Two reports, two figures: build dashboards people can reconcile

Sales presents one total, operations another, and the meeting becomes a debate about spreadsheets. Disagreement often starts before visualisation: periods, states, cancellations, duplicate records and sources were defined differently. Another dashboard can create a third number unless those definitions are resolved.

Reconcile one measure before expanding

Write the definition of a single indicator. In a hypothetical monthly order count, decide whether to count creation or confirmation, how cancellations work and which time zone defines the period. There is no universally correct choice; the measure should support a stated decision.

Compare a small sample that responsible people can check. Identify each difference and classify its cause. Changing a total merely to match another spreadsheet removes the explanation needed for trust.

Present limitations alongside the figure

Show refresh timing when the source updates daily. If a location has not supplied its records, disclose coverage rather than present the total as complete. Missing records can mean missing information, not inactivity.

A summary may provide access to underlying records, subject to permissions. That supports reconciliation but must not expose personal information to every reader. Aggregation and access rules belong in the visualisation brief.

Scope the engineering behind the chart

Through dashboards and BI, I can define contracts, connect sources and implement reconciliation checks. The dashboard demonstration uses fictional records and a filter-based total. It is not a production data pipeline or evidence of business outcomes.

Bring the conflicting reports, their definition owners and the decision the indicator supports. Initial examples can be anonymised. A useful implementation may start with source quality and an agreed contract before any new screen is built.