Subscription apps and the invisible leak

When each install is worth thirty dollars over its life, a tap that never arrives is not a rounding error.

6 August 2026 · PAD team

The economics make it worse

A subscription app can justify $5 to $15 per install because the lifetime value supports it. That also means every lost tap is worth more than it would be for an ad-supported product.

Lose a fifth of taps on traffic campaigns and you are not losing clicks, you are losing a fifth of a cohort that would have paid monthly for a year.

Why it stays hidden

Subscription funnels are measured deep: trial starts, trial conversion, retention, churn. All of those look fine, because they are calculated on people who actually installed. The missing people never enter any of those charts.

The only visible symptom is that acquisition feels expensive, which every subscription team believes anyway.

The paywall preview pattern

Many subscription apps run a web paywall or quiz before the download, precisely to qualify traffic. That funnel runs inside the in-app browser and works fine until the final Download button, which is exactly where the store link fails.

The user answered questions, saw a personalised plan, reached for the app. That is the most expensive moment in the funnel to lose someone.

What to measure

• Taps that reached the store, separately from clicks billed
• Click to install split by campaign objective
• iOS against Android for the same creative
• The final button's escape rate, not just its press rate

One more thing specific to subscriptions

Campaign tokens matter more here than in most categories, because you need to compare cohort quality by source over months. Attach ct at the final store URL so App Store Connect can separate them, as described in the campaign token write-up.