we design conversion value schemas with you. We report what SKAN honestly shows and blend it with your first-party data.

Q.What should I prepare for SKAN campaigns?

An MMP with SKAN support, a conversion value schema mapped to your revenue events, budget consolidated enough to clear thresholds, and realistic expectations about delay and granularity.

Start advertising