Why this matters
Activation gets a user to value once. Retention depends on whether they come back without you having to buy them again. Products that rely on reminders and discounts for every return visit carry a cost on every use. Products that become a habit get repeat usage almost for free.
A habit is a behaviour done with little conscious thought, set off by a cue. For a product, that means the user reaches for it when a specific situation arises: the team message arrives, the month closes, the coffee runs low. The design question is how to get from “I have to remember to use this” to “this is what I do when that happens”.
The framework
The most widely used model for product habits is the Hook model, popularised by author Nir Eyal. In our own words, it describes four repeating phases.
| Phase | What happens | Design question |
|---|---|---|
| Trigger | Something prompts the behaviour. External triggers are notifications, emails, an icon on the home screen. Internal triggers are feelings or situations, such as boredom, uncertainty or a task deadline. | What situation should make the user think of us? |
| Action | The simplest behaviour done in expectation of a reward. | Can the user do it in seconds, on the device they are holding? |
| Reward | The payoff. Eyal stresses variable rewards, where the exact outcome is not fully predictable, which keeps attention. | Does the user get something they value every time, plus something new some of the time? |
| Investment | The user puts something in (data, content, effort, reputation) that makes the product better for them next time and sets up the next trigger. | What does the user leave behind that pulls them back? |
The loop matters because investment loads the next trigger. A saved draft creates a reason to return. A follower base brings notifications. A streak count becomes something to protect.
A daily learning habit loop
Each step feeds the next — step 4 (“Investment: streak and progress saved”) loops back into step 1 (“Trigger: evening practice reminder”). Tap a step to trace it.
Natural usage frequency
Before designing any loop, find how often your product is genuinely needed. This is the natural usage frequency. It is set by the user’s underlying problem, not by your wishes.
| Product | Natural frequency | Habit unit |
|---|---|---|
| Team chat, language learning | Daily | Active days per week |
| Project management, analytics | Several times a week | Active days per week |
| Expense reports, social scheduling | Weekly | Active weeks per month |
| Payroll, invoicing, accounting close | Monthly | Active months per quarter |
| Coffee, protein, skincare refills | Every 3 to 6 weeks | On-time reorders |
Habit frequency fit is the match between the loop you design and that natural rhythm. A daily loop on a monthly product creates empty visits and notification fatigue. A monthly loop on a daily product lets the habit fade.
Step-by-step playbook
-
Measure the natural interval. Take users retained at month 3. For each, compute the days between consecutive core actions (not logins). Take the median per user, then the median across users. For D2C, use days between orders for repeat buyers, or pack size divided by daily usage.
-
Choose the habit metric. Pick the period that matches the interval (day, week, month, reorder cycle) and set a consistency threshold, such as active in 3 of the first 4 weeks. This is your habit rate.
-
Map the internal trigger. Interview retained users: “What was happening right before you last opened the product?” Look for a repeating situation or feeling. That is the internal trigger you want to own.
-
Design the external trigger to point at it. Time reminders to the situation, not to your calendar. A finance tool reminds on the last working day of the month. A coffee brand messages when the bag is likely running low. Ask permission well, cap frequency, and give each message one clear action.
-
Make the action tiny. Deep link straight to the action. Remove logins where safe. Aim for a first meaningful action in under 60 seconds from the trigger.
-
Design the reward. A reliable core payoff (the report is ready, the order arrives on time) plus some variety (a new insight, a reply from a teammate, a milestone). Celebrate progress, especially early, when each step is a large share of the total.
-
Ask for an investment after the reward. Save a preference, add a teammate, rate the product, set the next goal. Ask after the reward, when goodwill is highest.
-
Add flexibility and an exit. Let users pause, skip a cycle, freeze a streak or change frequency. A habit that breaks completely after one miss turns into a churn event.
-
Run the ethics check (below) before launch, then test against a holdout and watch both habit rate and opt-outs.
Worked example
Say a B2B expense app wants a weekly habit. Its retained users submit receipts a median of every 6.5 days, so the natural unit is weekly.
Baseline, for a cohort of 2,000 new users:
- Active in at least 3 of their first 4 weeks: 520 users. Habit rate = 520 ÷ 2,000 = 26%.
- Month-3 retention among habit-formers: 74%. Among the other 1,480 users: 12%.
- Month-3 retained users: (520 × 0.74) + (1,480 × 0.12) = 384.8 + 177.6 = 562.4, or about 28% of the cohort.
The team designs a Friday loop. Trigger: a 4 pm Friday push, “3 receipts waiting from this week” (the app reads card transactions). Action: snap receipts from a single screen. Reward: “Your claim is ready to submit” and an estimate of when the reimbursement lands, plus a monthly summary of savings. Investment: the user’s categories and approvers are remembered, so next week is faster.
After a test, the habit rate in the test group rises to 33%. If the retention rates for each group hold (a check you must run, not assume):
- Habit-formers: 2,000 × 0.33 = 660. Retained: 660 × 0.74 = 488.4.
- Others: 1,340. Retained: 1,340 × 0.12 = 160.8.
- Total retained: 649.2, or about 32.5% of the cohort, versus about 28% at baseline.
That is roughly 87 more retained accounts per 2,000 signups (649.2 − 562.4 = 86.8). The team also watches push opt-out rate. It rose from 4% to 5%, inside the guardrail they set before the test, so the loop ships.
B2B SaaS vs D2C: what changes
| Element | B2B SaaS | D2C / e-commerce |
|---|---|---|
| Natural frequency | Set by work rhythm: daily standups, weekly reports, monthly close | Set by consumption: pack size and usage rate |
| Best triggers | Teammate activity, workflow deadlines, integrations (a Slack or email digest) | Depletion timing, delivery updates, WhatsApp and SMS reminders |
| Action | Do the work in the product, ideally from the notification | One-tap reorder, saved address and payment |
| Reward | Work done faster, recognition from teammates, new insight | Product arrives on time, visible progress (routine results, points) |
| Investment | Data, templates, integrations, teammates invited | Preferences, subscription settings, loyalty points, reviews |
| Common failure | Daily nudges for weekly work | Discount-driven reorders that ignore the actual cycle |
In D2C, the habit is often the product’s own use, not visits to the store. A skincare brand’s job is to make the morning routine stick, and the reorder follows. In India, WhatsApp is often the most effective external trigger for D2C, but the same rules apply: get consent, time it to the cycle and give one clear action.
The ethics of habit design
Habit design can help people do things they already want to do more consistently. It can also be used to capture attention against their interests. Run these checks before shipping any loop:
- Goal alignment. Does the loop help the user reach a goal they already have? Would they agree if you described the loop to them plainly?
- Would you use it? Would the team be comfortable with their own families in this loop?
- Easy exit. Can users pause, change frequency or turn off triggers in two taps?
- Honest triggers. No fake urgency, no invented social activity, no guilt-tripping copy.
- Outcome metric. Pair every engagement metric with an outcome metric (lessons learned, expenses filed, skin goals reached) and review both together.
Duolingo’s own published work shows the useful direction: making the streak more flexible and easier to keep (Streak Freezes, separating streak from daily goal) increased engagement rather than reducing it.
How to measure it
- Cohort and window. Measure habit rate on signup or first-order cohorts, over the first 4 to 8 natural periods.
- Active-days distribution. For daily or several-times-a-week products, plot how many users were active 1, 2, 3 and up to 7 days in a week. A healthy product shows a meaningful group at the high end that grows over time.
- Internal trigger signal. Track the share of sessions not preceded by a notification or email in the previous hour. A rising share suggests users come back on their own.
- Guardrails. Notification opt-out rate, unsubscribe rate, uninstall rate and support complaints. Set thresholds before you launch.
- Segments. Persona, plan, channel, platform. Habits often form in one segment first.