Most dashboard advice starts with “what metrics should I show?”, which I think is backwards. The question I’d start with is which meeting the dashboard will be opened in and what decision needs to get made there.

Most dashboards go unused within months, and there’s a good chance yours has already landed in the graveyard. The data is usually fine and so is the tool; what’s missing is design for the moment of decision.

The same pattern shows up in just about any marketing department. A dashboard was built six months ago with great fanfare, and someone spent weeks on it configuring metrics, choosing colors, and adding widgets. Today nobody opens it, so the team exports to Excel, runs manual reports, and makes decisions on gut feel.

The dashboard shows how things performed but gives no direction. Most get built to report what happened and stop there, so teams keep exporting to spreadsheets when a dashboard doesn’t answer their actual questions.

What meeting is it for?

The reframe that changed how I think about this is to design backwards from the meeting. Every dashboard exists to answer specific questions in specific moments, so before adding a single metric I’d ask;

  • What meeting will this be opened in?
  • What question needs to get answered?
  • What decision will someone make based on this?

A campaign performance dashboard opened in a Monday standup needs to answer “Which campaigns do we pause before 11 AM?” That’s a very different design from a quarterly board deck asking “Is marketing working?”, even though the data underneath is the same. Dashboards that try to serve everyone end up serving no one, and the fix there is a sharper purpose rather than better visualizations.

Which metrics make the cut?

Every metric needs to pass one test: “so what?” If I can’t immediately connect a metric to a decision someone needs to make, it gets cut. Vanity metrics like impressions, follower counts, and page views feel good but drive nothing.

Two that fail the test;

  • “Sessions increased 15% this month.” So what do we do about it?
  • “Email open rate is 22%.” Is that good, and should we change something?

And two that pass;

  • “CAC is $45, up from $38 target, so we need to cut spend on underperforming channels.”
  • “Pipeline from paid is down 20% vs goal, so adjust budget toward content.”

The ones that pass carry context and a consequence, and every metric needs both. Most dashboards also have too many metrics. The limit is 5-7 KPIs per view before cognitive overload kicks in, and a dashboard with 50 widgets is a data dump.

How many layers does it need?

Different people need different depths, which is why I’d build dashboards in three layers;

  1. Executive summary. One view with three to five KPIs max that answers “Is marketing working?” This is what gets pulled into board decks and leadership standups: revenue impact, pipeline contribution, and efficiency metrics, shown as trends over time and always compared to targets (never naked numbers).
  2. Operational metrics. The next level down covers channel performance, campaign results, and conversion rates for the weekly team meeting. It shows what’s working, what isn’t, and where to shift budget, and it should be actionable enough to decide in the meeting.
  3. Debug layer. This is the granular detail for when something looks weird, like “Why did CAC spike last week?” It exists for investigation instead of regular consumption, and most people never see it, which is fine.

The mistake is building one dashboard that tries to do all three. Executives drown in detail, analysts can’t find what they need, and everyone ends up exporting to Excel.

Where to start?

A quick audit for each existing dashboard;

  • Is the purpose clear? Can you name the specific meeting or decision it serves?
  • Does every metric pass the “so what” test and connect to an action?
  • Is it 5-7 metrics max, or is it a data dump?
  • Is context included (targets, benchmarks, trends) instead of naked numbers?
  • Is it actually opened? When’s the last time someone used it?

If a dashboard hasn’t been opened in 30 days, it’s dead and should be archived or rebuilt with a sharper purpose. The 30-day rule is real, and I’d make quarterly audits standard practice, because a dashboard nobody uses isn’t connecting data to decisions.

For Monday, I’d start with three things;

  1. Pick your most-used dashboard and run it through the audit to see how many metrics actually pass the “so what” test.
  2. Name the specific meeting each dashboard you own serves. If you can’t name one, that’s the problem.
  3. Cut ruthlessly and aim for 5-7 KPIs per view. If a metric doesn’t drive a decision, it doesn’t belong.

Companies with effective dashboards make faster decisions, and ultimately the difference comes down to design: start from the purpose and the decision, then keep it ruthlessly simple.

This piece covers what to measure and why. For how to display it (chart types, visual hierarchy, the mechanics of making data readable), check out our guide to dashboard data visualization.

Curious how you handle this in your org. What’s the state of your dashboard graveyard?