Why habits outlast campaigns

The spike

Every spring, the team at a mid-sized meal-planning app ran the same campaign. A discount email, a push notification and a banner in the app, all timed for the first warm weekend. And every spring the dashboard did the same thing: a tall, satisfying spike in weekly active users, followed by a slope that ran straight back to where it started.

“We were very good at getting people to open the app once,” their head of lifecycle told me. “We were not good at getting them to open it the next Tuesday.” The campaign worked, in the narrow sense. It just didn’t last.

That gap — between a behavior you can provoke and a behavior that repeats on its own — is the most expensive misunderstanding in retention work. Campaigns are rented attention. Habits are owned.

What a habit is

In the research literature a habit is not simply a frequent behavior. It is a behavior that has become tied to a cue: a time, a place, a preceding action. Once the link is strong enough, the cue triggers the behavior with very little deliberation. You don’t decide to check the weather app with your first coffee. The coffee decides for you.

That definition has a practical consequence. If you want customers to return without being asked, you are not trying to make the product more exciting. You are trying to attach it to something that already happens in their week.

For the meal-planning team, that something turned out to be Sunday afternoon, when most of their loyal users were already writing a shopping list. The customers who stayed a year had, almost without exception, planned at least three Sundays in a row in their first month.

“

We stopped asking how to bring people back and started asking what they were already doing when they came back on their own.

Designing for the cue

The team made three changes. None of them was a campaign. First, the onboarding flow asked one new question: when do you usually do your food shopping? Second, the weekly reminder moved from a fixed Monday send to an hour before the customer’s own answer. Third, the plan screen opened on the shopping list, the thing people were there to make, rather than on a gallery of recipes.

The reminder changed the most. A notification that arrives at the moment the customer is already reaching for a pen is not an interruption. It is a shortcut. Open rates on the new reminder were lower than the old campaign emails, but the share of customers planning in three consecutive weeks rose steadily for the next two quarters.

What the curve shows

Habits are slow to show up in metrics. The first month after the change looked flat; weekly actives barely moved. The retention curve was where it showed: the line for new cohorts stopped falling at week six and stayed there, a few points higher than any cohort before it.

That flattening is the signature of a habit. A campaign lifts the whole curve for a week. A habit changes its shape.

The honest caveats

Not every product has a natural cue, and some should not try to manufacture one. A tax tool used once a year does not need a weekly habit; it needs to be easy to find again in April. And habit design slides easily into compulsion design. The team drew their line at a simple question: would the customer, asked directly, want to be reminded at this moment? If not, the reminder went.

The spring campaign still runs. It is smaller now, and it sends people straight to the Sunday plan.