July 2026 · Anthony Sogg

DAU Is a Retention Metric Wearing a Growth Costume

Daily active users look like a growth metric. They're a retention metric wearing a costume, and three public companies prove it in their own numbers.

The most expensive confusion in consumer software is treating daily active users as something you acquire. You don't acquire a daily user. You acquire an installer. The daily part is manufactured later, or it never happens.

Three public companies make the case better than any framework I could draw, so let's read their own numbers.

Exhibit one: the greatest acquisition event in consumer history, and what it bought

In July 2023, Threads signed up 100 million people in five days. Meta said so itself, proudly, and it was true. It's still the fastest product launch ever recorded. Within a month, third-party trackers like Sensor Tower and SimilarWeb were reporting that daily activity had collapsed by roughly four-fifths.

One hundred million people walked into the store, looked around, and left. The store called it a grand opening.

Nothing about the acquisition failed. Distribution was perfect: a billion-user parent app, one-tap signup, wall-to-wall press. What failed was the part nobody had built yet: a reason to come back tomorrow. Acquisition without a retention loop isn't growth. It's a parade.

Exhibit two: the company that does it on purpose

Duolingo reported 56.5 million daily users against 137.8 million monthlies in its most recent quarterly letter, a daily-to-monthly ratio north of 40 percent, which is absurd for a product whose core activity is homework.

None of that is luck, and very little of it is advertising. It's engineered: streaks, leagues, widgets, the passive-aggressive owl. Duolingo treats the daily habit as the product and the lesson as the payload. Their growth team is, functionally, a retention team with better branding. The market keeps calling this gamification. I'd call it lifecycle marketing that finally got a seat at the product table.

Exhibit three: half a billion people in the waiting room

Snap's Q1 2026 investor letter reports 956 million monthly actives and 483 million dailies. Subtract. Roughly 473 million people already installed Snapchat, already have friends on it, and don't open it on any given day. Convert one percent of that pool to a daily habit and you've added about 4.7 million daily users, real growth, without spending one acquisition dollar.

The cheapest daily user on earth already has your app. They're not a prospect. They're a lapsed relationship, and lapsed relationships aren't won back with reach. They're won back with propensity models, triggers worth answering, and the discipline to stay silent when silence is what keeps the door open. In my world we call that win-back and suppression. Growth teams tend to call it boring, right up until the quarter it carries.

The costume comes off

Here's the uncomfortable arithmetic underneath all three stories. A daily active user is a compound object: acquisition times activation times habit times survival. Most growth orgs are staffed, budgeted, and promoted against the first term, because the first term is the easiest to buy and the easiest to screenshot. The other three terms are where the money lives, and they're classical CRM disciplines wearing nobody's job title: onboarding journeys, behavioral triggers, defection-risk modeling, reactivation economics, measured against holdouts or not measured at all.

I've run those disciplines for years at tens-of-millions-of-customers scale in industries where a customer's next purchase was years away and the relationship had to survive the gap. Long-cycle categories learned this lesson early because they had no choice: you can't re-acquire your way out of a relationship business. Consumer software is learning it now, one flat DAU chart at a time.

So the next time a growth deck opens with an acquisition funnel, ask the question the funnel is dressed up to avoid: of the people who already came, how many came back yesterday, and what, specifically, did we do to deserve it, and how do we do more of it better?

That answer is the growth strategy. Everything else is the costume.

Anthony

← More writing · If this was useful, or wrong in an interesting way, tell me: hireme@crmjedi.com.