Notes that find themselves

First week

A four-week sprint only works when the question is narrow. We agreed on one: how does a researcher get back to a note they half remember? I interviewed eight beta users on video and asked each of them to find three notes while thinking out loud.

Nobody searched by title. They searched by when, by who was in the meeting, or by a word they remembered from the middle of the note.

That changed the plan for the sprint. Instead of a new folder structure, which the founders had expected, we spent the second week on search and on getting tags into the note without leaving it. The founders agreed in one call, because the recordings made the case better than any slide.

Decisions

Search became one field that understands dates and people, with results grouped by week. Tagging moved into the note itself: type a hash and pick from the list, no separate screen. I tested two prototypes with the same eight people in week three and kept the one that was faster for six of them.

A tag that costs three screens is a tag nobody adds.

The handover file had every state named the way the code names it: empty, loading, no results, offline. The engineers built from it without a single call that week.

Launch

The beta opened on the planned Friday. Time to find a note fell by 35% in the first fortnight, the beta rating rose from 3.8 to 4.6, and tagging went from three screens to one. Notewell came back in the summer for the sharing flow.