Every organisation of a certain size has a dashboard, and most of them are quietly ignored. Not because the charts are bad, but because the first time a figure on the dashboard disagreed with the figure in the finance system, everyone learned to go and ask finance instead.
A dashboard earns trust on three properties, and none of them are visual.
**It reads from the source, and only reads.** The indicators are computed from the modules that own the data — the ledger owns the money, the register owns the headcount — rather than from a copy that is refreshed by a job somebody has to remember. A reporting layer that can also write is a reporting layer that can diverge.
**Every figure opens.** A number you cannot drill into is a claim, not evidence. If the operating surplus looks wrong, the useful dashboard is the one where you click it and arrive at the journals behind it, in the same session, with the same permissions you already have.
**It is scoped like the organisation is.** A site director should see their site, a group executive the consolidation, and the dashboard should be the same dashboard — not a second product bought for the boardroom. Where several entities roll up, consolidation has to handle the unglamorous parts: intercompany elimination, currency translation, and a period close that locks.
What that looks like in our platform: a metric catalogue that defines each indicator once, widgets built from it, period comparison, alerts when an indicator crosses a threshold, and board packs and scheduled reports so the monthly pack assembles itself instead of being rebuilt by hand every month.
The test to apply to any vendor, including us: ask for a figure on the dashboard, then ask to be taken to the transactions that produce it. If that journey needs an export, the dashboard is decoration.