Push notification opt-in rates have been declining for years, and the reason is not that users hate being notified. It is that most apps have trained them to expect noise instead of value. A retention feature that was supposed to bring users back has, for many products, become the reason they disable notifications entirely within the first week. Fixing this is less about clever copywriting and more about rethinking when, why, and how often a notification actually earns its place on someone's lock screen.
The instinct to send more notifications to drive more opens is understandable and usually backfires. Every notification a user finds irrelevant makes the next one, even a genuinely useful one, more likely to be ignored or to trigger an opt-out. Push strategy is really an attention budget problem: each app gets a limited number of notifications a user will tolerate before disabling them entirely, and the job of a good strategy is to spend that budget only on notifications the user would genuinely want, even if they never say so explicitly.
This budget is not fixed across users. A highly engaged daily user might tolerate several relevant notifications a week without friction, while a casual user might only tolerate one meaningful nudge a month. Treating both users identically is one of the most common reasons push campaigns underperform.
The best push notification is the one a user is glad they saw. Every other kind is spending down a budget that does not refill.
For example, a food delivery app might send a generic "we miss you, order now" notification to every user inactive for seven days, regardless of why they went quiet. A more targeted approach could instead differentiate: a user who abandoned a cart mid-checkout gets a notification about their specific unfinished order, a user whose favorite restaurant just came back online gets a notification tied to that restaurant, and a user who churned because of a bad delivery experience gets no promotional push at all until a service recovery message has been sent first. This kind of segmentation, tying the notification to a specific, relevant trigger rather than a blanket re-engagement message, is generally associated with meaningfully better response rates than one-size-fits-all campaigns, though the exact improvement depends heavily on the product and audience.
None of this works without solid technical foundations. Reliable delivery depends on correctly implemented platform-specific push services (APNs for iOS, FCM for Android), a permission request flow that asks at the right moment in the user journey rather than immediately on first launch, and, increasingly, deep linking so a tapped notification takes the user directly to the relevant screen instead of a generic app home. A notification that requires three extra taps to reach its actual content loses much of its value regardless of how well it was targeted. Teams building this from scratch benefit from getting the underlying mobile app architecture right early, since retrofitting proper deep linking and segmentation infrastructure onto an app that was not built with it in mind is considerably more work than including it from the first release.
iOS requires an explicit opt-in prompt, and once a user declines, re-prompting requires them to manually change a system setting. This makes the moment of the first permission request disproportionately important; apps that ask immediately on launch, before demonstrating any value, generally see substantially lower opt-in rates than apps that wait for a meaningful first moment.
Android has historically granted notification permission by default in many versions, though this has been tightening in recent releases to more closely resemble iOS. Strategies built assuming Android users are automatically opted in should be revisited regularly as platform defaults continue to shift.
Modern push notification systems support more than a title and a body. Rich media, an image thumbnail of the product a user abandoned in their cart, or a preview of new content, tends to draw more attention on a crowded lock screen than plain text. Interactive actions, buttons that let a user confirm an order, snooze a reminder, or dismiss an offer directly from the notification, reduce the friction of acting on it, which matters because every additional tap between a notification and its outcome is a chance for the user to lose intent and move on. These features are easy to under-use simply because they require slightly more implementation effort than a basic text push, but for high-value notification types, checkout reminders, appointment confirmations, delivery updates, the extra effort tends to be worth it.
A push strategy is only as good as the team's ability to measure its actual effect, not just its surface-level engagement numbers. Beyond open rate and opt-out rate, useful metrics include the conversion rate of the specific action a notification was designed to drive (completing a purchase, returning to a lapsed session, confirming a booking) and the incremental lift measured against the holdout group described earlier. For example, a team might discover that a notification type has a respectable thirty percent open rate but drives almost no incremental purchases beyond what would have happened anyway, in which case continuing to send it mostly just spends down the user's attention budget for no measurable return. Retiring low-ROI notification types, even ones with decent open rates, is often one of the highest-leverage changes a growth team can make to its push program.
Push notifications remain one of the most powerful retention tools available to a mobile product, but only for teams disciplined enough to treat the channel as a scarce resource rather than a free marketing megaphone. The strategy that wins is rarely the one that sends the most notifications; it is the one that sends the fewest notifications a user would genuinely regret missing. Getting that balance right requires segmentation, timing, and a willingness to measure opt-out rates as seriously as open rates, but the payoff is a channel that keeps working for years instead of one that quietly gets muted within the first month of a user's relationship with the app.