Most apps get downloaded, opened once, and never touched again. That's the normal outcome, not the failure case, and it happens to good apps with real features.
The apps that survive have something specific going on underneath, and it's the same shape in almost every one of them.
The loop
Four parts, in a circle.
Something brings the person back. A notification, an email, a habit they've formed, or an itch they get at a particular moment. Without this, they simply forget you exist, and forgetting is the default.
They do the thing. Whatever your app is actually for. Log the workout, send the invoice, check the number.
They get something out of it. Progress, relief, a small hit of satisfaction, information they wanted. If this part is weak the loop breaks here, because there's no reason to run it again.
And they leave something behind. Data, history, a streak, a customised setup, saved work. This is the part beginners skip, and it's the one that makes the next visit more likely than the last.
That fourth part is worth sitting with. Each time someone uses your app it should become slightly more theirs, and slightly more annoying to abandon. Not through lock-in tricks, but because they've accumulated something real. Three months of data is genuinely valuable to the person who created it.
Where the loop lives
Here's the bit that connects to Chapter 2. The loop isn't a design decision you apply later, it's mostly a data decision you make early.
If your app is meant to show someone their progress over time, that requires recording what they did and when, from day one. Bolting it on afterwards means you have no history to show, and you can't invent it.
So when you're sketching tables and someone asks what you need to store, the answer includes "whatever the loop needs to show them next week."
When there's no natural loop
Some apps genuinely don't have a daily reason to return, and pretending otherwise produces the app everyone mutes.
A tax calculator gets used once a year. A tool for renaming files gets used when you have files to rename. These aren't broken, they're differently shaped, and the mistake is forcing engagement onto them.
For those, the honest move is to be excellent at the moment of need and easy to find again. The measure of success isn't daily use, it's whether people come back the next time the need arises, and whether they tell someone.
Trying to make a once-a-year tool into a daily habit produces notifications people resent, and that's worse than a low but genuine usage pattern.
The line you shouldn't cross
Everything above can be used well or badly, and the difference is whether the person is better off.
A streak that motivates someone to do a thing they actually want to do is good design. A streak engineered so that breaking it feels like loss, on something they were ambivalent about anyway, is manipulation.
The test I'd use: does this help them get the thing they came for, or does it work against their interest to serve my numbers?
Beyond ethics, the manipulative version doesn't even work for long. Retention bought with guilt collapses the moment someone's attention lapses, and it comes with one star reviews explaining exactly why. Retention that comes from the thing being useful is the only kind that compounds.
A quick story. My own app was built almost entirely around a loop. Daily lessons, streaks, visible progress over ninety days. The whole product thesis was that people would come back because they could see themselves changing.
I built all of it carefully. Then I never launched it for long enough to find out whether any of it worked.
Which is the part worth taking from this. You can design a loop beautifully on paper, and it stays a theory until real people either come back or don't. The design is a hypothesis. The retention numbers are the answer, and you only get them by shipping.
If you want to work out your loop
Point three is the one people skip. Plenty of good software gets used rarely, and knowing that early changes what you build and how you'd measure success.
Next: The first minute someone uses your app →
Go deeper: Chapter 9, The Core Loop and Chapter 14, Behavioral Design and Ethics