Most design effort goes into the first five minutes. Onboarding, empty states, welcome tours, the first successful action. It makes sense: that's where users are lost, and it's where a design team's work is most visible. But a product is used for years, and by the second year the first five minutes are a rounding error in the total time spent inside it.
The visit nobody designs
Consider a customer who opens your product on a Tuesday afternoon, eighteen months after signing up. They know where everything is. They have strong habits and slightly resentful workarounds. They aren't exploring; they're finishing a task so they can get back to something else. Every pixel you spend trying to delight them is a pixel in the way.

Four principles we've adopted
- Respect muscle memory. If a button moves, you'd better have a reason that survives contact with someone who clicks it two hundred times a week.
- Make the rare thing findable, not prominent. Advanced features should be a short search away, never in the way.
- Let the interface get quieter. Tips, banners and nudges should retire themselves once they've done their job.
- Design the error you'll see most. In year two, people don't hit exotic failures; they hit the same three, repeatedly.
Onboarding gets someone in the door. Everything after that is a long conversation about whether they made the right choice.
What we changed
Last quarter we ran a small experiment. For one month, every design review began with a single question: what does this look like for someone who has used it 500 times? It was uncomfortable. It killed a handful of well-loved ideas, and it pushed us toward improvements that were far less photogenic and far more valuable — keyboard shortcuts, saved views, better defaults and shorter error messages.
Customer feedback on those changes was quieter than for any launch we've done, and the retention curve for the cohorts that got them is the best we've measured. Nobody wrote a thread about it. That's roughly how you can tell it worked.



