Experimentation
PXL prioritisation scoring
Potential, Exposure and Level of effort, each scored one to five, resolved into one number the backlog sorts on.
PXL is a prioritisation model for experiment ideas. Potential asks how much the change could move the metric. Exposure asks how many people would see it. Level of effort asks what it costs to build. Each is scored from one to five.
The three scores resolve into a single importance from zero to ten, which is what the experiment list sorts on. One number is what makes a backlog orderable; three numbers is a discussion.
Scoring is deliberately coarse. A five-point scale on three dimensions is quick enough that ideas actually get scored, and precise enough to separate the obvious work from the rest, which is all prioritisation needs to do.
The highest-scoring draft is surfaced on the home screen as the thing to pick up when nothing is broken and nothing is waiting on a decision.