Chart wallpaper
Teams buy BI — business intelligence software such as Power BI, Tableau, or Looker — when Excel feels embarrassing, then build twelve tabs nobody opens because the numbers do not match the Tuesday huddle. The dashboard becomes wallpaper: pretty, ignored, occasionally screenshot for a deck. As the one report the room trusts piece argues, trust comes before chrome.
If nobody opens the dashboard, it is not a UX problem. It is a trust and relevance problem. One metric, one owner, one weekly ritual beats twelve tabs refreshed at different times with different definitions.
Fix order: decision first, chart second
Ask what decision happens weekly — hire temp staff, release a PO, chase a debtor, pause a production line. Ask what input must be true for that decision — stock on hand, cash runway, open jobs over SLA. Ask who owns the refresh and who answers when it looks wrong. Delete charts that do not serve that decision — yes, delete. Shelfware inside a BI tool is still shelfware.
Do not add filters until the base number matches finance or ops' system of record. A drill-down on a wrong total is faster failure, not insight.
Distrust has causes you can name
Duplicate fields in CRM and finance. Manual adjustments made in a spreadsheet after the dashboard pulls. Different date rules — booked versus billed versus collected. Stale connectors nobody monitors — this is where naming owners for every alert becomes relevant. Nobody ever agreed that this was the official view.
Fix the definition in a shared document before you fix the colour palette. Run one side-by-side session with finance and ops until one line matches tolerance. Then build the smallest view that supports the weekly decision — not the wish list from a workshop.
When a BI platform is the wrong buy
If the problem is one number, a scheduled report from the system of record may be sufficient. If the problem is process, no chart fixes everyone busy, nothing moving. A quarterly systems check-in can catch drift before you buy more login fatigue.
Start with one tile, not a portal
Ship one trusted view — open jobs over SLA, cash this week, stock below minimum — emailed or bookmarked at 07:00. Add a second tile only after the first survives a bad week. Portal projects die when tile twelve disagrees with tile three and nobody owns the refresh.
Colour and layout come last. Definition, owner, and ritual come first — the same playbook as one report the room trusts. A BI dashboard without those is expensive wallpaper.
Ask users why they still open Excel — the answer is usually "I trust mine" or "yours is stale." Fix trust and refresh before you buy another connector. When the stack is heavier than the work often includes business intelligence tools nobody uses.
Involve the Tuesday huddle owner in dashboard design. If they will not present from it weekly, it is not the dashboard you need. Delete vanity metrics that look impressive in sales demos but never change a decision on your floor.
Book a review — Teesside and the North East. A process audit when you need the map before another tool. Build when the fix is a thin integration, not a platform.