Set up a clean experiment loop; the winning changes lifted 30-day retention 65%.
See the full storyProof of work
Requirements arrived vague and rework was constant, because nobody found the missing decision until the build was half done. I turned intent into crisp, testable specs and a review that surfaced the hard questions while they were still cheap to answer.
Set up a clean experiment loop; the winning changes lifted 30-day retention 65%.
We were shipping features nobody used and treating the launch as the finish line. I got close to the data and to actual users, changed what we were betting on, and made the case with evidence rather than the loudest opinion in the room.
Introduced crisp, testable specs and a review ritual; decisions that took weeks now take days.
Growth had stalled and everyone had a theory, none of them testable as stated. I set up the experiment loop that turned theories into hypotheses with a metric attached, then ran them cleanly enough that the results actually settled the argument.
Took a 0-to-1 surface from concept to 10k weekly active users in three quarters.
A cross-team launch had ten dependencies and no single owner, so every team assumed someone else was tracking the critical path. I ran the program end to end — surfaced the real sequence, drove the weekly decisions, and kept the date honest instead of optimistic.
Owned the program across 8 teams and 9 dependencies; it shipped on the committed date.
Bring craft and measurement to work that usually gets neither.
Off the clock
In public
Practical bits
Unlock with a company account to message directly and reveal:
Jump back to the board with this background already filtered in.
