SKAN and Post-ATT Measurement

Mandatory literacy. What SKAN is, versions, CV mapping, postbacks, strategy, modeled conversions.

Why SKAN literacy

SKAN is Apple's privacy-preserving attribution framework for iOS post-ATT. Most iOS measurement now runs through SKAN. Misconfiguration produces useless data; proper configuration produces actionable insights despite signal loss.

What SKAN is

• SKAdNetwork: Apple's built-in attribution framework.
• Privacy-preserving: aggregate signal only.
• Postback from device to ad network after install.
• Limited conversion value bits and time windows.
• No user-level data exposed.

SKAN versions

• SKAN 1.0–3.0: 6-bit CV; single postback within 24h after install.
• SKAN 4.0+: 3 postback windows (0–2d, 3–7d, 8–35d).
• SKAN 4.0+: coarse-grained CV expanded.
• SKAN 5.0: forthcoming with additional features.

CV mapping

• 6 bits = 64 possible values to encode events.
• Trade-off: granularity vs flexibility.
• Common designs: bitmap (each bit = event), enum (each value = event), revenue tier.
• Conversion value mapping is the critical design decision.
• Re-mapping requires app update.

Postbacks

• Sent from device to ad network.
• Aggregated by Apple before sharing.
• Timing: random delay (privacy).
• Threshold: minimum installs before sharing (privacy).
• Three windows in SKAN 4.0+.

Measurement strategy

• CV mapping designed for your LTV signal.
• Modeled conversions supplement SKAN.
• MMP integration for aggregated reporting.
• Network-specific SKAN setup verification.
• SKAN data combined with deterministic (ASA, in-app).

Modeled conversions

• Networks model conversions they can't observe (consent declined, SKAN limitations).
• Statistical models project missing data.
• Combined with SKAN observed for fuller picture.
• Each platform's modeling differs.
• Less precise than deterministic.

Advanced playbook

• CV schema designed for predictive LTV.
• SKAN 4.0+ multi-window strategy.
• Modeled conversion integration.
• Cross-platform measurement strategy.
• Network-level SKAN setup verification.
• Annual CV schema review.
• iOS vs Android measurement reconciled.
• SKAN data in warehouse for analysis.
• Stakeholder education on SKAN limits.
• Apple roadmap monitoring.

Common mistakes

• CV mapping designed without LTV signal in mind.
• SKAN 3.0 mapping used when 4.0+ available.
• Modeled conversions treated as observed.
• SKAN data interpreted at user level.
• Cross-network SKAN setup inconsistent.
• iOS measurement abandoned; Android-only strategy.
• Stakeholders not educated on limits.
• CV re-mapping bottlenecked on app updates.
• Network-specific setup neglected.
• SKAN combined incorrectly with deterministic.

Operating checklist

• CV mapping designed for LTV signal
• SKAN 4.0+ multi-window strategy
• Network setup verified per platform
• Modeled conversions integrated
• MMP aggregation
• Stakeholder education on SKAN limits
• Annual CV schema review
• Cross-platform measurement strategy
• SKAN data warehoused for analysis
• Apple roadmap monitored

Part of the Mobile App Growth series.