Mobile game UA strategy after IDFA: what you still control
Gaming ATT opt-in sits at 39%, the EU prompt changes with iOS 27.2, and Meta and Google now choose the audience themselves. What a UA team still controls in 2026, and how to plan around it.
A mobile game UA strategy in 2026 starts from one fact: the ad networks pick the audience now. They do it with less user-level data than they had before Apple's App Tracking Transparency (ATT). Five years on, a UA team controls four things: the conversion signal it sends back, how it splits budget across networks, where the player pays, and the creative it feeds the auction. The fourth is the one most teams are short of. This guide covers all four, plus the one change still ahead: from iOS 27.2, the ATT prompt itself changes in the EU.
This post is about paid acquisition. For the wider playbook (store listing, channel order, who owns creative in a studio), start with how to market a mobile game.
Where mobile game UA stands in 2026#
The worst of the ATT shock is over. AppsFlyer puts 2025 app UA spend at $78 billion, up 13% on 2024. All of the growth came from iOS, up 35%, while Android slipped 1% (AppsFlyer). For games specifically, UA spend grew 6% on iOS and 2% on Android. That suggests iOS measurement is good enough again for budgets to follow it.
Installs got more expensive at the same time. Adjust's 2026 gaming report puts the global gaming cost per install (CPI) at $0.56, up 30% in a year (GameDevReports). The spread between regions is wide enough to plan around:
That spread is why a soft launch in a cheaper market still makes sense before a US launch. You can buy enough installs to read retention and test early creatives for a fraction of the US price. Just don't carry a LATAM CPI into a US budget: the creative that wins there gets re-tested at US prices.
ATT opt-in has settled at about 39% for games#
Most game players still say no to tracking, but the share saying yes keeps creeping up. Adjust measured gaming opt-in at 39% in Q1 2026, up from 38% a year earlier and above the 38% average across all apps. Adjust measures it among users shown the prompt.
| Sports | 54% |
|---|---|
| Action | 50% |
| Arcade | 46% |
| All games | 39% |
| All apps | 38% |
Genre matters more than any prompt trick. A sports game can expect half its prompted users to opt in; an arcade game, under half. When you model iOS performance, use your own genre's rate. The blended industry number will flatter some games and punish others.
The EU's ATT prompt changes with iOS 27.2#
This is the part of "UA after IDFA" that is still moving. On 17 August 2026 Germany's competition authority, the Bundeskartellamt, closed its case against Apple by accepting binding commitments rather than issuing a fine. Apple will remove what the authority called "possibly discouraging symbols and wording" from the prompt. The prompt must be neutral in content, wording and layout. Publishers also get more freedom to combine Apple's request with their own consent request. Apple has four months to implement it, and the commitments run for seven years under a monitoring trustee (Bundeskartellamt).
Apple followed on 16 September: from iOS and iPadOS 27.2, developers can use an alternative ATT prompt in the EU. In Germany, France, Italy, Poland and Romania only the alternative version is available (Apple). TechCrunch reports that the decline button changes from "Ask App Not to Track" to "Reject", and that apps may ask again a year after the user's last answer (TechCrunch).
Eric Seufert expects EU opt-in rates to rise "by a non-trivial amount" and doubts the US will follow (Mobile Dev Memo). What we would do now:
- Plan the EU prompt as its own test. Your consent screen and the ATT request can now be designed together. Test when you ask and what comes before it, country by country.
- Report EU iOS separately. If opt-in rises in five countries, their measured performance improves without anything else changing. Don't let that read as a creative or bidding win.
- Put the one-year re-prompt in the calendar. Users who declined at install can be asked again. Decide which moment in the game you'll ask at.
Measurement after IDFA: three layers, three jobs#
No single number tells you what an install was worth any more. A workable setup has three layers, each answering a different question.
| Postbacks and attribution | Media mix model (MMM) | Incrementality tests | |
|---|---|---|---|
| The question it answers | Which campaign and creative drove this install? | Where should next quarter's budget go? | Would these installs have happened anyway? |
| Speed | Days | Weeks, refit monthly or quarterly | Weeks per test |
| iOS input | AdAttributionKit or SKAdNetwork postbacks, plus opted-in users | Spend and outcome time series | Geo or holdout splits |
| Use it for | Daily bids and creative decisions | Splitting spend across networks | Settling disputes between the other two |
Apple's attribution stack is stable. Apple now tells developers to "use AdAttributionKit for app ad campaigns." The June 2025 update added postbacks in multiple conversion windows, a country code in postbacks, and configurable attribution rules and cooldown windows (Apple). SKAdNetwork isn't deprecated, though two of its methods are (Apple). AdAttributionKit's update log has no 2026 entry. If your conversion-value schema dates from the SKAN 4 rollout, revisit it against today's monetisation events rather than waiting for Apple to force the issue.
Android keeps its ad ID. In October 2025 Google said it would retire most Privacy Sandbox technologies, including the Android versions of Attribution Reporting, Protected Audience and Topics, citing low adoption (Google). The Google advertising ID remains. A user who deletes it returns a string of zeros, and apps targeting Android 13 or later must declare the AD_ID permission (Google Play). Android remains the place where deterministic attribution still works for most users.
MMM costs less to start than it used to. Google opened Meridian, its open-source media mix model, to everyone in January 2025, with more than 20 certified measurement partners (Google). A studio spending across several networks no longer needs a consultancy to start modelling budget allocation.
The networks do the targeting now#
Meta's chief executive has been clear about where ads are going. Asked by Stratechery about the future of Meta's ads, Mark Zuckerberg described advertisers setting an objective and a budget while the system handles creative, targeting and measurement (Stratechery). For app campaigns, most of that has already happened.
- Meta moved app campaigns to Advantage+. An Advantage+ app campaign takes up to 50 creatives, and Meta reports a 7% average improvement in cost per action for advertisers who used them (Meta).
- Google App campaigns take up to 4 text lines, 20 images and 20 videos. Google's guidance says asset variety lets the campaign build more ads for different users, and steadies performance when you refresh creative (Google Ads Help).
- The in-app networks run on machine-learned bidding too. AppsFlyer's 2025 Performance Index has Apple Ads as the top media source for iOS gaming and Google Ads leading Android gaming. AppLovin is narrowing the gap and already leads in tier-1 North America and Western Europe (AppsFlyer).
What's left for the advertiser is the input: the creative. Each concept is a hypothesis about who the game is for. A battle-heavy ad finds a different player from a base-building ad, and the network can only find the players your creatives speak to.
Channel mix: concentrate, then branch out#
Spend is concentrating on the biggest networks. In AppsFlyer's index, 60% of the global top-5 media sources grew spend year on year, and 80% of those ranked 6 to 10. Only 30% of those ranked 11 to 20 did (AppsFlyer).
The in-app networks are where the market moved fastest. AppLovin's revenue rose 53% to $1.92 billion in Q2 2026 (AppLovin, SEC filing). In Q1 its revenue per install rose 93% while install volume fell 18% (AppLovin 10-Q). Advertisers are paying more for each install the network delivers, not buying more of them. Unity's Grow revenue rose 35%, driven by its Vector ad model, and the ironSource Ads network was shut down on 30 April 2026 (Unity). If your plan still lists ironSource Ads as a separate line, it's out of date.
A sensible order for most studios:
- Meta and Google App campaigns first. They carry the most reach, and both are now set up to take many creatives.
- Apple Ads on iOS. It tops AppsFlyer's index for iOS gaming, and it reaches players who are already searching the store.
- AppLovin, then Unity and Mintegral, once you have a game whose early retention justifies their return-on-spend bidding.
- TikTok and YouTube for reach, treated as creative tests as much as channels.
Web shops change what a player is worth#
Your UA ceiling is set by what a player pays you, and that number moved in 2025 and 2026.
| 2019 | 12% |
|---|---|
| 2024 | 44% |
| 2025 | 74% |
Bain found that 47% of 5,339 players surveyed had bought from a game's web store in the past year, and puts the gain at 15 to 30 extra margin points per direct sale (Bain). The rules underneath are still settling:
- iOS (US): after the Epic v. Apple contempt ruling, Apple had not charged commission on link-out payments for nearly a year as of May 2026. The court is still deciding what fee Apple may charge, so don't plan on 0% being permanent (The Next Web).
- Google Play (US, UK, EEA): from 30 June 2026 the service fee starts at 10% on a developer's first $1 million a year, and the extra 5% billing fee applies only to Google Play billing, not to alternative billing or web links (Android Developers Blog).
For a UA team this is a lifetime value (LTV) question before it is a payments one. If 15 to 30 more margin points of a payer's spend reach you, the same CPI pays back sooner.
Liftoff says in-app-purchase games can take 90 to 180 days to pay back (Liftoff). Rebuild your return-on-ad-spend (ROAS) targets on net revenue after fees, by OS and country. And note what developers say holds direct sales back. In a GDC and Appcharge survey of 1,200+ developers, player awareness (50%) and player acquisition (41%) top the list (Payments Dive). Telling your existing players about the shop is a marketing job too.
Creative is where the plan usually breaks#
Once the networks do the targeting, the creative plan is most of the targeting plan, and that is the part production limits. The volume is already high. Singular's 2026 benchmark has the median app advertiser launching 13 new creatives a week, with the top quarter at 53 or more. AppsFlyer counts 2,400 to 2,600 variations a quarter from the top gaming advertisers, up 25 to 30% in a year (AppsFlyer).
Volume alone isn't the answer, though. In the same Singular data, the top 10% of creatives take up to 95% of spend, and creative volume on its own doesn't move click-through, conversion or return (Singular). The point of a batch is coverage: distinct concepts, each a real hypothesis, and then controlled variations of the ones that win.
2 × 3 × 2 × 2 =24versions of one ad, each changing only what you chose
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
Clash of ClansStrategy · Free- Build & defend
- Raid for loot
- Join a clan
That is the shape of a good week's batch: a few concepts, each broken into the parts you can vary (character, headline, button, size). The variations stay on-model because they recombine the game's own pieces rather than inventing new ones. For what to vary first, testing the first 3 seconds covers the opening, where most of the signal sits. Mobile game ad sizes covers the formats every concept has to ship in.
AI is already part of this at the top of the market: 56% of the 100 top-grossing games used AI in ad creation in 2025, by AppMagic's count (Gamigion). The research on how to use it is telling. In a field study Liftoff cites, AI-built ads made from human briefs drew 19% higher click-through than ads built by human experts alone. AI edits of existing human ads showed no significant lift (Liftoff). The person still sets the brief. The AI does the building.
That is the step Vidscape covers. Paste a store link and it turns the game's own art into layered mobile game ads, in every size the networks take, with each headline, character and button still editable. If your team is measured on creative throughput, Vidscape for growth teams shows how a weekly batch fits a UA calendar.
A mobile game UA plan for 2026#
- 1Fix the signal before the spend
Map your AdAttributionKit or SKAdNetwork conversion values to the events that predict revenue today. Check that Android events reach every network.
- 2Split the EU out on iOS
Report the five neutral-prompt countries separately from iOS 27.2 onwards, test your consent and ATT screens together, and schedule the one-year re-prompt.
- 3Rebuild targets on net revenue
Recalculate LTV and ROAS targets after store or web fees, by OS and country. Higher net margin per payer shortens payback.
- 4Concentrate, then add in-app networks
Meta and Google first, Apple Ads on iOS, then AppLovin and Unity once early retention supports return-on-spend bidding.
- 5Plan creative as concepts, not files
Ship several distinct concepts a week and vary the winners' parts. Fill the creative slots the campaigns allow: 50 on Meta Advantage+, 20 images and 20 videos on Google.
- 6Check the model against a test
Fit an MMM for budget, and settle the disagreements between it and attribution with a holdout or geo test each quarter.
Frequently asked questions#
What is a mobile game UA strategy?
The plan for buying players profitably: which networks to buy on, what to pay per install or per action, how to measure the result, and which creatives to run. In 2026 the networks handle most targeting, so the strategy rests on the conversion signal, the budget split, the payment route and the creative.
How does ATT affect mobile game UA in 2026?
About 39% of prompted game users opt in to tracking (Adjust, Q1 2026), so most iOS measurement runs through AdAttributionKit or SKAdNetwork postbacks and modelled data. iOS budgets have recovered anyway: iOS drove all of 2025's UA spend growth, according to AppsFlyer.
Is SKAdNetwork deprecated?
Not formally. Apple recommends AdAttributionKit for app ad campaigns and has deprecated two SKAdNetwork methods, but the framework itself is not deprecated. AdAttributionKit's update log has no 2026 entry.
What happened to Privacy Sandbox on Android?
Google announced in October 2025 that it is retiring most Privacy Sandbox technologies, including the Android versions of Attribution Reporting, Protected Audience and Topics, citing low adoption. The Google advertising ID remains available.
Should a mobile game soft launch outside the US?
Usually, yes. Gaming CPI in 2025 was $1.68 in North America against $0.53 in Europe and $0.14 in Latin America (Adjust), so retention and early creative tests cost far less elsewhere. Re-test winners at US prices before you scale them there.
- Top 5 Data Trends of 2025 and Predictions for 2026 — AppsFlyer · 2026
- Gaming App Insights Report 2026 (PDF) — Adjust · Apr 2026
- Adjust Gaming App Insights report (summary) — GameDevReports · Apr 2026
- Adjust's 2026 mobile app report — PPC Land · Feb 2026
- Press release on Apple's App Tracking Transparency Framework — Bundeskartellamt · Aug 2026
- Updates to App Tracking Transparency in the European Union — Apple Developer · Sep 2026
- Apple will let EU apps use less alarming tracking consent screens — TechCrunch · Sep 2026
- Apple forced to restructure ATT — Mobile Dev Memo · Aug 2026
- AdAttributionKit updates — Apple Developer Documentation · 2025
- SKAdNetwork — Apple Developer Documentation · 2026
- Update on plans for Privacy Sandbox technologies — Google · Oct 2025
- Advertising ID — Google Play Console Help · 2026
- Meridian marketing mix model is now open to everyone — Google · Jan 2025
- An interview with Meta CEO Mark Zuckerberg about AI and the evolution of social media — Stratechery · May 2025
- Advantage+ app campaigns — Meta for Business · 2026
- Make better creative assets for App campaigns — Google Ads Help · 2026
- AppsFlyer releases 2025 Performance Index — AppsFlyer · Dec 2025
- AppLovin second quarter 2026 earnings release — AppLovin (SEC filing) · Aug 2026
- AppLovin Form 10-Q, Q1 2026 — AppLovin (SEC filing) · May 2026
- Unity reports second quarter 2026 financial results — Unity · Aug 2026
- Game distribution: half your players already buy direct — Bain & Company · Aug 2026
- Supreme Court denies Apple's stay in Epic contempt case — The Next Web · May 2026
- Expanded billing choice and lower fees on Google Play — Android Developers Blog · Jun 2026
- Direct-to-consumer is already a $17bn market in mobile gaming — Payments Dive (GDC and Appcharge) · Jun 2026
- What is a good ROAS? 2026 benchmarks — Liftoff · Apr 2026
- Creative Benchmark Report 2026 — Singular · 2026
- State of Gaming for Marketers 2026 — AppsFlyer · Jan 2026
- Mobile Market Landscape 2026 by AppMagic — Gamigion · Feb 2026
- Creative in the AI era — Liftoff · 2026
