SKAdNetwork (SKAN) is Apple's privacy-safe attribution framework that measures iOS app installs and post-install conversions without exposing individual users or the IDFA. Since App Tracking Transparency (ATT) made device-level tracking opt-in, SKAN is how advertisers, ad networks, and measurement partners attribute iPhone and iPad campaign performance. Here is how SKAN 4 works and what performance marketing teams must change to keep measuring iOS growth.
TL;DR: Skadnetwork at a Glance
- SKAdNetwork is Apple's built-in attribution system for iOS; it reports campaign results without user-level identifiers.
- It exists because ATT ended broad IDFA access, so most third-party and platform tracking lost granularity on iOS.
- SKAN 4 sends up to three postbacks with coarse and fine conversion values, plus support for web and omni conversions.
- You encode a 6-bit conversion value (0 to 63) to signal quality, such as revenue tier or retention, to the ad network.
- It is delayed, aggregated, and blind to user journeys, so it cannot replace full funnel or multi-touch attribution on its own.
- Teams adapt by combining SKAN with server-side signals, media-mix modeling, and incrementality testing.
What Is Skadnetwork and Why Does It Exist?
SKAdNetwork is a framework inside iOS that lets an ad network receive a signed notification when a user installs or takes an in-app action after seeing an ad, without revealing who the user is. Apple introduced it ahead of ATT, and it became the default measurement path once ATT made the IDFA unavailable unless a person explicitly opted in.
The design goal is privacy by construction. Instead of a pixel or SDK that follows a person across apps, SKAN keeps attribution on the device and sends only an aggregated, delayed postback to the ad network that served the ad. That protects the user but removes the campaign-level and user-level detail marketers relied on for optimization and attribution. For a broader view of how iOS changed measurement, our notes on Facebook Ads iOS attribution workarounds cover what the platform changes broke.
How Does Skadnetwork Attribution Work Step by Step?
The flow is smaller than a classic pixel and deliberately blind. The standard path looks like this.
- A developer registers their app and ad networks with Apple, and the ad network signs its impressions and clicks.
- A user sees or taps an ad, then installs and opens the app within a defined window.
- The device confirms the install with Apple, which validates the ad signature and attribution window.
- Apple sends a postback to the ad network that won the last click or view, containing the campaign, a coarse or fine conversion value, and a source identifier.
- With SKAN 4, the advertiser can also receive a copy of the postback, and additional postbacks arrive as the user keeps engaging.
Notice what is missing: no user ID, no device graph, no timestamp precision, and no path between the ad and the eventual purchase beyond what you encode into the conversion value. That is the trade Apple made for privacy.
What Changed in SKAN 4 Versus SKAN 3?
SKAN 3 was a single postback with one fine conversion value and limited source data. SKAN 4, released by Apple in 2023, expanded measurement in four ways that matter to marketers.
- Multiple postbacks: up to three postbacks across the early, middle, and later windows, instead of one delayed signal.
- Coarse conversion values: when the install count is low, Apple replaces the fine 0 to 63 value with a low, medium, or high bucket to preserve privacy.
- Web and omni conversions: SKAN can now attribute on the open web and across app and web, not only inside apps.
- Hierarchical source identifiers and crowdsourced learning: more campaign granularity as volume grows, and Apple can surface aggregated learning signals.
The practical effect is that SKAN 4 gives you more than one snapshot of the user's value, but still on Apple's clock and inside Apple's privacy envelope. If you want the full mechanics, Apple's SKAdNetwork developer documentation is the source of truth for each version.
What Is the SKAN Conversion Value and How Do You Use It?
The conversion value is a 6-bit integer from 0 to 63 that you control inside your app. Apple sends it to the ad network in the postback, and it is the only rich signal you get about what the install was worth. Because you have only 64 possible values, the strategy is everything.
Common encodings include:
- Revenue tiers: map 0 to 63 onto increasing purchase value bands.
- Engagement or retention: encode day-1, day-7, or day-30 retention as the value.
- Funnel stage: which onboarding or activation step the user reached.
- Hybrid schemas: split the 64 values across several signals, trading resolution for breadth.
You update the conversion value within a lock window; once locked, later activity cannot change that postback's value. Teams usually run a single schema across the app and revisit it quarterly as priorities shift from installs to activation to revenue.
How Do You Measure Campaigns with SKAN Postbacks and Hierarchy?
SKAN reporting lands in your ad platform's SKAN dashboard, where Apple's hierarchical source IDs map back to campaign and ad-set identifiers as volume allows. At low volume you see coarse buckets; at higher volume you get finer campaign breakdown. This is why creative and audience testing on iOS needs more spend and patience than on Android or web.
A clean operating model is to treat SKAN as one of three inputs, not the whole truth. The other two are media-mix modeling for channel-level effects and incrementality or geo tests for causal lift. Our marketing attribution guide for startups lays out how to combine these into one defensible number.
How Does SKAN Affect Meta, TikTok, and Google iOS Attribution?
Every major channel rebuilt its iOS measurement around SKAN. Meta uses Aggregated Event Measurement and SKAN postbacks with modeled conversions; TikTok and Google Ads similarly ingest SKAN signals and fill gaps with modeling. The shared result is lower match rates, noiser optimization, and a push toward server-side and API-based events.
That is why many teams pair SKAN with a server-side setup. Sending events through a controlled server container rather than only the browser improves signal quality and resilience, and it dovetails with SKAN's limitations. Our server-side tagging guide explains the infrastructure behind that approach.
What Can SKAN Not Measure?
SKAN is deliberately narrow, and the blind spots are real.
| Blind spot | What it means for marketers |
|---|---|
| User-level data | No ability to tie a conversion back to a specific person or device for retargeting as before. |
| Timing precision | Postbacks are delayed and windowed, so you lose hour-by-hour or day-by-day journey detail. |
| Full funnel paths | You see an encoded value, not the steps a user took between ad and conversion. |
| Cross-channel comparison | Each network reports SKAN differently, making like-for-like comparison hard. |
| View-through and assisted touch | SKAN is mostly last-touch by click or view, so assist value is invisible. |
How Should a Performance Marketing Team Adapt to SKAN?
Adapting to SKAN is less about a new tool and more about a new measurement posture. A practical checklist:
- Own your conversion-value schema: decide what 0 to 63 represents and align product and growth on it.
- Move eligible events server-side so signal quality does not depend on the browser alone.
- Run SKAN alongside media-mix modeling and at least one incrementality test per quarter.
- Set platform expectations: iOS optimization will be noiser, so widen learning phases and avoid daily budget thrash.
- Watch creative and product-market fit; with less targeting data, the creative and the offer do more of the work.
- Document a single source of truth that combines SKAN, platform modeling, and your own tests so leadership sees one number, not three.
Frequently Asked Questions
What Is the Difference Between Skadnetwork and the IDFA?
The IDFA was a device-level identifier that let apps and ad networks track a specific person across apps for attribution and retargeting. SKAdNetwork removes that identifier and instead sends an aggregated, delayed, privacy-safe postback to the winning ad network. With ATT, the IDFA is only available when a user opts in, so SKAN is the default measurement path for most iOS campaigns.
Does Skadnetwork Work for Web Campaigns or Only Apps?
SKAN 4 extended the framework to support web and omni conversions, so attribution is no longer limited to in-app installs. Earlier versions were app-only. Even with web support, the reporting remains aggregated and coarse at low volume, so web campaigns on iOS still rely partly on modeled and server-side signals for full measurement.
How Long Does a SKAN Postback Take to Arrive?
Postbacks are deliberately delayed. Under SKAN 3 a single postback typically arrived 24 to 72 hours or more after install. Under SKAN 4 you can receive up to three postbacks across early, middle, and later windows, each following its own delayed and locked schedule. Plan reporting and optimization around that lag rather than expecting real-time feedback.
Can You Still Do Multi-Touch Attribution on iOS with SKAN?
Not with SKAN alone. SKAN is essentially last-touch and aggregated, so it cannot reconstruct a multi-step journey. To approximate multi-touch and causal effects on iOS, teams combine SKAN with media-mix modeling, incrementality or geo experiments, and platform-provided modeled conversions, then reconcile them into one measurement view.
Key Takeaways
- SKAdNetwork is Apple's privacy-safe, aggregated attribution system for iOS installs and conversions.
- SKAN 4 added multiple postbacks, coarse values, and web or omni support over the older single-postback design.
- Your conversion value is the only rich signal you send, so the encoding schema is a core growth decision.
- SKAN is delayed and blind to journeys, so pair it with modeling and incrementality testing.
- Treat SKAN as one input in a combined measurement stack, not the entire source of truth.