Glossary
SKAdNetwork (SKAN)
Apple's privacy-preserving ad attribution framework that allows install attribution without IDFA. The current major version is SKAdNetwork 4.0, introduced in iOS 16.1.
Also known as: SKAN, SKAN 4.0, SKAd Network
What is SKAdNetwork?
SKAdNetwork (SKAN) is Apple’s privacy-preserving ad attribution framework, introduced in iOS 14 and significantly enhanced in iOS 16.1 with SKAN 4.0. It allows ad networks to attribute installs back to a specific campaign without exposing user-level identifiers like the IDFA (Identifier for Advertisers).
SKAN exists because of Apple’s App Tracking Transparency (ATT) framework, which requires explicit user consent before an app can access the IDFA. Most users (>70% in many markets) decline, breaking traditional deterministic attribution. SKAN is Apple’s prescribed alternative.
How SKAN works (simplified)
The flow involves three parties: the ad network, the source app showing the ad, and the target app being advertised.
- Impression. When the source app shows an ad, it registers a SKAN signed payload containing the campaign ID and ad network identifier.
- Click or view-through. If the user taps (or, with SKAN 4.0, views) the ad and then installs the app, iOS associates the install with the signed payload.
- Conversion value updates. After install, the target app can update a “conversion value” each time the user completes a meaningful event (signup, purchase, level completion). The value is stored client-side.
- Postback. After randomized delay windows, Apple posts a winning install + conversion value back to the ad network, with no user identifier — only campaign-level aggregation.
SKAN 4.0 — what changed
SKAN 4.0 introduced four major improvements:
| Feature | SKAN 3.0 | SKAN 4.0 |
|---|---|---|
| Postback windows | 1 (within 24-48 hours) | 3 (0-2, 3-7, 8-35 days post-install) |
| Conversion value granularity | 6 bits (single coarse value) | Fine + coarse (high/medium/low) |
| Source identifiers | 100 campaign IDs | 4-byte source identifier (10,000+ values) |
| Crowd anonymity | All-or-nothing | Tiered (Tier 0 / 1 / 2 / 3 by volume) |
The three postback windows are the most strategically important change. They let advertisers measure not just installs but post-install retention and revenue events across the first 35 days — making LTV-based optimization possible under privacy constraints.
Crowd anonymity tiers (SKAN 4.0)
To prevent reverse-identification through small-sample inference, Apple suppresses campaign-level detail when conversion volume is low:
| Tier | Volume threshold | What gets shared |
|---|---|---|
| Tier 0 | <5 conversions | Only source app name + zero campaign detail |
| Tier 1 | 5+ conversions | Source app + coarse conversion value |
| Tier 2 | 25+ conversions | + Fine conversion value |
| Tier 3 | 100+ conversions | + Source identifier |
This is the dominant reason small advertisers see “Tier 0” or “null” SKAN data — their campaigns simply don’t generate enough installs to clear the threshold.
How SKAN interacts with Apple Search Ads
Apple Search Ads is in an unusual position: it has two attribution paths.
- Ad Services API (first-party). ASA installs are attributed deterministically inside Apple’s ecosystem without IDFA, because both the ad surface and the App Store are operated by Apple. Tap, install, and re-download events report in near-real-time.
- SKAdNetwork (third-party). When your MMP (AppsFlyer, Adjust, Kochava, Branch) observes attribution across multiple networks, it uses SKAN postbacks. ASA shows up as one source in the SKAN mix.
Most ASA-only operators rely on the Ad Services API directly through the Apple Search Ads dashboard (or tools like ASAPilot that wrap it). Cross-channel buyers also use SKAN through their MMP for portfolio-level attribution.
Practical implications for ASA operators
- ASA’s Ad Services API attribution is more granular and faster than SKAN — use it as the primary source for ASA-specific optimization.
- Conversion value mapping is your responsibility. Decide what events (purchase, trial, paywall view) map to what values. Most teams use a 64-value coarse mapping for SKAN 4.0.
- Small campaigns will be Tier 0. If you’re spending under ~$1,000/day on a single SKAN source, expect heavy noise. Consolidate campaigns where possible.
- LTV-aware bidding is now possible under SKAN 4.0 because the 3-window postbacks let you compute Day 35 LTV per campaign instead of just Day 1 install rate.
How ASAPilot uses this
ASAPilot reads ASA attribution via the Ad Services API (the first-party path). For cross-channel SKAN reporting we recommend pairing ASAPilot with your existing MMP — we don’t duplicate attribution infrastructure.
See Apple Search Ads for how ASA fits into the broader iOS UA stack, or pricing for plan tiers.