A dashboard that looked great at launch can quietly stop working long before anyone officially decides to replace it. The tell isn't always obvious. It's rarely "the colors look dated." It's usually one of these five things, and every one of them points to the same root cause: the underlying data model, not the visuals on top of it.
1. Nobody opens it anymore
Adoption is the most honest metric a dashboard has. If usage has quietly dropped off and people are back to asking a person for the number instead of checking the report, that's not a training problem. It's a trust problem. Somewhere along the way, the dashboard stopped being the fastest, most reliable way to get an answer.
2. It takes forever to load, or times out
Slow dashboards are almost never a "the tool is slow" problem. They're a data model problem: too much raw, un-aggregated data being pulled live instead of a properly built semantic layer doing the heavy lifting ahead of time. A refresh of the visuals won't fix this. The model underneath needs to be rebuilt.
3. Numbers don't match between reports
This is the one that actually damages trust in the data team. If revenue on one dashboard doesn't match revenue on another, it almost always means there's no single governed source of truth. Each report is calculating its own version of the same metric. The fix isn't a reconciliation spreadsheet; it's one semantic model that every dashboard pulls from.
4. Every new question requires a developer
A well-built dashboard lets a business user answer their own follow-up question with a filter or a drill-down. If every new question means opening a ticket and waiting a week for someone to edit the underlying query, the dashboard was built as a fixed report, not a flexible tool, and that's a design decision made early in the build, not something a cosmetic update can undo.
5. It was built by someone who's gone
Undocumented DAX measures, calculated fields, and table calcs are common technical debt in BI tools, and when the person who wrote them leaves, nobody wants to touch the dashboard for fear of breaking it. If your team is afraid to edit its own dashboard, it has effectively already stopped being useful, even if it's still technically running.
What a rebuild actually means
None of this means throwing out everything and starting from a blank page. It usually means rebuilding the data model: the star schema, the relationships, the governed measures, while keeping the parts of the visual design that already work. That's a fundamentally different (and usually cheaper) project than most teams expect once they assume "rebuild" means starting over.
If two or more of these sound familiar, it's worth a conversation before you spend more budget patching the symptoms.