We finished rebuilding our dashboard. The full before and after, and what it did to engagement
For months we let our marketing site look great while our actual product, the dashboard, quietly looked like enterprise software from 2009. Last month we rebuilt it. It is now done and live for everyone, and the engagement data is in.
What was wrong with the old one
- 14 metric cards, all the same size. When everything is equally loud, nothing stands out
- color on everything, so no number drew the eye
- labels written by engineers for engineers (returning unique sessions)
- a blank second on load while it fetched
- mobile was just desktop cards stacked into an endless scroll
The principles we rebuilt around
- one hero number at the top (visitors and revenue), everything else smaller
- one accent color, the rest greyscale
- plain language a non technical user can read in five seconds
- progressive disclosure: the overview answers "is it up or down", you click in for "why"
- a real mobile view with the three numbers that matter, not all 14
What it did to engagement
I am not going to throw exact percentages at you, partly because a clean 30 day window is still settling and I would rather not quote a number I have to walk back later. But the direction is clear across the three things we actually instrument.
Sessions per active user went up. People open the dashboard more often now, which makes sense once it stopped being a wall of equally loud cards and started answering "up or down" in the first second.
Time to first insight came down, and this is the one I care about most. We instrument the moment a user opens a detail view, the first time they go from "looking at the overview" to "actually digging into something". That gap got noticeably shorter. Part of that is the redesign, but part of it is a streak feature we added: the dashboard now rewards users for coming back, and a small daily nudge turns out to pull people into the detail views faster than the old layout ever did. The streak gave them a reason to come back, and once they are back they get to the insight quicker.
Returning users came back more often too. Same mechanism. A screen people can read, plus a reason to return, beats a prettier screen with no hook.
The hardest part
The hardest part was not design. It was deciding what to cut. Engineers want to show every metric because we built them all. The rebuild was mostly deletion, not addition.
The thing I underestimated: a better dashboard is a retention feature, not just a pretty one. People come back to a screen they can read. They avoid a screen that makes them feel dumb.
Two co founders, bootstrapped. The tool is Zenovay (web analytics, leans on revenue not pageviews) if you want to see it.
For people who have redesigned a data heavy product: how did you decide what to hide? That was 80% of the work for us.
