Past soft launch, a mobile game needs one person who owns the whole acquisition loop: the Meta and AppLovin budgets, the iOS conversion value schema, the creative pipeline and cohort ROAS read at day 7 and later. For most studios with live revenue and a growing budget, that is either an in house UA lead with engineering and creative support, or a fractional Head of UA directing a small team or an agency; an agency or a network’s own team works on its own only when someone senior on your side already reads the numbers. Workload decides it: once the work fills a full week and you can recruit someone who has scaled a game before, hire that person in house.
Scope: mobile games with live in app purchase, ad or hybrid revenue, past soft launch and scaling paid spend on iOS and Android, any region. Platform facts were checked against Apple, AppLovin, Meta and AppsFlyer pages on 1 October 2026. I sell one of these options, the fractional one, so weigh that. The table says where it breaks.
What changes for UA after soft launch?
Soft launch answers whether the game keeps and monetizes players in a test market. I have written out what a soft launch has to prove separately. Scaling asks different questions, and each one needs an owner:
- More channels. Budget spreads across Meta and ad networks such as AppLovin, and each has its own optimizer, goals and setup rules.
- Thinner iOS signal. Apple’s User Privacy and Data Use page says that without permission through App Tracking Transparency, the advertising identifier is all zeros and you may not track the user. For those users, measurement leans on AdAttributionKit and SKAdNetwork postbacks and, on Meta, on Aggregated Event Measurement, which carry far less than a user level event stream.
- Creative volume. Most game ads lose, so scale depends on how many new concepts reach a fair test each week.
- Longer value curves. A game that looked fine on day 0 ROAS can fail at day 28, or the reverse, depending on how its revenue arrives.
No single specialist covers all four, so the person you put in charge has to connect them.
Who are the options, and where does each break?
The table compares four options a studio can put in charge at this stage.
| Option | Good at | Where it breaks | Fits when |
|---|---|---|---|
| In house UA lead | Full time presence, direct access to product, engineering and data, knowledge that stays in the studio | The hire needs to have scaled a game across Meta, networks and iOS measurement before, and that person is hard to find; one hire cannot also supply creative volume and engineering time | The work fills a full week and you can find and afford someone who has done it before |
| UA agency | Execution capacity, channel coverage, often creative production | An account manager often sits between you and the buyer; the signal layer, the LTV model and cross channel budget calls stay unowned unless the contract gives them to someone | You already have a senior owner in house and need more hands |
| AppLovin’s or another network’s own team | Knowledge of that network’s goals, setup and delivery | It sees one network and earns from your spend on it; it cannot move budget to Meta or define your conversion values | You want help running that network, with the cross channel decisions made elsewhere |
| Fractional Head of UA | Senior ownership of channels, signal and creative testing, part time; the person reading the cohort data is the person changing the campaign | Stops fitting when the work fills a full week, when you need daily people leadership in the building, or when nobody on your side can implement event changes | Live revenue, a growing budget, and no senior owner yet, or an agency with nobody senior on the client side |
The network row has changed recently. AppLovin’s own announcement, AppLovin Ads is now open to all advertisers (22 June 2026), says advertising with AppLovin required a relationship with its team for 14 years, then a referral code for eight months, and now neither. So access is no longer the reason to lean on a network’s team. What that team cannot do is weigh AppLovin against Meta for you.
Which option fits at which stage?
These are my criteria, not industry rules.
- Leaving soft launch, one or two channels, spend rising in steps. This is where the schema, the event values and the cohort reporting get set, and mistakes here compound later. A senior owner matters more than execution capacity. If nobody in house has done it before, bring in a fractional Head of UA, or start with a growth audit if you are not sure what is broken.
- Scaling into Tier 1 markets and adding a network. Budget now has to be split between platforms that report differently. Keep one owner for the decisions; add an agency or a buyer for execution if the volume calls for it.
- The work fills a full week. Many markets, daily creative deployment, several buyers to manage. Hire a permanent UA lead. A fractional lead should help you recruit that person and hand over the documentation.
What must the person be able to do for a game?
Whichever option you pick, check these five abilities. Titles tell you little.
Run AppLovin and Meta together. AppLovin’s Setting up your app campaign page lists four goals: IAP ROAS for apps that earn mainly from purchases, IAA ROAS for apps that earn mainly from ads, Blended ROAS or IAP ROAS for a mix of both, and Cost Per Purchase, which it describes as best when every purchaser has the same value. It says the app must be on MAX mediation for IAA ROAS or Blended ROAS, offers Day 7 and Day 28 optimization windows, and says longer windows generally target the highest value users, with lower early ROAS. Its Scale your campaign page recommends a daily budget that drives at least 15 to 20 conversions a day, all the countries you actively support under one global budget, a mix of video, playable and static assets, and all installs and opens passed to AppLovin. Its Set up MMP tracking page says to configure tracking before any campaign goes live. Meta has its own optimizer and its own setup rules. The person you put in charge has to decide where each platform’s defaults suit the game and where they do not.
Design the iOS conversion value schema for how the game earns. Apple’s Receiving postbacks in multiple conversion windows gives three windows, days 0 to 2, 3 to 7 and 8 to 35, the same structure the SKAdNetwork 4 release notes introduced for iOS 16.1. A fine value is available only in the first postback and only in privacy tiers 2 and 3; Tier 1 and the later postbacks carry a coarse value, and at Tier 0 only the first postback arrives, with no conversion value in it. Apple’s reference for updatePostbackConversionValue defines the fine value as a 6 bit number from 0 to 63 and the coarse value as low, medium or high, and leaves their meaning to the app. For an IAP game the schema has to separate payers early; for an ad revenue game it has to carry ad revenue from the mediation layer; for a hybrid game it has to do both inside 64 values. Apple also says the postback tier depends on crowd size for the app or site showing the ad, the advertised app, the country of install and the network’s source identifier. So the more thinly you split spend across countries and campaigns, the more postbacks are likely to arrive in the tiers with a coarse value or none. My post on what SKAN can carry after a free trial covers what changes for a game, including how a hybrid game can split the 64 values.
Set up Meta’s side of iOS measurement. Meta’s About Meta’s Aggregated Event Measurement describes AEM as the protocol that measures web and app events from people on iOS 14 and later, and says Meta has sent AEM reporting to mobile measurement partners since 9 October 2024. The same page says Meta is gradually extending AEM to app promotion campaigns, which then run without app events configured for SKAdNetwork; no action is needed to get access, though the app may need changes to be eligible, which Events Manager shows. It also says AEM does not currently support custom event optimization. The person should know which app events Meta can optimize on for your game and how those results appear in your MMP next to SKAN.
Read cohort ROAS by day 7 and later, by revenue model. AppsFlyer’s The State of App Monetization, 2026 Edition, covering January 2025 to March 2026 with all apps pooled, found ad revenue reached 89% of its day 60 total by day 7 and purchase revenue 60%. So a day 7 target means different things for an ad revenue game and a purchase game, and the person has to set targets on the window that matches the revenue curve, not the one the dashboard shows first.
Keep the creative pipeline full. A game is understood on sight, and most tested ads lose. The owner should run a testing method with a spend rule per ad, read the results in writing and brief the next round from them. My version of that loop is on the service page, and game ad production is available when production is in scope.
What is the best way to improve signal quality after ATT for a game scaling past soft launch?
Fix the data before the bidding, in this order:
- Agree the source of truth. Store proceeds and your backend for revenue, the MMP for attribution across networks, the platforms for what their optimizers see. Write down how far apart they normally are.
- Check every event value. On a subscription app’s Google App campaigns in 2026, a revenue event sent $1 instead of $0 when there was no value, which inflated the ROAS Google reported for those campaigns, mostly on iOS. That was not a game, but the check is the same, and I wrote up what those campaigns taught me.
- Rebuild the conversion value schema for your revenue model and your volume, as above.
- Confirm each network’s postbacks before spend goes up. AppLovin asks for all installs and opens to be passed to it and for tracking to be verified before launch; on Meta, check in Events Manager that the app is eligible for AEM.
- Then scale on value. On Horse Racing Solitaire, value optimization on Meta went on only once the signal could carry it.
What does Horse Racing Solitaire show?
For two years I was the fractional UA lead for Salto Games on Horse Racing Solitaire, from soft launch through scale. The work was value optimization on Meta across iOS and Android, SKAN and AEM set up so iOS conversions carried value instead of counts, a cohort stack on Adjust and Tableau, and structured creative concept tests.
Over those two years, D0 ROAS rose from a range of 8 to 12% to a range of 17 to 23%, roughly double, and day 1 payer conversion rose 50% across the account. Blended CPI went from £9.43 to £5.38 in one month. From December 2025 to April 2026, 175 ads went through one install campaign in one country, and 14 came in under £1.00 per install.
An outside studio’s ads joined the same account in March 2026. When its best ads moved into the value optimization campaigns, Meta gave them 30% of the day’s budget on 28 March and had cut that to 4.5% by 12 April, inside a like for like comparison of 22 days. What matters for hiring is that I owned measurement, bidding and creative testing together, so I could read each change against the others. It is one account, and your game, economics and audience will produce their own numbers.
What should you ask any candidate?
Ask these whether the candidate is a person, an agency or a network’s team.
- Which cohort ROAS target would you set, at which day, and why that day for our revenue model?
- How would you split budget between Meta and AppLovin, and what would make you move it? Attributed ROAS shows credit, not cause, so a good answer includes a test. I wrote up how I would test whether AppLovin is incremental.
- Walk me through a conversion value schema you designed. What did each value mean, and what happened in the low privacy tiers?
- What do you do when Meta, AppLovin, the MMP and our backend disagree? Look for a reconciliation method.
- How many new concepts a week does our budget need, and who makes them?
- Can you show a game you scaled, with a metric, a time window and what you personally owned?
- What will you do yourself, and what do you need from our engineers?
Where do I fit?
I work as a fractional Head of UA: I buy the media, own the signal and run the creative loop, next to your team or your agency. Retainers are for games spending around $100K a month on paid UA, or funded to get there; paid UA for mobile games describes the work. If you want a diagnosis before deciding who to hire, the growth audit reads the account on its own, at any spend level.
Sources
All checked on 1 October 2026.
- AppLovin: AppLovin Ads is now open to all advertisers, published 22 June 2026.
- AppLovin: Setting up your app campaign, no page date.
- AppLovin Support: Scale your campaign, no page date.
- AppLovin Support: Set up MMP tracking, no page date.
- Apple Developer: Receiving postbacks in multiple conversion windows, no page date.
- Apple Developer: updatePostbackConversionValue, no page date.
- Apple Developer: SKAdNetwork 4 release notes, no page date.
- Apple Developer: User Privacy and Data Use, no page date.
- Meta Business Help Center: About Meta’s Aggregated Event Measurement, no page date.
- AppsFlyer: The State of App Monetization, 2026 Edition, data from January 2025 to March 2026.
Questions people ask
Can AppLovin's own team run our UA instead of hiring someone?
It can help you run AppLovin, but it cannot own your paid UA. AppLovin says advertising with it required a relationship with its team for 14 years and has been open to every advertiser without one since 22 June 2026. A network's team sees one network, earns from your spend on it and cannot move budget to Meta or decide what your iOS conversion values mean. Someone on your side still has to do that.
Should the same person run AppLovin and Meta for a mobile game?
The same person should own the decisions across both, even if different people press the buttons. Each platform has its own optimizer and setup rules: AppLovin, for example, recommends selecting every country you actively support, one global budget and a budget that drives at least 15 to 20 conversions a day. Only one owner can judge both platforms against the same cohort ROAS and decide where the next dollar goes.
What should an iOS conversion value schema encode for a hybrid game?
Both revenue streams, in a form the first postback can carry. Apple gives three windows, days 0 to 2, 3 to 7 and 8 to 35, with a fine value from 0 to 63 only in the first postback and only in privacy tiers 2 and 3; Tier 1 and the later postbacks carry a coarse low, medium or high value instead, and Tier 0 carries no value at all. The app decides what the values mean, so purchase and ad revenue have to be mapped into the same buckets, or the network undervalues the stream it cannot see.
How much should a studio spend before hiring a fractional Head of UA?
My retainers are for games spending around $100K a month on paid UA, or funded to get there. Below that, or before deciding who to hire, the growth audit reads the account on its own and says what is broken and in what order to fix it. It can be booked at any spend level.