At 9:14 on Monday, Maya signs up for a planning app. At 9:18, she has chosen a workspace colour, invited nobody, and earned a shower of confetti for completing onboarding.
On Tuesday morning, she writes her tasks on a sticky note.
The app got her through its checklist. It never gave her a reason to put it in her life.
Maya is an imagined user. The studies and product examples below are real.
That quiet gap is where a lot of products lose people. The first visit can run on curiosity, a friend's recommendation, or a persuasive landing page. The second needs something the first visit actually delivered.
By then, the interesting question has changed. It was “What does this do?” Now it is “Why would I reach for this again?”
A second visit is too early to call it a habit
Imagine trying a café for the first time. The room looks beautiful. The menu is clever. None of that tells us whether you will come back on Thursday.
But perhaps the coffee was excellent, the barista made ordering easy, and the café is on the way to your train. Thursday now has somewhere to go.
Products need an equivalent: an experience worth repeating, a moment when it becomes relevant, and an easy way back. This is a design lens, not a validated psychological formula. It helps separate three problems that often get bundled into “engagement”.
Habit formation takes repetition. In Lally and colleagues' study, 96 volunteers chose an everyday behaviour to repeat in a consistent context over 12 weeks. Among participants whose data fitted the model well, estimated time to near-maximum automaticity varied widely, from 18 to 254 days. The upper end was a model estimate beyond the observation period, not 254 days of tracking.
A second visit can be deliberate, tentative, and easily interrupted. Design for someone who is still deciding whether this belongs in their day.
- This helped.Remembered value
- I need it now.A relevant moment
- Right where I left it.An easy return
The first win has to belong to the user
“Account created” is progress for your database. “I sent the invoice I was avoiding” is progress for the person.
That distinction sounds obvious until you look at onboarding. Upload a photo. Connect three services. Invite a teammate. Watch the tour. The product keeps asking for small commitments while the promised payoff moves further down the corridor.
In a 2001 study of online banking users, Anol Bhattacherjee found that satisfaction and perceived usefulness were associated with people's intention to keep using the service. Whether the experience confirmed their expectations mattered to that satisfaction. It was a survey about continued-use intentions, not a randomized test of onboarding or observed second visits.
The practical reading: people leave their first session with evidence about your promise. If the promise was “get your week under control” and the evidence is a decorated empty workspace, the next marketing email has a difficult job.
For Maya, a better first session might end with tomorrow's three priorities arranged and saved. She can close the tab with one small piece of tomorrow already handled.
What can the person do, know, or feel confident about now that they couldn't before opening your product?
Duolingo made the return smaller
There is a tempting response to weak retention: ask people to commit harder. Pick a bigger goal. Make a promise. Aim higher.
Duolingo tested a different move. Its streak used to depend on meeting a daily XP goal. The team separated the two, so completing one lesson could keep a streak alive even without reaching that goal.
In its published account of the A/B test, Duolingo reported a 3.3% relative increase in day-14 retention. That means a proportional lift, not 3.3 percentage points. The company also reported a cost: fewer learners reached their daily goals.
The useful lesson is narrower than “streaks work”. Reducing the minimum action made returning more achievable in that experiment. The results are company-reported, and day-14 retention is a different measure from a second visit.
Look at your own invitation back. Does it sound like a quick, useful action, or like taking on a second job?
Where do I even start?
I can do that before my coffee.
Give tomorrow a cue
Someone can like your product, intend to return, and forget it completely. Tomorrow contains meetings, shopping, messages, and the same old shortcuts they already use.
Psychologists Peter Gollwitzer and Paschal Sheeran studied a way to bridge that gap: implementation intentions. Instead of simply wanting to do something, a person links an action to a situation: “After breakfast, I will review five words.” Their 2006 meta-analysis of 94 studies found that these plans helped people achieve goals across varied settings. It does not tell us how much a reminder feature will lift app retention.
The design opportunity is to help a willing user choose a moment that already exists in their life. For Maya, that could be reviewing her priorities when she sits down at her desk. For an invoicing tool, it might be the end of the working week. Daily use is only appropriate when the need is daily.
Spotify gives discovery a day of the week
Discover Weekly refreshes every Monday. A recurring collection turns an open-ended promise, “there is music here”, into something with a date: there will be a new selection to explore.
That is an example of a predictable reason to return, not proof that Monday causes retention. It is also a mature-use example: Spotify says new listeners may need a few weeks of listening before personalized playlists appear. A product needs an earlier payoff while personalization is still learning.
For your own product, try finishing this sentence in the user's voice: “When ___ happens, I'll open this to ___.” If the only answer is “when we send a push notification”, look for the underlying need.
Leave something worth coming back to
There is a particular pleasure in opening a tool and finding your place. The draft is there. The shortlist is there. Yesterday's decisions still count.
Now imagine the opposite: a welcome tour you have already watched, an empty dashboard, or a saved project buried behind a generic home screen. Returning requires reconstructing what you were doing before you can do anything useful.
Readwise makes yesterday's effort useful again
Readwise's Daily Review resurfaces saved highlights through email or the app. The material waiting for you can include passages you already chose while reading. The product creates a new encounter with something you considered worth keeping.
Its review controls also let people adjust how frequently documents appear. Personal relevance and user control work together here. This is an observation about the feature's design; public documentation does not establish its effect on second-visit retention.
The broader opportunity is to preserve useful effort. A recipe app can keep the meal you saved for Saturday. A design tool can reopen the canvas at the right place. Maya's planning app can put her actual priorities on screen before offering another feature.
This is not a reason to make leaving costly. A useful saved object should help someone continue, with access and export appropriate to the product. Its job is to reward the return.
Start where you left off.
Yesterday, you chose what matters today.
A return can be earned. It can also be pressured.
Two people open an app at 10pm. One is curious about the next lesson. The other is afraid of losing a streak. An analytics event can make those visits look identical.
Ryan and Deci's self-determination theory distinguishes motivation people endorse from motivation driven by external pressure. Their account emphasizes autonomy, competence, and relatedness as important psychological needs. It gives us a reason to ask what kind of motivation an experience supports, rather than treating every extra opening as equally desirable.
In practice, let people choose a manageable goal, change reminders, and return after a missed day without a scolding. Lally's study also found that missing one opportunity did not materially disrupt habit formation. A broken streak counter and a broken habit are different things.
There is room for anticipation, too. Finish today's useful task, then make the next opportunity visible. A completed lesson can point to the next skill. A finished report can offer a comparison when new data arrives. You do not need to withhold today's answer to create tomorrow's reason.
Fix the final screen before adding another reminder
Open your product as a new user. Complete one useful task. Then inspect the moment just before you leave.
Show what changed
Replace “You're all set” with a concrete result: “Your first invoice is sent” or “Tomorrow's three priorities are saved.” Make the first win visible.
Make the next use specific
Name a useful next action and the situation where it belongs. For a weekly task, a weekly return can be success. Avoid manufacturing a daily obligation.
Check the actual way back
Close the tab and return later through the routes people really use. Check saved state, sign-in, and deep links. Does the person reach their work, or have to find it again?
Then test the change. Randomly assign new users to the current ending or the revised one. Define a meaningful second session in advance: a separate visit in a suitable time window that includes a useful action. A refreshed tab or a notification opened and immediately dismissed should not count as the same success.
Keep the denominator to all eligible users assigned to each version, allow the full observation window, and look at useful outcomes alongside return rate. Choose a duration and sample size suited to your traffic and the product's natural rhythm. People who save something may already be more motivated; comparing savers with non-savers alone will not show that saving caused the return.
Back to Maya. This time, she closes the app with tomorrow's plan saved. On Tuesday, her first task is waiting where she expects it. The product has made a small promise about her next working morning, then kept it.
That is a reason she can recognize the next time she sits down at her desk.
What the evidence can tell us
These sources study different things: habits in everyday life, intentions to keep using online banking, goal attainment, motivation, and a company's retention experiment. Together they inform the design argument; they do not validate a universal recipe for second visits. The Spotify and Readwise examples illustrate product mechanisms, with no claimed causal retention lift.
Read the research
- Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W. & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40, 998–1009. Consistent contexts, automaticity, and missed opportunities.
- Bhattacherjee, A. (2001). Understanding Information Systems Continuance: An Expectation-Confirmation Model. MIS Quarterly, 25(3), 351–370. A survey of online banking users; satisfaction, usefulness, and continuance intentions.
- Gollwitzer, P. M. & Sheeran, P. (2006). Implementation Intentions and Goal Achievement: A Meta-analysis of Effects and Processes. Advances in Experimental Social Psychology, 38, 69–119. Situation-linked plans across 94 studies.
- Ryan, R. M. & Deci, E. L. (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being. American Psychologist, 55(1), 68–78. A theoretical review of motivation and psychological needs.
- Duolingo (2020). Improving the streak. A company-reported A/B test; all reported lifts are relative.
- Spotify. Find playlists on Spotify. Official documentation of playlist schedules and personalization availability.
- Readwise. Reader and Daily Review; Reviewing Your Highlights. Official product and review-control documentation.
Sources checked 18 October 2026. Product examples describe documented features. Interface illustrations are original examples, not screenshots.