Dashboards: what people actually look at

Dashboards get admired once and then ignored. What the ones people return to have in common, and how to design for a decision instead of a display.

3 minread 607words last updated

The short answer

Dashboards get admired at launch and ignored by the second month. The ones people return to have one thing in common: looking at them changes a decision someone makes regularly. People do not open a dashboard to see data; they open it to answer three questions. What needs my attention now? How are we doing against what we expected? What changed since I last looked? A dashboard that answers those for a named person on a named schedule, on one screen, with few numbers, each with a comparison and a reason to care, gets used. One that shows everything the data could show gets a screenshot in a presentation and then nothing.

What people actually look at

QuestionWhat answers itWhat does not
What needs attention now?Exceptions, alerts, items past a threshold, with a link to actA grid of everything, all green
How are we doing against expectation?A number with its target or previous period and the differenceA number alone
What changed since last time?Deltas, new items, movements since the last visitStatic totals
Why?One click from a number to the records behind itA chart with no drill-down

Designing one that gets used

  1. Name the person and the rhythm: the owner on Monday morning, the team lead at the daily stand-up, the finance lead at month end.
  2. List the decisions they make at that moment and the question each decision needs answered.
  3. Pick one metric per question, with its comparison and its threshold for attention.
  4. Order by urgency: attention items first, performance against expectation second, changes third.
  5. One screen, readable on the device they actually use, often a phone.
  6. Drill-down from every number to the records behind it.
  7. Measure use after three months: if they stopped opening it, find out which question it stopped answering.

Numbers, charts and colour

Numbers with comparisons for values. Charts where shape matters: trends over time, distributions, comparisons across many items. Colour only for meaning: attention, good, bad, and used consistently. No decorative gauges, no three-dimensional anything, no colour for its own sake. The test of every element is whether removing it would make a decision harder. If not, remove it.

What this means for you

Build dashboards for decisions, not displays: a named person, a rhythm, the three questions, a handful of numbers with comparisons, one screen, drill-down, and a check after three months. The dashboards that survive are the ones somebody would miss on Monday morning, and that is the only standard worth designing to.

Written by the CivSec S.M.A.R.T team

We build and run websites, software and AI systems for businesses. We write about what we see in that work, in plain language, and we update articles when things change.

Last checked . Spotted something outdated? Tell us.

Frequently asked questions

Why do our dashboards stop being used?

Because they were built to show what the data could show rather than to answer what someone needed to decide. After the novelty, a screen of numbers with no comparison and no consequence is skipped. Rebuild from the questions a specific person asks on a specific day, and remove everything that does not answer one.

How many metrics should a dashboard have?

As few as answer the decisions it serves, usually a handful, each with a comparison against a target or a previous period and a way to drill into why. If there are more questions than fit, there are more dashboards, each for a person and a rhythm, not one screen for everyone.

Charts or numbers?

Numbers with comparisons for anything where the value is the point: revenue this month against target, open tickets against last week. Charts where the shape is the point: a trend, a distribution, a comparison across many items. The common fault is too many charts showing values that a number and an arrow would show better.