Mobile game UA after IDFA is a different job from the one most studios learned in 2019, but it is not a broken one. Apple’s App Tracking Transparency prompt arrived in 2021 and ended the easy, user-level view of which ad produced which install on iOS. Four years on, the teams that kept scaling are the ones that stopped waiting for that view to come back and rebuilt their process around what is measurable today.

This post covers a post-IDFA UA strategy in practical terms: what you can still measure, how to read it, and how to prepare for whatever privacy rules arrive next. It is written for studios running paid acquisition, not for privacy lawyers, so we keep the mechanics light.

What actually changed with IDFA

Before App Tracking Transparency, an ad network could match an install to a specific device and report back, per user, what that player did next. That made it possible to optimise a campaign toward the players who spent or watched the most ads within days.

After the prompt, that matching only works for players who opt in to tracking. Many do not, so on iOS a large share of installs arrive with no device-level signal. Apple replaced the old model with its own aggregated attribution framework. It reports installs and a small amount of in-game behaviour back to the network, with delays and with privacy thresholds that hide data on small campaigns.

The practical result is less data, later. Android did not change in the same way, and that difference shapes how studios split their work between the two platforms.

How measurement works on iOS today

Apple’s attribution system started as SKAdNetwork and has since been succeeded by AdAttributionKit. For a UA team, the ideas that matter are the same in both.

Postbacks arrive late and in pieces. You do not see a player’s full journey. You see a limited number of reports over a set period after install.

You choose which signals to send. Studios map a small set of in-game events, or a bucketed revenue level, to a conversion value. That choice decides what the ad networks can optimise toward, so it is one of the most important decisions you make on iOS.

Privacy thresholds hide small campaigns. When a campaign has too few installs, the report can come back with little or no detail. Fewer, larger campaigns usually give you more readable data than many small ones. Apple has also added more flexibility in newer versions, including configurable attribution windows, so it is worth checking the current documentation with your attribution partner rather than relying on notes from a few years ago.

For the ad networks, a weaker signal means optimisation is slower and less precise on iOS. That is true for every network, including Applovin, Google Ads, Mintegral and Unity. It is a platform limit, not a network flaw.

What still works: a post-IDFA UA playbook

Studios that still scale on iOS tend to do the same handful of things.

Design your conversion events on purpose. Pick the early signals that best predict long-term value in your game: a tutorial milestone, a level reached, a first purchase, or an ad engagement threshold. Test whether the signal you chose really separates good cohorts from poor ones, and revise it when your game changes.

Lean on Android for fast learning. Android still offers richer, quicker data. Many teams test new creatives, geos and bids on Android first, then carry the winners over to iOS. It is not a perfect mapping, because audiences differ, but it saves a lot of iOS spend on ideas that were never going to work.

Judge results at cohort level. Stop looking for a clean answer for each campaign on each day. Track blended performance by install week and by platform, and compare it with payback over time. Our guide to what a good ROAS looks like explains why blended, patient reading beats day-one panic.

Use incrementality and holdout checks. When attribution is blurry, a simple test can answer the question that matters: did this channel bring players you would not have had anyway? Pausing a channel in one region for a short period, or comparing similar regions, gives you directional proof. Treat it as a sanity check, not a precise number.

Let creative do more of the work. When targeting signals shrink, the ad itself decides who clicks and who installs. Strong hooks and honest previews filter for the right players before any algorithm does. That is why creative output matters more now, and why creative fatigue hits harder when you cannot rely on precise targeting.

Keep your budget flexible. With delayed data, you cannot move spend daily on a noisy signal. Set review points weekly, give each change time to read out, and use a clear budget plan so you are not tempted to overreact.

What about Android and the Privacy Sandbox?

This is the question behind “how are gaming studios preparing their UA strategies for the Privacy Sandbox rollout”, and the answer has changed since the original plan.

Google announced the Privacy Sandbox for Android in 2022, with the aim of moving away from the Google advertising ID towards on-device, aggregated attribution, much like Apple. The plan was slow and many companies were cautious about adopting it. In October 2025 Google announced that it was retiring several Privacy Sandbox technologies, including its attribution reporting API, on both Chrome and Android, citing low adoption. Google said it would carry its attribution work forward through a web standards group.

What this means in practice is that the Android advertising ID is still in use, and attribution on Android works much as it did. We would not read that as permanent. Privacy rules and platform policies keep moving, and a retired project does not stop the direction of travel. Studios are best placed when they plan for less user-level data, not more.

If you want the latest detail, check Google’s own Android developer documentation and ask your attribution partner what they expect. Do not trust any article, including this one, as the final word on dates.

How to prepare for the next round of privacy changes

You cannot predict the rules, but you can reduce how much your results depend on them.

  1. Own your first-party data. Know your own install, retention and revenue numbers from inside the game, independent of any ad network. It is your reference when network reports disagree.
  2. Keep your event design flexible. Make sure you can change which events you send and how they are bucketed without a major release.
  3. Stay on the current SDKs. Attribution and ad SDKs are updated to match platform changes. Falling behind on versions is a quiet source of lost data.
  4. Do not depend on one channel. A mix of networks, with Applovin as the lead in many of our own campaigns and Google Ads, Mintegral and Unity alongside, protects you if one channel loses signal. Our scaling roadmap shows how we phase that out.
  5. Revisit your assumptions every quarter. Check that your payback model still matches what the data shows, especially on iOS.

The short version

Mobile game UA after IDFA rewards teams that accept a lower resolution view and build habits around it: purposeful conversion events, cohort-level judgement, Android-first testing, incrementality checks and strong creative. The rules will keep shifting, so a flexible process matters more than any single tool.

If your iOS numbers feel unreadable and you want a second opinion on how your events and campaigns are set up, see how we work on the services page. We would rather help you build a sound measurement setup than sell you a promise.