Apple and Google take up to 30% of subscriptions sold in the app, and they hide your funnel data on top of it. A deep dive on when to route growth around the stores.
If your product is a mobile app, you have a landlord. Two of them.
The rent is visible: up to 30% of every subscription sold inside the app. The second cost is quieter and often bigger. The stores sit in the middle of your funnel and show you almost nothing about what happens inside them.
This article is the deep dive behind my YouMail case study: why the app-store funnel was failing, what the web funnel changed, and how to decide whether your app should make the same move. It maps to lens three of the 30-day growth audit.
The two taxes
Everyone prices in the commission. Fewer teams price in the measurement tax.
When your ad sends someone to the app store, your visibility ends at the store's front door. You cannot see which visitors looked and left. You cannot A/B test the store page the way you test your own site. And connecting a specific ad to a specific subscriber becomes guesswork.
That last part quietly breaks paid acquisition. Ad platforms optimize toward events you send back to them. If subscriptions happen inside an app you cannot instrument properly, the platforms optimize toward installs. Installs are the app world's version of signups: an event with no money attached.
What it looked like at YouMail
YouMail makes voicemail, spam-call blocking, and call-handling tools. Freemium model, subscription upgrades, and at the time, a growth engine pointed entirely at the app stores.
Paid campaigns sent clicks to the store listings. 76% of that traffic bounced there. Of the people who stayed, few installed, fewer registered, and fewer still ever paid. Visitor-to-customer conversion sat around 9%.
Inside the company, the conclusion was that paid acquisition had stopped working. The channel was taking the blame for a funnel defect. That misdiagnosis is common, and it is exactly why the growth audit grades the funnel before judging any channel.
The web funnel
The fix reversed the order of operations. Ads stopped pointing at the stores and started pointing at web landing pages built for each campaign. Visitors registered on the web, saw the product's value, and subscribed on web checkout. The app install came last, after the customer already existed.
Every step became measurable and testable. Landing pages could be iterated weekly. Registration became the optimization event ad platforms could actually learn from, and later, subscription events took over.
The results, from the case study: bounce fell from 76% to 25%. Visitor-to-customer conversion rose from 9% to 22%. Registrations grew 102%. Premium upgrades grew 2.5x. The cost of acquiring a customer fell 77%.
And the channel that had been written off recovered completely: ad-driven signups went from roughly zero to 15% of new business.
Who should make this move
The web funnel fits when three things are true:
- Your revenue is subscriptions, so the 30% commission compounds every renewal and the math strongly favors web checkout.
- Your value can be shown, started, or delivered on the web, even partially. A dashboard, a report, a working demo. The visitor needs a reason to register before installing.
- Your growth depends on paid traffic, which needs the measurable funnel to optimize against. If all your installs come free from app-store search, the store is earning its cut.
It fits poorly for games and for products whose entire value lives in the moment of phone use. Forcing a web detour on those buyers subtracts value instead of adding measurement.
Worth noting: the rules have been moving in the web's favor. Court rulings in 2025 forced Apple to allow US apps to link out to external purchase pages. The direction of travel makes the web funnel more attractive every year, and it was already winning on the math before any of that.
Test it without betting the company
This does not require abandoning the stores. It requires one controlled experiment:
- Pick one paid campaign that currently points at the app stores.
- Build one landing page for it, with web registration and web checkout.
- Send half the budget to each destination and instrument every web step.
- Compare on cost per paying subscriber. Cost per install is the metric that hid the problem in the first place.
- Hand the app to customers after purchase and watch retention. Paying users who chose you tend to stick.
If the web side wins, scale it one campaign at a time. At YouMail, that experiment became the front door for the whole business.
Running an app business and suspicious your funnel is the real problem? Let's talk, or email me at karran@karrangupta.com.