How to build executive dashboards that actually get used

Most executive dashboards fail quietly. Someone spends two weeks wiring up thirty metrics, the CEO looks at it in the Monday meeting, asks "so are we on track?", and nobody can answer from the screen. By the third week the dashboard is a bookmark nobody clicks.

The cause is almost always the same: the dashboard was built for completeness, and executives use dashboards for decisions. This guide covers how to build the second kind.

Start from decisions, not from data

Before you open any tool, ask the executive four questions and write down the answers:

1. What decisions do you make every week, month and quarter?
2. Which numbers would change those decisions?
3. At what level would each number make you act? ("If gross margin drops below 38%, I want to know.")
4. What do you currently ask someone to pull for you by hand?

The answers give you the metric list, the thresholds and the comparison each number needs. Anything that does not connect to a decision belongs on an operational dashboard, not this one.

Keep it to a handful of numbers

The "rule of seven" comes from George Miller's 1956 paper on short-term memory ("seven, plus or minus two"). Later research, such as Nelson Cowan's work in 2001, puts the real limit closer to four chunks. The practical reading is the same either way: 4 to 7 headline numbers, with trends and detail below them.

If you have more candidates than that, rank them by how often they change a decision. The rest can live one click away.

What each executive usually looks at

These are starting points, not templates. Replace anything the business does not act on.

CEO: revenue against plan, growth rate, cash runway, one customer health number (churn or retention), one pipeline number.

CFO: cash and runway, gross margin, operating expenses against budget, accounts receivable days. Runway is simple arithmetic: cash divided by monthly net burn. With $600,000 in the bank and a net burn of $50,000 a month, runway is 12 months.

COO: throughput (orders shipped, tickets closed), cycle time, on-time rate, cost per unit or per order.

Head of marketing or sales: qualified pipeline, customer acquisition cost, conversion rate by stage, spend against budget. Our post on customer acquisition cost covers which CAC formula to show.

Give every number a comparison

"Revenue: $184,000" means nothing on its own. Every headline number needs at least one of:

• Against target: $184,000 against a plan of $200,000 is 92% of plan.
• Against the previous period: up from $170,000 last month is +8.2%.
• Against the same period last year: the fairest comparison for seasonal businesses.

Pick one primary comparison per metric and use it consistently. Mixing month-over-month on one card and year-over-year on the next is how meetings turn into arguments about definitions.

Status colors people can trust

Red, amber and green work only if the thresholds are written down and agreed. Three rules:

• Define thresholds in advance, in the metric definition, not in the chart settings of whoever built it.
• Avoid a sea of green. If everything is always green, the thresholds are too loose and people stop looking.
• Do not rely on color alone. About 1 in 12 men has some form of color vision deficiency. Pair color with an arrow, a sign or a label such as "below plan". Our list of data visualization mistakes covers this and other chart problems.

Layout: headline, trend, detail

Put the headline numbers in the top row, trend charts in the middle and tables at the bottom for the people who want detail. In clariBI the grid is four columns wide, so four small metric cards fill the top row and two medium charts fill the next. On a phone the grid drops to a single column and keeps the same order, so the top row is still what an executive sees first. The dashboard layout guide has the widget sizes.

Annotate what the chart cannot say. A one-line note such as "March dip: warehouse move" saves ten minutes of questions.

Match the refresh to the decision

Executives rarely decide anything hourly. A dashboard that refreshes daily is enough for most leadership use, and weekly suits a review dashboard. In clariBI each dashboard has a refresh frequency (manual, hourly, daily, weekly or monthly), and sources sync every 24 hours by default. Refreshing an AI dashboard hourly mostly spends credits on the same data. The refresh schedule docs explain the trade-off.

Push the numbers, do not just publish them

A dashboard waits to be opened. A weekly email does not. On Professional and Enterprise you can schedule a report from the same data and send it to the leadership team every Monday morning; see scheduling reports. The dashboard answers follow-up questions, the email makes sure the numbers are seen. It pairs well with a weekly business review.

Treat the dashboard as a product

Dashboards decay. Metrics stop mattering, sources change their columns, someone adds "just one more chart". Four habits keep it useful:

• One owner who can say no to additions.
• A quarterly review of each metric: did it change a decision in the last three months? If not, remove it.
• A data check before important meetings: when did each source last sync?
• A short change log in the dashboard description, so people know why a number moved when a definition changed.

Four patterns that kill executive dashboards

• The everything dashboard. Forty widgets, no hierarchy. Nobody knows where to look.
• The pretty one. Gauges and 3D pies that look good in a screenshot and answer nothing.
• The set-and-forget. Built for last year's priorities and never updated.
• The committee dashboard. Every executive gets their favorite metric, so none of them get a clear view. Build one shared headline row and separate pages for each function.

Building one in clariBI

1. Connect the sources the metrics come from. Most sources get an automatic Data Insights dashboard, which is a useful place to see what columns you have.
2. Open AI Dashboards, click Create Dashboard and use the Dashboard Builder: pick the sources, the business area and then edit the suggested metrics down to your 4 to 7. On Lite, use manual mode and bind each widget to columns yourself.
3. Arrange the layout: metric cards on top, trends below.
4. Use Link to Goal on a headline widget so the target lives next to the number. See connecting goals to dashboard data.
5. Set the refresh to daily and share it with the leadership team.

Then sit with the executive for ten minutes and watch them use it. Whatever they ask that the screen cannot answer is your next change.