DS-201a · Module 2

Dashboards That Drive Decisions

3 min read

3. That is the maximum number of metrics that should appear on a single dashboard view. Not 12. Not 8. Three.

There are two types of dashboards: reporting dashboards and decision dashboards. A reporting dashboard answers "what happened?" with comprehensive data. A decision dashboard answers "what should we do?" with focused, actionable metrics. 94% of dashboards in production are reporting dashboards. The executives looking at them need decision dashboards. That gap is why data teams get the complaint: "I have all this data and still can't make a decision."

DECISION DASHBOARD FRAMEWORK

┌─────────────────────────────────────────────┐
│  PRIMARY METRIC        [value]   [delta]    │
│  ─────────────────────────────────────────   │
│  What it means: [1-sentence interpretation]  │
│  Action threshold: [when to act]             │
├─────────────────────────────────────────────┤
│  SUPPORTING METRIC 1   [value]   [delta]    │
│  Why it moved: [causal explanation]          │
├─────────────────────────────────────────────┤
│  SUPPORTING METRIC 2   [value]   [delta]    │
│  Why it moved: [causal explanation]          │
├─────────────────────────────────────────────┤
│  ACTION REQUIRED: [Yes/No]                  │
│  If yes: [specific recommended action]       │
└─────────────────────────────────────────────┘

Drill-down available for: [list of context metrics]

The decision dashboard has four properties that reporting dashboards lack.

First, action thresholds. Every metric has a defined threshold that triggers action. Pipeline velocity drops below 45 days? Alert the sales manager. Win rate drops below 25%? Trigger deal review. Without thresholds, a dashboard is just a screen saver with numbers.

Second, causal context. When a metric moves, the dashboard explains why — not just that it moved. LEDGER built this into our revenue dashboard: when pipeline coverage drops, it automatically surfaces which segments declined and which rep territories are below target. The "why" is built into the view.

Third, color coding that works. Green-yellow-red based on defined thresholds, not arbitrary ranges. And accessible to colorblind users — use icons or patterns alongside color. 8% of male executives are red-green colorblind. Your stoplight dashboard is invisible to 1 in 12 of them.

Fourth, recommended actions. The best decision dashboards don't just show what happened — they suggest what to do. "Pipeline coverage at 2.1x (below 3x threshold). Recommended: increase HUNTER's outbound volume by 40% in underperforming segments."

Do This

  • Define action thresholds for every metric BEFORE building the dashboard
  • Include causal context — why did the number move, not just that it moved
  • Limit to 3 metrics per view with drill-down for detail

Avoid This

  • Put 15 charts on one screen and call it a "comprehensive view"
  • Use red/green color coding without alternative indicators for accessibility
  • Build dashboards without defined thresholds — that is a report, not a tool