BI dashboards that survive the Monday meeting

Three out of every four BI dashboards we’re asked to “fix” were built by competent teams. The data is correct. The visuals are tasteful. The pages are well-organised. And nobody looks at them anymore — sometimes within two weeks of the launch demo.

There is no mystery to this. Dashboards die because they were built metric-first instead of decision-first. Someone — usually a finance or ops lead — wrote down 30 metrics they’d like to see, and a BI team rendered them. The result is a beautiful screen that nobody can act on.

The fix is small but ruthless. Before any chart goes on a dashboard, we run it through four questions:

Question 1 — What decision does this chart change? If the answer is “none — we’d just like to know”, the chart should not be on an executive dashboard. It can live in a data exploration notebook. The wall is for decisions.

Question 2 — Who owns the decision? If a chart can change a decision but the owner of that decision is not in the meeting where the dashboard is reviewed, the chart is wasting space. Move it.

Question 4 — Is the underlying data fresh enough for the decision frequency? A chart that updates daily is wasted if the decision it informs is taken weekly — or, worse, if the data feeding it is stale by 3 days. Match the freshness to the decision.

Apply those four questions to a typical 30-metric exec dashboard and you usually end up with 7 or 8 charts. That is the right number. That is what gets used.

There is a corollary to this principle that tends to surprise clients: the best dashboards have an *actions* column visible per chart. Not just the number, but the suggested next step at the current value. A churn chart shipping at 4.7% (the threshold) literally has the line next to it: “At threshold — review the Tier-2 cohort’s last 14 days of NPS.” That sounds bossy. It is also exactly what a tired CFO at 7am wants to see.

And the final principle, which only matters once you’ve done the others: every metric on every chart needs a written definition, source, and freshness indicator visible to the viewer. Either inline or one click away. The moment a metric definition gets ambiguous — different people interpreting it differently — the dashboard loses authority. Authority lost is hard to win back.

If you run an organisation and your exec dashboard is on its second or third reincarnation, try the four questions before you commission a fourth. Most of the time the rebuild is unnecessary — the redesign is.