Dashboard Audit Checklist

01

Purpose & Alignment

A dashboard without a clear decision it supports is decoration. These questions test whether this dashboard earns its place.

Question 1 of 30

Does this dashboard directly support a named, recurring decision?

If nobody can name the decision this dashboard supports, it likely does not belong in the reporting estate.

Define the primary decision this dashboard supports before spending more time on it. Read more →

Question 2 of 30

Can you identify which meeting or workflow this dashboard is used in?

A dashboard not attached to a specific cadence is rarely used when it matters.

Map this dashboard to a specific meeting or decision cycle — or consider retiring it. Read more →

Question 3 of 30

Is the intended audience clearly defined and limited to people who act on this data?

Dashboards built for everybody are optimised for nobody.

Name the primary audience and ask if everyone with access actually uses it. Read more →

Question 4 of 30

Do recipients use this dashboard to make decisions — not just to check in?

There is a meaningful difference between a dashboard people consult before decisions and one they glance at and close.

Ask the last three people who opened this dashboard what decision they were making. Read more →

Question 5 of 30

Are the metrics connected to documented business goals or OKRs?

Metrics that float free from business goals tend to measure activity, not impact.

Trace each metric to a business objective. Remove the ones that do not connect. Read more →

02

Metric Definitions

Poorly defined metrics undermine every decision this dashboard supports. These questions check whether the numbers mean what people think they mean.

Question 6 of 30

Are all metrics defined in writing — business definition, formula, and source?

If the definition lives only in someone's head, the dashboard will produce disagreements, not decisions.

Start a KPI dictionary entry for every metric on this dashboard. Read more →

Question 7 of 30

Does every metric have a single named business owner?

A metric owned by everybody is owned by nobody.

Assign a named owner to each metric. One person, not a team. Read more →

Question 8 of 30

Is the authoritative source system documented for each metric?

Without a documented source, reconciliation disputes are inevitable.

Document the source system for each metric and treat it as the single authoritative source. Read more →

Question 9 of 30

Are the filters, exclusions, and grain documented for each metric?

Most metric disagreements are about what is included and excluded, not the formula.

Add explicit filter and exclusion documentation to each KPI definition. Read more →

Question 10 of 30

Would Finance, Operations, and Sales agree on the key metric definitions?

Cross-functional disagreement on definitions is the leading cause of untrustworthy reporting.

Run a definition alignment session. Surface the conflicts now. Read more →

03

Data Trust & Quality

Technically correct data is not the same as trusted data. These questions probe whether people actually rely on what they see.

Question 11 of 30

Do recipients trust these numbers enough to use them without verifying elsewhere?

If people run parallel spreadsheets to check the dashboard, it has a trust problem regardless of accuracy.

Run a trust survey with the primary audience. Low trust is a separate problem from low accuracy. Read more →

Question 12 of 30

Is there a clear process for who investigates when a metric looks wrong?

Without a defined escalation path, data errors fester. The same issue surfaces in every review.

Define a data error escalation process. Name the technical owner for each metric. Read more →

Question 13 of 30

Is the data refresh time documented and visible to the audience?

Stakeholders making decisions with stale data do not know what they do not know.

Add a visible last-updated timestamp and document the refresh cadence. Read more →

Question 14 of 30

Are there known data quality issues affecting any metric?

Undocumented known issues are trust destroyers. Transparency builds more trust than silence.

Document known issues visibly on or near the affected metric. Read more →

Question 15 of 30

Has the data powering this dashboard been validated in the past 12 months?

Data pipelines drift. A validation check once a year is minimum hygiene.

Schedule an annual data validation. Log and share the results. Read more →

04

Design & Usability

A well-governed dashboard that is confusing to read will not change behaviour. These questions assess whether the design serves the decision.

Question 16 of 30

Can a new team member understand this dashboard without a walkthrough?

If the dashboard requires institutional knowledge to interpret, it depends on people, not systems.

Add a one-paragraph description of what the dashboard is for and how to read it. Read more →

Question 17 of 30

Does the dashboard have a clear visual hierarchy — the most important metric is most prominent?

When everything is equally prominent, nothing is.

Identify the one metric that matters most. Make it visually dominant. Read more →

Question 18 of 30

Is the dashboard free of decorative elements that do not drive decisions?

Chartjunk and redundant visualisations dilute attention.

Remove any chart or metric the audience has not used in a decision in the past quarter. Read more →

Question 19 of 30

Are chart types appropriate for the data being shown?

Mismatched chart types obscure insight. Pie charts for comparisons, bar charts for rankings.

Review each visualisation against the question it answers. Replace mismatched chart types. Read more →

Question 20 of 30

Does the dashboard lead the reader toward a clear conclusion or next step?

A dashboard that requires the audience to figure out what it means has done half the job.

Add thresholds or conditional formatting that shows whether each metric is healthy or needs attention. Read more →

05

Governance & Ownership

Governance is what separates a dashboard that ages well from one that silently drifts into unreliability.

Question 21 of 30

Is there a single named owner responsible for this dashboard's accuracy?

Shared ownership is no ownership. When a number is wrong, one person should be the obvious call.

Name a single dashboard owner. Document the name on or near the dashboard. Read more →

Question 22 of 30

Is there a defined process for requesting changes or reporting data issues?

Without an intake process, change requests come through ad hoc channels and do not get prioritised.

Create a simple intake form or documented process for change requests. Read more →

Question 23 of 30

Has this dashboard been reviewed and updated within the past six months?

A dashboard not reviewed in six months is likely drifting from its original purpose.

Schedule a semi-annual review. Verify the metrics still reflect the decisions being made. Read more →

Question 24 of 30

Is this dashboard listed in a reporting inventory with owner and purpose documented?

If it is not in a reporting inventory, it is invisible to governance.

Add this dashboard to a reporting inventory. Read more →

Question 25 of 30

Is there a change log for significant changes made to this dashboard?

Without a change log, unexplained metric shifts create confusion.

Start a simple change log: date, description, and the person who made the change. Read more →

06

Operational Health

A well-designed, trusted dashboard that breaks or requires manual effort creates operational debt.

Question 26 of 30

Is this dashboard fully automated — no manual steps to produce or refresh it?

Manual reporting is brittle. Every manual step is a failure point and a recurring time cost.

Identify which manual steps exist and prioritise eliminating the highest-effort ones first. Read more →

Question 27 of 30

Does the dashboard load and refresh within a reasonable time for its intended use?

A dashboard that takes five minutes to load will not be used in a meeting.

Measure load time. If it exceeds 30 seconds, investigate query optimisation. Read more →

Question 28 of 30

Do the right people have access — and only those people?

Overly broad access creates security risk. Overly narrow access creates workarounds.

Review the access list. Remove access for people who have not used it in six months. Read more →

Question 29 of 30

Does this dashboard have a clear owner for the underlying data pipeline?

Dashboard ownership and pipeline ownership are separate. Both need a named person.

Identify and document the technical owner of the data pipeline. Read more →

Question 30 of 30

Does this dashboard have fewer than 20 metrics or visualisations?

Dashboards with more than 20 elements are almost never used as intended.