A useful dashboard helps a named person make a decision. Choosing attractive charts is only part of the work. Definitions, data quality, refresh behaviour and the relationship between activity and customer outcomes determine whether the report can be trusted.
Define the audience and action
Record who uses the dashboard, the question it answers and what action might follow. A weekly capacity review needs different information from an investigation into failed customer journeys. Give each measure a business owner who can explain its meaning.
Write down the calculation, population, dates, timezone, exclusions and filters. Keep authoritative sources identifiable. If a number changes meaning, annotate that change, so readers do not interpret a definition change as operational improvement.
Check quality against the intended use
Assess whether required records and fields are present, values are correct, duplicates are controlled and data is timely enough. Completeness and accuracy are different: a filled telephone-number field can contain the wrong number.
Use checks that suit the consequences of a decision. Compare representative records with the operational source and document known limitations. Assign responsibility for business meaning, data quality and technical maintenance rather than assuming one person covers every issue.
Power Query's profiling tools inspect the first 1,000 rows by default. Select profiling over the entire dataset when that is needed, or state the actual sample coverage. A clean preview does not establish that all rows meet the same conditions.
Explain refresh and availability
Show when source data was updated and when the report obtained it. In Power BI, imported data needs a source refresh to include changes. DirectQuery queries the underlying source rather than importing the same point-in-time copy.
Refreshing report visuals or dashboard tiles is separate from refreshing imported source data. Check the actual storage mode and refresh history before calling a report live. Provide a clear response when a refresh fails or a source becomes unavailable.
Present gaps and changes honestly
For regularly spaced observations, leave a visible gap where a value is missing rather than drawing a line as though it were known. Distinguish a recorded zero from unavailable data. Explain significant changes near the affected chart and in the surrounding text.
Use readable labels, accessible contrast and a meaningful alternative where a chart cannot convey its information to a reader. Avoid relying on colour alone to distinguish outcomes. Make filters and units visible in exports as well as on the interactive page.
Measure whole customer journeys
Define the genuine start and completion of a task. Registrations, clicks and form interactions may occur before the customer's task is complete. Include assisted and offline routes where they belong in the business question.
Gather satisfaction evidence across the journey, including follow-up where appropriate. A completed transaction does not necessarily mean the customer found the process easy. In Clarity, rage clicks and dead clicks are observations to investigate, not automatic proof of a person's emotion or the cause of a problem.
Separate capacity from cash savings
Time released by a process improvement can support additional capacity, faster service or reduced effort. It does not automatically reduce cash expenditure. Define which outcome was achieved and avoid counting the same benefit under several headings.
Use report usage information with care. Embedded experiences or blocked collection can affect available usage records. A missing view in one report is insufficient evidence that a dashboard is unused; check the relevant context and users.
Dashboard review checklist
- Identify the audience, action and measure owners.
- Document calculations, coverage and exclusions.
- Check correctness as well as completeness.
- Show refresh status and missing observations.
- Connect activity to complete customer outcomes.
- Keep accessible explanations and limitations available.
The attribution guide explains credit and reporting windows. Our analytics guide covers validation before reporting.
