Designing a Smarter Experiment Dashboard

I feel like experiment dashboards could be improved to better set expectations: what decisions are allowed, what could be expected, and how much longer a sequential experiment might need to get there. So here's me putting my experiment + UI design hat back on for a playful prototype.

The Prototype

The key ingredients I was going for:

1. Spatial and colour-coded action recommendation Each row sits in the column of the decision it is heading for (Wait, Abort, Give Up, Ship), so position and colour alone tell you where it stands.
2. Visualizing change since last visit An effect that strengthened arrives larger, one that weakened arrives smaller, and a decision that flips columns entirely travels across and leaves a coloured trail behind it.
3. Flexibility for different test designs Sequential, fixed-horizon, and hybrid test designs are allowed and codified as icons, alongside one- and two-tailed and non-inferiority goals.
4. Show "how much longer" tests need to run Sequential rows still gathering data carry a running estimate of the days left to a decision.
5. Visualize that some tests cannot be aborted by design A fixed-horizon test only reads once, at completion, so it has no early abort to offer — shown as a greyed-out, dashed placeholder instead of a live column.

This is just a prototype for now, but if you're working on experiment platform-level improvements, like better decision UI or clearer test status, I'd be glad to help design or build it with you.

Want to Build This Together? Let's Chat
A friendly 25-min call about experiment & UI design