Another dashboard is not the answer
Almost every engagement starts the same way. Someone shows us a dashboard. It is not a bad dashboard. The charts are fine, the filters work, somebody clearly cared. And nobody is using it.
The reflex is to blame adoption — training, buy-in, a champion in the team. It is almost never adoption.
The actual failure
Ask what happens when the dashboard says one thing and the finance export says another. In a healthy system, that is a bug with an owner. In most of the operations we see, it is a known condition. Everyone has learned which number to trust for which conversation, and that knowledge lives in three people's heads.
Once that is true, the dashboard cannot win. It is one more claimant to the truth in a room that already has several, and it has less standing than the spreadsheet the head of operations has maintained for six years — because her spreadsheet has been wrong in public and survived it.
A dashboard presents numbers. It does not settle them. If your problem is that the numbers disagree, building a faster way to display the disagreement is motion, not progress.
What settles a number
Three things, in order, and none of them is a chart.
One definition, written down. Not "revenue" — revenue recognised when, net of what, excluding which entities, in which currency at which rate on which date. Most reconciliation work is two correct calculations of two different things that share a name.
One place it is computed. Not one database — one computation. The moment the same figure is derived in the warehouse and again in a spreadsheet and again in the billing system, you have three figures that will diverge, and the divergence will be discovered by a customer.
A visible lineage. Anyone who doubts the number should be able to get from it to its inputs without asking a person. Not because they often will, but because the possibility is what makes it credible. A number nobody can audit is a number nobody defends in a meeting.
Do those three and the dashboard becomes almost incidental. People will read the figure off a plain table, or a daily email, and act on it. We have shipped exactly that and watched a six-figure BI project get quietly shelved next to it.
Why this is unpopular
Because it is not a deliverable you can screenshot. "We agreed what revenue means and moved the calculation into one place" does not photograph well in a board pack. A dashboard does.
That asymmetry is the whole reason the industry keeps building the wrong thing. Someone has to be willing to sell the unglamorous version.
The test
Before commissioning anything that visualises data, ask one question of the people who would use it:
When this disagrees with the other number, which one do we act on, and who decides?
If the answer comes back immediately and everyone in the room gives the same one, build your dashboard — it will get used. If the room goes quiet, or you get two answers, you do not have a reporting problem. You have a definitions problem, and it will follow you into whatever you build next.