iOS attribution after ATT: what you can still know, and what you have to accept

2 min read

The era of knowing exactly which ad produced which paying user on iOS is over for most channels. That is not the end of measurement; it is the end of pretending measurement was ever as precise as the dashboards suggested. Here is what you can still know, and how we set it up.

Apple Search Ads: still clean

Apple's own attribution for Search Ads works without tracking consent and tells you, deterministically, which campaign and keyword produced an install. Connected to your subscription events, this gives you trial and renewal by keyword. It is the most reliable signal in iOS marketing, and one of the reasons Search Ads is usually the first channel we turn on.

Social channels: aggregated and delayed

Meta, TikTok and others on iOS are measured largely through Apple's privacy-preserving frameworks, which report in aggregate, with delays and with limits on how much post-install detail can be sent back. Campaign-level numbers are usable; keyword-level or creative-level precision is not what it was. Treat the numbers as directional and compare channels on the same basis.

Your source of truth

Subscription events from your own backend, sent server-side to your attribution tool: trial started, converted, renewed, refunded. These are complete and accurate regardless of consent, because they are your data about your customers. Everything else is a lens on top of this ledger.

Incrementality

When attribution cannot tell you whether a channel is working, run the experiment instead. Pause the channel in one set of countries for two weeks, keep it running in a comparable set, and compare total trials and renewals. It is cruder than a dashboard and far more honest.

The setup we build

  • Apple Search Ads attribution connected to subscription events.
  • Server-side subscription events as the ledger.
  • Social channels measured at campaign level with conversion values configured for trial and paid events.
  • A monthly incrementality check on the largest spend.

It will never be as neat as the old dashboards. It is a lot closer to the truth.

Written by , founder of KNT App Studio. He designs and ships the studio's iOS apps (Dreamify, Viola and GERG) and runs Apple Search Ads and ASO for the studio and the teams it works with. Posts describe what we did and what it cost; when a number is an estimate, the text says so.