App user acquisition · iOS & Android
Installs are not the product. Paying users are.
We run paid acquisition for subscription apps on iOS and Android, and we report on the cohort – install to trial, trial to paid, renewal to churn – because that is where the money either appears or does not.
- Platforms
- UAC · ASA · Meta · TikTok
- Attribution
- AppsFlyer · Singular
- iOS measurement
- SKAdNetwork 4
- From
- $5,000 / mo
Why most app UA reporting is wrong
A cost per install of $2.40 tells you almost nothing. The same $2.40 can be an excellent buy or a slow bankruptcy depending on what share of those installs start a trial, what share of trials convert, and how many of those convert again at renewal. Optimising campaigns against CPI reliably buys the cheapest users, who are usually the ones least likely to pay.
So the first thing we change on almost every account is what the campaigns are told to value. Then the structure. Then the creative. Bidding comes last, which is the reverse of how most accounts have been tuned.
What we actually do in an account
Fix the measurement layer first
In-app events mapped to what the business cares about, purchase and subscription events firing once rather than three times, deep links that survive, and a SKAdNetwork conversion-value schema designed around a decision instead of copied from a template. On iOS the schema is the single highest-leverage technical artefact in the account: it determines what you are able to know at all.
Rebuild the structure around decisions
Separate what you need to control separately – geography, platform, campaign objective – and merge everything else so each unit gets enough conversion signal to leave the learning phase. Fragmenting a $15,000 budget across nineteen campaigns is the most common way accounts stall.
Test creative on a schedule, not on a hunch
In app UA creative is the main lever and it decays fast. We run a standing cadence with a written hypothesis per concept, hold the audience constant while creative is the variable, and kill on stated criteria rather than on the fourth day when someone gets nervous.
Report to payback
Cohorts by source, campaign and creative, tracked to the month they turn profitable. That is the number that decides whether to scale, and it is not visible in any ad platform's own dashboard.
By channel
Where the budget goes
Google App Campaigns
The channel with the least manual control and the most ways to waste money. Structure, bid target choice and asset discipline are the whole job.
Read more →Apple Search Ads
Highest-intent inventory on iOS and the cheapest place to learn what your buyers actually call your product.
Read more →Subscription apps
Where install volume stops being the answer and trial start, trial-to-paid, renewal and churn become the whole conversation.
Read more →Who this works for
Subscription apps in health and fitness, productivity, education, utilities or fintech, targeting Tier-1 markets, ideally with revenue data we can reach – RevenueCat, Adapty, or your own backend. There is no minimum spend, though on iOS the measurement gets meaningfully sharper once you are past roughly $10,000 a month, because that is where SKAdNetwork's privacy thresholds stop swallowing your conversion values.
It works less well for pre-launch apps with no conversion history, for games, and for anything where the monetisation model is ad revenue rather than subscription. We will tell you at the first call if you are in that group.
Questions people actually ask
- What do you need access to before starting?
- Your ad accounts with admin rights, your MMP (AppsFlyer or Singular) with reporting access, App Store Connect and Play Console for organic and store data, and whatever revenue reporting you have – RevenueCat, Adapty, or your own backend. Without revenue we can only optimise towards installs, which is the problem we are trying to fix.
- Can you work without an MMP?
- On Android, partially. On iOS, not usefully. SKAdNetwork postbacks need somewhere to land and a schema someone designed on purpose, and Apple Search Ads plus Google App Campaigns plus Meta cannot be reconciled without one. If you do not have an MMP we will set one up before spending.
- What size of app do you take on?
- Any size – there is no minimum budget. What we will tell you honestly is that below roughly $10,000 a month of iOS spend, SKAdNetwork postbacks get sparse enough that cohort decisions take much longer to become reliable. That is a fact about Apple's privacy thresholds, not a rule of ours, and it changes what the first few months are realistically for.
- Do you do ASO as well?
- No. We will tell you when store conversion is your bottleneck rather than acquisition cost, because that changes what you should spend on, but keyword optimisation and store page testing are somebody else's craft and we will point you at people who do it properly.
- How long until the numbers mean anything?
- Campaigns exit learning in two to three weeks. Cohort data that tells you about payback needs the cohort to age – for a monthly subscription that is a minimum of five to six weeks before renewal behaviour is visible. Anyone quoting you a verdict at week two is reading noise.
- Games or non-games?
- Non-games. Subscription utilities, health and fitness, productivity, education, fintech apps. Games are a different economy – ad monetisation, IAP whales, playable creative pipelines – and there are agencies who do only that.
Free
Get the app UA audit
We look at your campaign structure, your SKAdNetwork schema, your MMP setup and your cohort reporting, then write up what we would change.
Or read the numbers first.
The benchmarks and the subscription guides are free, need no email, and will tell you more about how we think than any sales page.