For most of the last decade, consumer product teams have been trained to think about retention through one model: the habit loop. A cue triggers a routine, the routine produces a reward, and over enough repetitions the user stops thinking about the action altogether. The model is elegant, easy to teach, and has shipped billions of dollars of revenue. It has also produced a generation of products that are easier to start than to stop.
We think it is time to retire the habit loop as the default mental model for engagement, and to replace it with something older and more honest: the ritual.
Habits are automatic. Rituals are chosen.
A habit is something you do without deciding. A ritual is something you decide to do, often the same way, often at the same time, because the doing of it means something. Brushing your teeth is a habit. Sunday dinner is a ritual. The difference is not the frequency or the consistency — both can be daily, both can persist for decades. The difference is agency.
When a product is designed as a habit, the user's attention is the raw material and the loop is the machine that consumes it. The metric of success is whether the user keeps showing up, ideally without thinking. When a product is designed as a ritual, the user's intention is the raw material and the loop becomes a container for something the user wants to be doing. The metric of success is whether the user keeps choosing to show up, with their thinking intact.
There is a meaningful difference between a user who couldn't stop and a user who wouldn't want to. We have spent a decade building for the first. We should spend the next building for the second.
Where the habit model breaks
The habit-loop framing was useful when the dominant question was 'how do we get people to come back at all.' It is less useful, and increasingly counterproductive, in a market where attention is contested and trust is scarce. Streaks generate anxiety, not loyalty. Variable rewards train users to discount the product the moment the variability stops paying off. Engagement-as-addiction is a posture the culture is no longer willing to absorb without comment, and regulators are catching up to the pattern.
More practically, habit-driven products tend to plateau in a particular way: usage is high, satisfaction is flat, and the moment a credible alternative appears, the cohort is gone.
The behavior was sticky; the relationship was not.
Designing for ritual
Designing for ritual is not a softer version of designing for habit. It is a different brief, and it asks for different choices:
- Make the opt-in explicit. Ritual begins with consent. The user should know what they are signing up for, and the product should treat that knowing as a feature, not a friction.
- Honour pause states. A ritual that cannot be skipped without guilt is a coercion. Build graceful 'not today' affordances and treat them as healthy, not as churn risk.
- Build narrative continuity. Rituals carry meaning across sessions. Show the user the thread — what they did last time, what this session is part of — so the action accumulates into a story rather than a streak.
- Design real off-ramps. A product the user can leave without punishment is a product the user is choosing to stay in. The off-ramp is a feature of the relationship.
- Reinforce identity, not usage. The reward at the end of a ritual is who the user is becoming, not how many sessions they have logged.
What changes when you measure rituals instead of habits
If you take this seriously, your dashboard changes. You stop optimising for daily active sessions and start watching the share of returns that are unprompted by a notification. You watch the proportion of sessions that end at a natural stopping point rather than a forced one. You watch recommendations and word-of-mouth, because rituals are the kind of behaviour people talk about. You watch depth on return, not just frequency of return. The numbers move more slowly, and they tell you something more truthful when they move.
We are not arguing that habits are bad or that loops are wrong. We are arguing that the loop is a tool, and that a tool optimised for one outcome — automaticity — is a poor default for products that want to be in their users' lives for a long time. Rituals are slower to build and harder to A/B test. They are also what users mean when they talk about the products they actually love.
The next decade of consumer product design will, we suspect, be a slow migration from one model to the other. The teams that get there first will discover that the most retentive thing a product can do is be worth choosing.