Apple retention messaging now sits in the App Store cancel flow, and the early benchmark is a median 17% save rate against roughly 1% offer acceptance. Developers can show text, an image, or a qualifying offer before a subscriber finishes leaving, earlier than 2024 win-back offers for people who had already canceled. That split means a default discount is a weak use of a reviewed slot; the save is usually a non-cancel, not an accepted offer.
Updated November 6, 2026
Key Takeaways
- Median save rate was 17% across 309 apps with at least 100 messages.
- Promotional offers were accepted about 1.1% of the time; plan switches, 0.6%.
- Most recorded saves happen without the subscriber taking an offer.
- Early data shows no clear cancellation drop versus comparable apps.
What Changed in the Cancel Flow?
On November 4, 2026, Apple opened retention messages for auto-renewable subscriptions, as MacRumors reported. Announced at WWDC earlier in 2026, the messages appear in the cancellation flow and can include text, images, or a special offer. App Store Connect sets eligibility and the wait between messages.
With the Retention Messaging API, available by request, developers can choose which message to show a subscriber based on subscription details. The Mac Observer adds the rules that shape the design: up to ten messages per app, only one active per subscription, and no scare tactics or unprovable claims. The text itself cannot mention specific prices, because the prompt might be shown to people who do not qualify. A lower price, pay-as-you-go discount, or free period appears apart from the main copy.
How Do Apple Retention Messaging Types Compare?
RevenueCat’s analysis covers March 22 to September 20, 2026, ahead of the wider rollout. The save-rate set is 309 apps with at least 100 messages — 21,960,020 messages, counted once per subscription per day, each app weighted equally. They are early access apps, not a representative sample, and the data does not establish that messages caused subscribers to stay or renew.
Save rate means no cancel within 24 hours, or an offer taken. The median was 17%. A quarter of apps saved fewer than 10 in 100; another quarter, more than 27 in 100; the top 10%, 42 or more. Some people would have stayed anyway, a cancel a few days later still counts, and renewal is not tracked.
| Comparison point | Figure |
|---|---|
| Text with promotional offers | 27% of apps |
| Primarily promotional offers | 22% of apps |
| Setups that include plan switches | 38% of apps |
| Primarily text | 13% of apps |
| Median save rate | 17% |
| Promotional offer acceptance | about 1.1% |
| Plan-switch offer acceptance | about 0.6% |
A type counts only if it is at least 15% of that app’s messages, and text may include an image. Together, discount and plan-switch setups cover 87% of apps. Subscribers saw about 4.3 million promotional offers and 1.3 million plan-switch offers. On June–September data, when tracking was complete, they accepted about 1.1% of promotional offers and 0.6% of plan switches.
Did Those Saves Cut Cancellations?
Most recorded saves happen with no offer accepted: the person does not cancel within 24 hours either. RevenueCat’s possible explanation is that subscribers are still deciding, so a message about features they would lose is worth testing next to a discount. RevenueCat also says low acceptance does not necessarily mean the discount needs to be bigger. Price may not be the issue.
RevenueCat looked at 64 apps with at least 500 paying customers and compared paid cancellations, excluding trials, for three months before launch versus the launch month and two months after. Rates fell at least 10% at 34% of apps using the feature, against 46% of comparable apps without it.
After launch, cancellation rates fell 3.6% with the feature and 7.3% without it. Both groups were already trending down, and RevenueCat says a gap this size could be chance. The data does not show a measurable effect either way.
What Should Developers Ship First?
Among App Store apps on RevenueCat, 44% of apps making $1 million or more a month have tried it, 17% in the $100,000–$1 million band, and below 5% of smaller apps. Smaller teams can request access. Once Apple approves access, RevenueCat handles the server side.
You still need to pass Apple’s performance test and get each message approved. App Store Connect alone does not offer the same targeting.
RevenueCat’s guidance for a first test is straightforward:
- Match the message to a known cancel reason from surveys, support, and usage.
- Test a feature reminder beside a discount and a plan switch.
- Do not show a discount to every cancel-screen visitor.
- Exclude free trials and prior offer redeemers when you can.
If every visitor gets a discount, RevenueCat notes, some may learn that starting to cancel is an easy way to a lower price.
Frequently asked questions
What is Apple retention messaging?
Developers can show text, an image, or a special offer in the auto-renewable cancel flow, set up in App Store Connect. API access is by request.
What save rate should teams expect?
The median among 309 apps was 17%, defined as no cancel within 24 hours or an offer taken.
How often are cancel-screen offers accepted?
About one in 100. Promotional offers were accepted about 1.1% of the time, and plan switches about 0.6%.
Did retention messaging cut cancellations?
No measurable effect showed up. Among 64 apps, rates fell 3.6% with the feature and 7.3% without it, and RevenueCat says that gap could be chance.
