A calm, opinionated cockpit for category managers — surfacing the next best action across 12K SKUs.

Discovery, framing, structure and validation — the work that decided what the screens had to be.
Observed category managers opening their day — ten tabs, three exports, no single answer to 'what needs me today?'.
Reframed the brief from 'show all the data' to 'shorten the list of what needs attention'.
Split the surface into two registers so scanning and deep-diving never fight for the same pixels.
Explored twelve above-the-fold signal layouts in wireframe before choosing chart types.
Six rounds with real managers, measuring seconds-to-answer rather than opinions.
Live captures of the shipped product and the design files behind it.
12,000 SKUs across nine categories. Managers were drowning in tabs. We built a dashboard that prioritises what's broken first and quietly stays out of the way when things are fine.
The previous tool treated every metric equally — a wall of charts that took thirty minutes to read. Managers needed a surface that triaged itself: what changed today, what's at risk this week, what's safe to ignore.
9 category managers · 4 buyers · 2 supply-chain leads
5 weeks discovery + biweekly co-design
"I open ten tabs before I've had coffee. I want one."
"I don't know which metric matters — I stare at all of them."
How the study actually ran — four frames from field to evidence.
Watched category managers open six tabs before deciding which enquiry to touch first.
Synthesis wall turned scattered observations into three repeatable triage themes.
Sketched a prioritised worklist instead of charts nobody acted on.
Benchmarked how fast a manager could act after login, round after round.
Managers didn't want more data — they wanted a shorter list of what needs attention.
Above-the-fold had to encode 12 signals without feeling like a cockpit.
'Change since yesterday' was the most-scanned value in every test.
Wrote an information-priority matrix with the head of category before any pixels.
Designed in two registers — 'glance' (status, deltas) and 'focus' (the one chart you came for).
Built a custom data-viz kit on top of an existing tokens layer, no third-party charts.
Iterated on density with real managers, not stakeholders — six rounds before launch.
Vijay.
Loading