Build for the second visit.

Returning customers bring knowledge of the old interface. A redesign needs to justify the understanding it asks them to replace.

A first-time user needs to learn where things are. A returning user hopes they still know.

That difference changes what counts as a better interface. A redesign can make a product clearer to a newcomer while making familiar work slower for everyone who already learned it. The team sees the improvement in front of them. The returning customer discovers the cost when a hand reaches for a control that has moved.

There are good reasons to move it. Perhaps the old arrangement caused mistakes or concealed an important action. But familiarity has value that a fresh design review can overlook. Changing an interface spends some of the understanding people built through earlier use. The improvement needs to repay that cost.

The rhythm of returning

Imagine a tool used once a month to prepare a report. The person returning to it is unlikely to remember every detail. They may remember the sequence: choose the period, review the figures, export the file. Keeping that sequence legible helps them recover their understanding without relearning the product.

A tutorial repeated on every visit could make matters worse. It asks them to stop doing work they recognize so the product can explain itself again. The team may have solved its fear that people will forget by interrupting the people who remembered enough.

Visit frequency also needs context. A monthly task has a different rhythm from a daily habit. Encouraging another session between reports may create a retention number without creating a reason to use the tool. A valuable product can be absent from someone's life until it is needed again.

That argues for judging a redesign across the relevant interval of use. A first-session test can show whether a new sequence is understandable. It cannot show whether returning customers can recover their place a month later. Follow both newcomers and people who learned the previous version and distinguish momentary hesitation from errors that continue after the new arrangement is understood.

What a change asks people to relearn

Familiarity should not protect a bad design forever. A known sequence may be causing serious mistakes and leaving it alone would preserve those mistakes along with the habit. The decision needs to name the problem the redesign removes and the relearning it introduces. "Cleaner" is too vague to settle that exchange.

For a change that earns its cost, preserve what still helps: the person's data, a recognizable task sequence where possible and an explanation at the point where an old expectation will fail. Running two versions indefinitely can create its own confusion, so a transition needs an end as well as a beginning.

There is something valuable in opening a product after a month away and still knowing where to begin. The earlier effort of learning it has lasted. We find that continuity easy to undervalue when a new design is in front of us. The customer returning to prepare a report brings a memory of the old product into every encounter with the new one.