Use case — BI & Reporting
Figures that everyone can explain.
Finance reports one revenue figure, sales reports another, and the meeting turns into a data investigation. The problem is rarely the visualization tool: it comes from diverging definitions, rules recreated in every report, and missing history.
The situation
Three reports, three answers
Each team built its own calculation logic in its own dashboard. None of them is wrong on its own; it's just that no one can say which one is authoritative, or why the results differ.
The second symptom is more subtle: without historization, a report rerun the following month no longer gives the same result, and it becomes impossible to say which one was correct. The third is lineage — when no one knows which source a column comes from, the debate over the figure becomes a matter of opinion.
What beVault provides
The calculation rule is written once, in the right place, and versioned.
Sources are loaded and historized in the Data Vault with no business transformation. Definitions — revenue, active customer, margin — are implemented in the Information Marts, where they are documented and tracked over time. Legitimate variants become explicit views, not unexplained discrepancies.
Historization makes yesterday's figure reproducible: you can restore the state of the data as of the relevant date, rather than recalculating from today's state. Lineage lets you trace a published metric back to the source fields that fed it.
Declared quality rules apply to every load: a partial load no longer produces a plausible but incorrect figure without anyone noticing.
See data qualityThe approach
Start with the metric that causes trouble
- 01
Choose the disputed indicator
One only, the one that sparks debate in meetings. Its competing definitions and their respective sources are documented side by side.
- 02
Integrate the relevant sources
Source systems are loaded as-is, historized and subjected to controls. No business rule is applied at this stage.
- 03
Publish a Governed Mart
The chosen definition is implemented once, documented and exposed to your existing BI tools in supported configurations.
- 04
Make reliability visible
Control results and the date of the last load accompany the figure: users know what they're relying on before commenting on it.
What changes
The executive committee discusses the decision, not the origin of the figure.
Manual consolidations decrease because the logic is no longer redone in every report. And when a discrepancy appears, it can be traced back to its source instead of remaining an open question.
Next step