Apple's grandfathering rules for subscription price changes — a refresher

Apple's subscription grandfathering policy is consistent but full of edge cases. With more apps experimenting with per-territory pricing in 2026, the rules are worth knowing exactly.

Apple's grandfathering policy for subscription price changes — the rules about what happens to existing subscribers when you change the new-signup price — is consistent across the App Store and well-documented in the App Store Connect Help. But the rules have edge cases that consistently surprise developers running their first price experiment.

The core rule

When you raise a subscription price, existing subscribers continue at their previous price automatically. Apple doesn't notify them or interrupt their subscription. New signups pay the new price. This is the default — no developer action needed.

The 50% threshold

If the price increase is more than ~50%, Apple's behavior changes: existing subscribers are opted out of auto-renewal and required to manually re-confirm the new price. The vast majority do not — most just churn. So a "raise prices 60% for new users" decision frequently produces a one-time MRR drop you didn't anticipate, because your existing subscribers got the opt-out treatment.

The per-territory complication

The 50% threshold is measured per-territory and per-product, not per-app. If you change pricing in 175 territories, each one carries its own grandfathering treatment. A 30% raise in Brazil + a 60% raise in Norway means Brazilian subscribers grandfather; Norwegian subscribers get opted out.

The price-decrease case

Price decreases never opt subscribers out. Existing subscribers continue at the old price; new signups get the new lower price. Apple doesn't auto-discount existing subscribers (you'd have to issue refunds or credits manually).

What this means for app builders

Before any subscription price increase above 30%, model the grandfathering impact: how many existing subscribers in each territory will be opted out, and what's the MRR effect of those churning? For Price Experiments specifically, this is why we recommend max 30% increases in early experiments — large enough to detect a signal, small enough to never trigger the 50% threshold.