A dashboard is easy to build and hard to get right. In more than a decade of building data products, I have seen the same small problems again and again: a bar chart that does not start at zero, a title that names the chart instead of saying what it found, a target that was in the brief but is not on the page. Nobody meant to make them, and nobody checked for them.
This checklist has 27 yes-or-no checks you can run on any dashboard, built in any tool: Power BI, Tableau, Looker, or a prompt to Claude. Tick each box as you go. Nothing you tick is saved or sent anywhere.
The first 24 checks cover design quality, in the same six categories that Meridian, the dashboard review tool I build, uses to score design. The last 3 cover business alignment: whether the dashboard answers the question it was built for. You need both. A beautiful dashboard that answers the wrong question still fails.
How do you review a dashboard?
Set aside ten minutes, open the dashboard on the screen people will read it on, and work through these five steps.
- Write down what the dashboard is for. In one sentence, write the business question the dashboard answers, who reads it and what decision they make with it. You need this sentence for the business alignment checks.
- Look at it for five seconds. Look at the dashboard for five seconds, then look away and write down what you remember. If you did not write down the most important number, the page needs work.
- Answer every check yes or no. Go through the 27 checks on this page and answer each one yes or no. If you want to answer “mostly”, answer no.
- Fix every no, starting with the ones that make the page wrong. Fix the Integrity & Honesty and Business Alignment checks first, because those problems make a dashboard give the wrong answer. The other checks make it harder to read.
- Ask someone who has never seen it. Give the dashboard to someone who did not build it, and ask what they would do next. If their answer is not the decision the dashboard was built for, go back to the first step.
What should a dashboard design checklist cover?
A dashboard design checklist should cover two things: whether the page is honest and easy to read, and whether it answers the question it was built for. The 24 design checks below sit in six groups. The business alignment checks follow in the next section.
Integrity & Honesty: does the dashboard tell the truth?
Clarity & Visual Design: can a reader see what matters first?
Structure & Navigation: does the page read in a sensible order?
Storytelling & Insight: does each chart say what it found?
Interaction & Usability: are the filters and charts easy to use?
Governance & Accessibility: can everyone read it, and trust where the numbers came from?
Is good design enough?
No. A dashboard can pass every check above and still fail. Good design shows that the page is easy to read. It does not show that the page answers the right question.
Meridian’s public sample certification shows one. A customer retention dashboard scores 84 out of 100 for design, which Meridian rates “Refine”. But its business question asks which customer segments are driving revenue churn, and the dashboard breaks churn down by the number of accounts, not by revenue. The question also asks about regions, and the page has no regional chart at all. Meridian scores it 67 for business alignment and holds the whole certification at “Revise”.
That is why Step 5 of my Decision Outcome Framework checks two things before a dashboard ships: design quality, and business alignment. Business alignment means the dashboard answers the question agreed at Step 1, for the people who asked it. These three checks cover that half.
Business Alignment: does it answer the question it was built for?
What are the most common dashboard design mistakes?
These are the six mistakes I have seen most often. Each one fails at least one check above. Each example shows the same numbers drawn badly and then drawn better. The numbers are made up.
1. A bar chart axis starts above zero, so small differences look large.
Start every bar chart at zero.
2. The page uses a different colour for every category, so nothing stands out.
Use grey for most things, and one strong colour for the thing you want people to see.
3. Chart titles name the chart, such as “Revenue by month”, instead of saying what it found.
Write the title as the finding.
4. Headline numbers have no target or comparison, so nobody can tell whether they are good or bad.
Show every headline number against a target, the last period or last year.
5. Red and green are the only way to tell good from bad.
Add a word or a symbol as well as the colour.
6. Every chart is the same size, so the reader cannot tell which one matters most.
Make the most important number the biggest thing on the page.
How can I get a free second opinion on my dashboard?
It is hard to check your own work, because you already know what the dashboard is meant to say. Someone reading it cold does not.
If you want a second opinion on the design, try Meridian Design Line on your own dashboard. Design Line is a free design review from Meridian. You drop in one screenshot and a sentence about what the dashboard is for, and you get a design review back in under a minute. You do not need to sign up, and Meridian does not keep your image.
Design Line reviews the design only. It does not have your business case, so it cannot check the business alignment half. A full Meridian certification checks both.
Try Design Line on your own screenshot.
It is free, you do not need to sign up, and you get the review in under a minute.
Review my dashboard with Design LineFrequently Asked Questions
What are dashboard design best practices?
Dashboard design best practices are the habits that make a dashboard honest and quick to read. Start bar charts at zero. Use one strong colour for the thing that matters and neutral colours for the rest. Put the most important number at the top left. Title each chart with what it found. Show every headline number against a target or a comparison. Never use colour alone to carry meaning. Above all, make sure the dashboard answers the business question it was built for.
How do you review a dashboard?
Write down in one sentence what the dashboard is for: the business question, who reads it and what decision they make. Look at it for five seconds and note what you remember. Answer each check on a dashboard design checklist yes or no. Fix every no, starting with anything that makes the dashboard give a wrong answer. Then ask someone who has never seen it what they would do next.
How many charts should a dashboard have?
There is no fixed number. Stephen Few defined a dashboard as the most important information needed to achieve one or more objectives, arranged on a single screen so it can be monitored at a glance. If a reader has to scroll to answer the main question, the page probably has too many charts, or it is trying to answer more than one question.
What is the five-second test for a dashboard?
Show the dashboard to someone for five seconds, then hide it and ask what they remember. If they cannot name the most important number, or they name the wrong one, the page does not show what matters first.
Should a bar chart axis always start at zero?
Yes. A bar shows its value through its length, so cutting the axis makes small gaps look large. Line charts are different. They show change over time, so an axis that does not start at zero can be fine, as long as it is clearly labelled.
How do I check a dashboard for accessibility?
Check three things. Under the WCAG accessibility standard, text needs a contrast ratio of at least 4.5 to 1 against its background, or 3 to 1 for large text. The lines, bars and shapes that carry meaning need at least 3 to 1. And nothing should rely on colour alone: if red and green mark good and bad, add a label or an icon as well, because about 1 in 12 men have a colour vision deficiency.
Notes
- Meridian's six design categories
- Meridian, “Dashboard Certification”. The six design groups on this page use the category names Meridian publishes on this page. The checks under each group are my own, written from general dashboard design practice. Source
- Meridian's sample certification
- Meridian, “A full Meridian certification, worked through”. A worked example on a demo dashboard, not a live result. It is the source of the customer retention example and of the three business alignment labels. Source
- Dashboard definition
- Stephen Few, “Dashboard Confusion”, Perceptual Edge, 20 March 2004. Source
- Pre-attentive processing
- Christopher G. Healey and James T. Enns, “Attention and Visual Memory in Visualization and Computer Graphics”, IEEE Transactions on Visualization and Computer Graphics, 2012. Source
- Angles and lengths
- William S. Cleveland and Robert McGill, “Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods”, Journal of the American Statistical Association, 1984. People in their experiments judged position and length more accurately than angle. Source
- Grouping by proximity
- Cole Nussbaumer Knaflic, Storytelling with Data: A Data Visualization Guide for Business Professionals, Wiley, 2015. Source
- Text contrast
- W3C, Understanding WCAG 2.2, Success Criterion 1.4.3: Contrast (Minimum). Source
- Chart contrast
- W3C, Understanding WCAG 2.2, Success Criterion 1.4.11: Non-text Contrast. Source
- Colour alone
- W3C, Understanding WCAG 2.2, Success Criterion 1.4.1: Use of Color. Source
- Colour vision deficiency
- Colour Blind Awareness, “Colour Blindness”. The charity puts colour vision deficiency at about 1 in 12 men and 1 in 200 women. Source
Where these checks sit in the framework.
The Decision Outcome Framework sets out all five steps, from curating the business question at Step 1 to certifying the dashboard at Step 5.
Read the Decision Outcome FrameworkMark Cunningham is the founder of Insights². He has spent more than a decade building data and analytics products. He writes at insights-2.com about how to make the meaningful measurable.