The Streak Is the Product

Every habit app worked something out years ago, and it is worth saying plainly because the whole industry is built on it. A result is expensive to deliver. Fitness takes months and most people give up, a language takes years and nearly everyone does, and focus is harder than either because nobody can quite agree what it looks like when you have it. A streak, on the other hand, costs nothing at all. It is a number that goes up by one each day, it requires no outcome of any kind, and the only thing it actually does is make you afraid of the day it stops.

That fear is the business model. I do not say that as a complaint, because I spent a long time on the other side of it. In boardrooms the number on the slide is called retention, and retention is the metric everything else is forgiven for: a product that keeps people coming back can be unprofitable, unfinished and faintly disliked by its own users and still be worth a great deal of money, while a product people love and use twice is worth almost nothing. I have sat through enough of those slides to know what a retention curve looks like from the inside, and to know what the team is asked to do when it sags. They are not asked to make people fitter. They are asked to make leaving feel expensive.

A streak is a retention metric with your face on it. It takes the thing the company most needs, which is for you to open the app tomorrow, and it hands the job of enforcing that to you, at no cost to them, using the one lever that works on capable people better than any other, which is the horror of a visible failure. The genius of it is that it never looks like coercion. It looks like your own discipline, celebrated. Seven days in a row, with a small animation. Thirty days, with a badge. A hundred, and a notification that you are in the top few per cent of users, as though that were a fact about your character rather than a fact about their funnel.

You will recognise what it does to the day, because it is the same thing in every household I know. It is eleven at night. You have not done the thing. You are tired, the thing would take ten minutes done properly, and so you do it in two, badly, half asleep, so that the number survives. The lesson is not learned and the run is not run. The streak, which was supposed to be a record of the habit, has quietly replaced the habit, and the two minutes at eleven were never for you. They were a payment, made in your attention, to keep a number alive on somebody else's dashboard.

I want to be honest about my own position here, because I have no business lecturing from a height. I took the trackers off and the rings and the targets went with them, but I kept exactly one streak, on Duolingo, and I am not proud of it. It survives because the cost of losing it has somehow become more than I am willing to pay, which is precisely the mechanism I have just described, working on a man who understands it perfectly well and has written about it in public. Knowing how the trick is done does not stop it working. That is worth sitting with, because it tells you that the answer is not going to be willpower.

The useful question, when something congratulates you on a run of days, is who the days were for. There is a simple test. Imagine the number reset to zero tonight, through no fault of yours, a server somewhere losing its memory. Would you do the thing tomorrow anyway? If the answer is yes, the streak was never doing anything; the habit is yours and the counter is decoration. If the answer is no, then you have learned something more important, which is that you do not have a habit at all. You have a hostage situation, and the thing being held is ten minutes of your evening, every evening, indefinitely.

I have come to think the people I work with are especially exposed to this, because they are already measured all day and have learned to respect a number that goes up. Revenue, pipeline, headcount, the market, their own ambition: every one of those is a counter, and most of them are legitimate. So when a counter appears in their private life, dressed as self-improvement, they extend it the same respect without asking the question they would ask of any metric at work, which is what behaviour this measure actually produces. At work they know that a measure which produces late-night, low-quality compliance is a bad measure and should be taken off the dashboard. At home they protect it.

I am building two small things at the moment for people with too much on, and I have said before that the design rule was fixed early. Neither of them has a streak. If you miss a day, nothing happens, and I mean that literally: no reset, no notification, no sad face, no number that was once large and is now one. That rule cost us something, because a streak is the cheapest engagement a product can buy and leaving it out is leaving money on the table. I would rather leave it there. A tool that is meant to give you your days back cannot also be the thing that holds them to ransom.

So the practice this week is a short piece of accounting. Write down every streak you are currently keeping, including the ones you would be slightly embarrassed to admit to: the language app, the meditation app, the step count, the closed rings, the daily anything. Next to each, answer the test without flattering yourself. Would you do it tomorrow if the number reset tonight? Then, for one of the ones that got a no, let it go. Not by deleting the app or making a speech about it. Just do not do the thing tomorrow, watch the number fall to zero, and notice, in the quiet that follows, that nothing at all has happened. The habit you actually wanted is still available to you on Thursday, and for the first time in a while it will be yours.

Or start the diagnosis: Take the free 10-minute Entrepreneur Gap scan →

The book: The Science of Getting Rich and How to Apply It →