Definition

SKAdNetwork (SKAN) is Apple’s privacy-preserving framework for attributing app installs and post-install events to ad campaigns without sharing user-level data. Ad networks report aggregated conversion values with timing noise.

Detailed Explanation

Campaign IDs map to ad sets; conversion value schema encodes funnel depth (install → trial → purchase). Postbacks arrive with a random delay, from about a day for the first window to several days for later windows.

Meta, Google, and TikTok provide SKAN dashboards; align CRM revenue with CV buckets for ROAS estimation.

SKAN 4 (iOS 16.1+) added up to three postbacks and coarse or fine conversion values depending on campaign volume. Apple introduced AdAttributionKit in iOS 17.4 as its successor, and it works alongside SKAN — keep ad network SDKs updated.

Nepal Context

Local fintech and ride-hail apps buying installs should define CV schema tied to NPR transaction thresholds, not only install counts.

Small install volumes in Nepal lengthen learning — avoid weekly creative pivots based on noisy SKAN data.

Practical Examples

  1. Beginner: Enable SKAN in Meta app settings, use default 6-bit conversion mapping.

  2. Intermediate: Custom CV mapping: install=0, KYC complete=32, first transaction=63.

  3. Advanced: Join SKAN postbacks with backend revenue in warehouse for true cohort ROAS by campaign ID.

How to Check It

  1. In your mobile measurement partner or ad network dashboard, confirm SKAN postbacks are arriving for each active campaign.
  2. Review the conversion value schema with the product team so each value maps to an event you can also see in your backend.
  3. Compare SKAN-reported installs with App Store Connect downloads over a month to see the size of the gap.

Key Takeaways

  • SKAN replaces granular IDFA attribution on iOS.
  • Conversion value schema must reflect business KPIs.
  • Expect delayed, probabilistic reporting.
  • Coordinate media buyer and app developer on SDK updates.
  • Nepal app markets need patience with small data volumes.

Common Mistakes

  1. Optimizing purely for installs when LTV lives in later CV bits.
  2. Ignoring SKAN 4 migration deadlines.
  3. Daily budget swings based on single postback days.