Every PM in mobile knows the feeling. The memories feature is on the roadmap. The business case is obvious. But each sprint, something more urgent takes priority: a bug that's affecting conversion, a platform deadline, a feature the sales team promised. The memories feature moves to next cycle.

The problem with deprioritising it isn't the feature itself. It's what happens to retention while you wait.

Users who have something to come back for, come back

Mobile retention follows a predictable curve. Day 1 is high. You've just acquired someone motivated enough to install. Day 7 starts to separate good products from average ones. By Day 30, most apps have lost 80–90% of their initial install base.

The apps that hold users past Day 30 share a trait: they give users a reason to return that isn't just "check for new content." A memories feed is one of the most effective versions of this. The content is the user's own moments, so the motivation to return is personal, not a response to being nudged by a notification.

Retention is comparative. You're not just competing with other apps in your category. You're competing with every app on the user's home screen for the habit of opening yours.

The gap compounds while you delay

While the feature sits on the backlog, the apps competing for the same users are shipping. Instagram, Snapchat, and a generation of newer apps have normalised the expectation: your content lives in the app, and you come back to see it.

Users who have that experience elsewhere become less tolerant of apps that don't offer it. The bar moves whether you build or not.

What deprioritisation actually costs

The cost isn't just a missing feature. It's the compounding effect on the metrics that drive every downstream decision in your company. Lower Day 30 retention means:

• Higher effective CAC. You're re-acquiring lapsed users instead of retaining active ones. The math on payback period gets worse.
• Weaker engagement data. Fundraising, partnership conversations, and growth projections all lean on retention numbers. Six months of 5% lower Day 30 retention compounds into a materially different story.
• Slower word-of-mouth. Users who return regularly tell people. Users who churned don't.

The decision hiding inside the build estimate

The reason the feature keeps getting pushed is almost always the same: the build cost looks prohibitive. A video memories feature done properly, with real caching, smooth seek, device fragmentation handled, and analytics wired up, is 3 to 4 months of genuine engineering work. That estimate is why it keeps losing to shorter-horizon items.

The decision you're actually making when you deprioritise it isn't "should we build this." It's "is the build the only way to ship this." Those are different questions.