01

Start with the decision, not the dashboard

A dashboard is useful only when its audience, decision, metric ownership, and expected refresh behavior are explicit. Begin by recording who uses the report, what action it supports, and which source is authoritative.

  • Named business owner
  • Defined audience and decisions
  • Documented refresh expectation
  • Known source and transformation path
02

Validate metric and filter logic

Compare displayed definitions with the underlying calculations, filters, hierarchy scope, date logic, and exclusion rules. Small differences in population logic are a common cause of lost trust.

  • Metric numerator and denominator
  • Date and timezone behavior
  • Hierarchy and role scope
  • Exclusions, suppression, and missing data
  • Comparison period logic
03

Test visibility and actionability

Review the report through representative roles and confirm that restricted users see the correct population. Check whether alerts, drill paths, exports, and follow-up actions behave consistently.

  • Representative role tests
  • Drill and export validation
  • Alert ownership
  • Empty and low-volume states
  • Leadership summary consistency
04

Create an evidence-based backlog

Record each finding with impact, evidence, owner, and a recommended action. Separate urgent data or access defects from usability improvements so remediation can be sequenced realistically.