Service

Mobile app user acquisition, planned around payback

Installs that turn into payers. I plan multi-network UA for iOS and Android apps from your price, store fee and install-to-paid rate, so every channel gets a maximum CPI before it gets a budget.

Judged on backend revenueFree diagnostic firstOne or two new clients a month
Brass coins slide down a chute and drop one by one into a clear tube, stacking up to a brass ring halfway up, then the full stack sinks out through the base and the tube starts filling again
What you get

Deliverables

For subscription apps, mobile games and app publishers with a working paywall or in-app revenue.

  • UA plan with maximum CPIs derived from price, store fee and install-to-paid conversion
  • Campaign structure per network and country tier
  • Event map in the MMP: one money event per app per network
  • Weekly creative tests with stop rules written before launch
  • Cohort and payback reporting by network, campaign and country

Platforms and tools

  • Meta Ads
  • Google App campaigns
  • TikTok Ads
  • Apple Ads
  • Moloco
  • AppsFlyer
  • Adjust
  • SKAN
  • AdAttributionKit
Diagram of a user acquisition loop: creative, install, payer, cohort payback and budget, with a kill rule gateA loop of five steps: creative, install, payer, cohort payback and budget, which feeds the next creative. Between cohort payback and budget sits a gate in the accent color: the kill rule written before launch.01Creative02Install03Payer04Cohort payback05BudgetKILL RULEPayback decideshow much goes back in
Diagram of a user acquisition loop: creative, install, payer, cohort payback and budget, with a kill rule gateA loop of five steps: creative, install, payer, cohort payback and budget, which feeds the next creative. Between cohort payback and budget sits a gate in the accent color: the kill rule written before launch.01Creative02Install03Payer04Cohortpayback05BudgetKILLRULEPayback decideshow much goes back in
The loop I run: payback decides how much goes back in, and the kill rule is written before launch.
Receipts

From my own work

Anonymized, with the denominator and the caveats left in.

The App Store fee on my own developer account in 2026

It moved twice in one year, and every max CPI starts from it

Sample1 developer account, my own iOS apps, 2026

  1. May15% feekeeps $85 of every $100
  2. From July30% feekeeps $70 of every $100
  3. From September15% feekeeps $85 of every $100

The thin line under each bar is the most an install can cost at the same install-to-paid rate, drawn to scale. Amounts before tax.

My own Apple developer account went from a 15% App Store commission to 30% and back to 15% in 2026. That is why the fee is checked again before any max CPI is set.
Cheaper CPI, worse users
15.1% → 9.9% day-one retention

On a DSP campaign for a gaming app, CPI fell from $4.74 to $1.66 and I didn't call it a win, because day-one retention fell from 15.1% to 9.9%.

Unit economics
11.9% vs 5.4% install-to-paid

At a 30% store fee, a $3 CPI needs about 11.9% install-to-paid at $35.99 a year, against 5.4% at $79.99.

How it starts

Diagnostic first, then a plan

  1. Free 30-minute diagnostic

    A short form, a pre-read built from public data and a call on the three biggest opportunities.

  2. A written plan

    Scope, tests in order, the stop rule for each test and what I need access to.

  3. The work

    Campaigns on each network within its maximum CPI, reported against your backend numbers with the stop rules agreed up front.

FAQ

Common questions

Do you work on iOS and Android?

Yes. iOS adds SKAN and AdAttributionKit to the measurement plan, so the event map is set up for each platform separately.

What do you need to start?

Your price and paywall setup, install-to-paid rate by cohort if you have it, access to your MMP and the ad accounts you already run. Without a working paywall or in-app revenue there is nothing to optimize toward yet.

Which networks do you start with?

It depends on the category, the platform and the budget. The UA plan ranks channels by expected payback and testing cost, and we start with the one or two that can be measured cleanly.

Free · 30 minutes

Want a second pair of eyes on this?

Send me the basics and get a pre-read plus 30 minutes on your three biggest opportunities, free.

Get a free 30-minute diagnostic