Here's what surprised me most: many large apps can't properly read their own profitability. They don't understand cohorts. They can't forecast future payback. This leads to absurd mistakes.
Take Cal AI. Great app, they disrupted the industry. But in the App Mafia course, its founder revealed they buy only 50% of impressions for their brand terms – despite running massive influencer campaigns.
They think brand traffic isn't profitable.
As a result, Cal AI’s clone happily buys the remaining 50% of brand impressions and generates $150-200k monthly revenue. There's even a clone of the clone doing the same thing.
The hilarious part: Cal AI's ongoing ROAS on brand campaigns is around 50%.
If you have an active campaign with 50% ongoing ROAS, it will become profitable on recharges within months. I can bet $1000 $100.
Not buying 50% of your brand impressions means you're literally throwing away money from your influencer campaigns and organic traffic.
The payback window problem
Forecasting payback is a non-trivial task – especially with iOS attribution issues post-14.5. Most small developers can't afford to hire an analyst who can build payback curves and tell them what KPIs to target on D3, D7, and D14 to forecast 100% payback at different time horizons.
Also, developers often aren't willing to wait more than 3-6 months for their cohorts to pay back.
I believe Cal AI’s brand campaigns will pay back. Maybe the payback will be slightly worse than their other channels. But if you want to scale further, if you want to keep buying influencers, you have to do two things:
buy 100% of your brand traffic
expand your payback window overall.
The most common answer to "how do I scale spend in any channel?" is: you need to expand your payback window. It's quite challenging to break even within 2-3 days.
Let me show you what this looks like in practice.
My app uses weekly subscriptions with trials, plus a yearly subscription without a trial
The advantages:
higher trial conversion rates, more conversions after the initial paywall (about 50% of our conversions happen after the first paywall)
users can test the product, see its value, hit free limits, then subscribe – our organic growth benefits from this approach.
The disadvantages:
with weekly subscriptions as your primary source of revenue, it's nearly impossible to be profitable in the first three days.
When we started buying traffic in 2023, competition was low.
We had a one-month payback. Coming from game dev, where 12-month payback is standard, this felt incredible.
I'd never worked on projects with such fast payback.
But even then, we only knew after waiting a full month, then another month to confirm profitability.
As competition increased, payback stretched to two months, then three, then six.
That 3-6 month range is the valley of death.
Most apps abandon traffic acquisition rather than wait that long for profitability.
If you want to stay in the game long enough, expanding your payback window is one of the few time-tested ways to keep competing for new users.
