Push Notification Strategy: Deliverability, Timing, and Opt-In Design
Most clients ask us to 'just add push notifications' as if it's a single feature rather than a system with three separate failure modes: users never opt in, opted-in users never receive the notification, or received notifications get ignored. We treat each as a distinct problem with its own metrics, because fixing one doesn't fix the others.
On opt-in, the single biggest lever we've found is not asking immediately on first launch. Every client we've A/B tested this with sees meaningfully higher opt-in rates — typically 15-25 percentage points — when we delay the system permission prompt until after a user completes a meaningful first action, paired with a pre-permission screen that explains the specific value (order updates, appointment reminders) rather than a generic 'stay in the loop' pitch.
Deliverability is where a lot of teams assume FCM and APNs are reliable pipes and stop investigating. They're not, particularly on Android where OEM battery optimization (Samsung and Xiaomi are the worst offenders we've measured) silently kills background delivery for apps not on the manufacturer's allowlist. For one client we added explicit onboarding guidance directing users on affected OEMs to disable battery optimization for the app, which recovered a measurable chunk of notifications that were previously just vanishing.
Timing and frequency decisions need actual data, not intuition. We instrument notification-level analytics — sent, delivered, opened, and the downstream action taken — for every notification type, then cut or consolidate categories that show high send volume with low open rates. One retail client was sending four separate promotional notification types per week; consolidating to one curated digest per week raised open rate from about 4% to 11% while cutting opt-out rate roughly in half.
Rich notifications and notification service extensions on iOS, and expandable/grouped notifications on Android, are worth the extra implementation time for anything transactional — order status, delivery tracking — because they let users act without opening the app, which we've found correlates with lower uninstall rates even though it's a harder metric to attribute directly.
The pattern across every client: push is a channel that degrades quickly if you treat every message as equally urgent. Segmenting by actual user intent and measuring at each stage of the funnel, rather than just tracking 'push enabled' as a vanity metric, is what actually moves retention.
