Glossary · Business Analytics

SKAdNetwork

also: SKAN · SKAdNetwork 4 · Apple privacy attribution · SKAN 4.0

Definition

SKAdNetwork is Apple's privacy-preserving attribution framework for iOS app install campaigns. It delivers aggregated, delayed, randomized-timing postbacks with a sparse conversion-value payload, and requires roughly 25 conversions in a privacy bucket before any data is released to the advertiser.

SKAdNetwork (SKAN) replaced the IDFA-based deterministic attribution model that dominated iOS user acquisition through 2020. SKAN 4 (introduced 2022) supports three postback windows, fine and coarse conversion values, and crowd anonymity thresholds. The privacy threshold (typically about 25 conversions in a bucket) means small campaigns or fine-segmented audiences return null data; the postback timing randomization (0-24 hours) breaks deterministic time-of-day join logic. Operators have responded by combining SKAN with geo-holdout incrementality designs, MMM, and CAPI server-side conversion forwarding to reconstruct what last-click attribution can no longer deliver.

Essays on this concept