The deposit and the retrieval

If you're building a product that asks the user to put something in regularly, a journal, a habit tracker, a mood log, a health metric, a workout, a meal, anything where the user is doing the work of recording, you're going to run into the same problem every founder in this category runs into. You'll get great signups, decent week-one retention, and a steep dropoff after that. You'll add streaks to fight the dropoff. The streaks will help a little. Then they'll stop helping. Then you'll spend a year trying to figure out what's wrong with the product, and the answer the whole time will have been that you confused the deposit with the value.

The deposit is the logging. It's the moment the user opens the app and types the thing or taps the thing or records the thing. From the product's point of view this looks like the activation event. From the user's point of view it's a cost. They had to remember, they had to open the app, they had to think for a second, they had to do the input. They paid for something. The product owes them something back.

The retrieval is what the product owes them. Retrieval is the moment the user gets back something they couldn't have seen for themselves. A pattern across weeks they wouldn't have noticed. A connection between two things they wouldn't have linked. A summary that's actually a summary, not just a list of what they typed in. The retrieval is the entire reason to log anything. Without it, the user is just maintaining a database for you and getting nothing back, and no streak in the world will keep them doing that past the point where the novelty wears off.

Streaks work in the early days because the user hasn't tested the retrieval yet. They're depositing in good faith, expecting the value to show up later. The streak gives them a reason to keep depositing while they wait. When the retrieval doesn't show up, the streak becomes the only reason to log, and once the user breaks the streak they stop. The streak was never the engagement mechanism. The streak was a bridge to the retrieval, and if the retrieval isn't on the other side of the bridge the bridge ends in a wall.

You can tell which side of this your product is on by looking at what your most engaged users do in the app after they log. If they log and leave, you have a deposit-only product and you're in trouble. If they log and then spend time reading something the product generated, scrolling through history, looking at a pattern, getting a recommendation that uses the data they just entered, you have retrieval and the streak is doing what it's supposed to do.

The mistake founders make is treating the retrieval as a feature to add later, after the logging is solid. The logic is that you need data before you can show patterns, so you build the input first and the output second. That logic is right about the data and wrong about the product. By the time you build the output, the user has churned. You needed the output on day one even if it was thin, because day one is when the user is deciding whether the deposit is worth making. A thin retrieval on day one beats a great retrieval in version three that nobody is around to see.

The smallest version of retrieval that works is usually not analytics. It's reflection. Show the user something they wrote a month ago that they'd forgotten. Surface a quiet pattern across the last two weeks in one sentence. Hand them back a question that connects today's entry to one from earlier. The mechanism doesn't have to be sophisticated. It has to make the user feel that the deposit produced something they couldn't have produced themselves, and that the something showed up at a moment that felt useful rather than scheduled.

If you take one thing from this, take this. Build the retrieval first, even if it's thin. Then build the input around it. The deposit serves the retrieval. When you build it the other way around, you get a product full of data and empty of users.

This connects to a question I answer in depth: What should I build next as a solo founder and Why do people say they like my idea but don't buy.