Lesson 4 12 min

Building Dashboards That Drive Decisions

Design executive dashboards that follow the What → Why → What to Do narrative — placing the right metrics in the right positions with the right comparisons to turn passive data displays into active decision tools.

Premium Course Content

This lesson is part of a premium course. Upgrade to Pro to unlock all premium courses and content.

Access all premium courses
1000+ AI skill templates included
New content added weekly

← Back to course overview

🔄 Quick Recall: In the previous lesson, you learned to design metrics that matter — pairing leading with lagging indicators, building KPI hierarchies that create clarity from the North Star down to diagnostic metrics, and using the vanity metric test to ensure every number drives decisions. Now you’ll take those metrics and arrange them into dashboards that actually drive action.

The Dashboard Problem

Most dashboards are data graveyards — rows of numbers, charts nobody reads, and reports that exist because someone once asked for them. Research shows that when dashboards try to show everything, they end up communicating nothing.

The fix is simple in concept but hard in practice: design dashboards for decisions, not data.

The Three-Layer Dashboard Narrative

Every effective dashboard tells a story in three layers:

Layer	Question It Answers	What It Shows
What	What’s happening right now?	Top KPIs with status indicators (green/yellow/red) and trend arrows
Why	Why are these numbers moving?	Breakdowns by segment, time trends, comparisons that explain changes
What to Do	What action should I take?	Recommended actions, responsible owners, links to detailed analysis

Layer 1: What is the executive summary. Five to seven numbers that answer “are we on track?” in under 10 seconds. If the viewer has to think about what a number means, the design failed.

Layer 2: Why is the analysis layer. When a Layer 1 metric turns yellow or red, the viewer clicks or scrolls to see why. Breakdowns by customer segment, product line, time period, or geography reveal where the problem lives.

Layer 3: What to Do turns insight into action. This layer might show specific accounts at risk, campaigns to adjust, or processes to review. It connects the data story to the work that needs to happen.

✅ Quick Check: Why do most dashboards fail to drive decisions? Because they stop at Layer 1 — showing what’s happening without explaining why or suggesting what to do. A red indicator gets attention, but without the why-layer and action-layer, it just creates anxiety. Effective dashboards make the next action obvious, not just the current problem visible.

Layout Principles

How you arrange information on a dashboard matters as much as which information you include.

The F-pattern rule: People scan screens in an F-shape — top-left first, across the top, then down the left side. Design for this:

Top-left quadrant: Your most important metric. Large, prominent, with trend comparison. This is the first thing anyone sees.
Top row: 3-5 strategic KPIs with sparkline trends. Answer “how are we doing?” in one glance.
Middle section: Charts and visualizations that explain the story behind the top numbers.
Bottom section: Detailed data tables for those who need to drill deeper.

Comparison is everything. A number without context is meaningless. Every metric needs at least one comparison:

Comparison Type	What It Reveals	Example
vs. Previous Period	Is the trend improving or declining?	Revenue this month vs. last month
vs. Target	Are we on track to hit our goals?	Conversion rate vs. quarterly OKR target
vs. Benchmark	How do we compare to the industry?	NPS vs. industry average
vs. Segment	Where are the differences?	Churn rate by customer tier

Help me design an executive dashboard.

Business: [describe your business or department]
Key decisions the dashboard should support:
1. [decision 1, e.g., "Where to allocate marketing budget"]
2. [decision 2, e.g., "Which customer segments need attention"]
3. [decision 3, e.g., "Whether we're on track for quarterly targets"]

Design a dashboard layout with:
1. Top-row KPIs (5-7 metrics with comparisons)
2. The three-layer narrative: What, Why, What to Do
3. Visual hierarchy — what goes where and why
4. Which comparisons each metric should include
5. What triggers a drill-down investigation

Common Dashboard Mistakes

Mistake	Why It Fails	Fix
Too many metrics	Viewer can’t find what matters	Ruthlessly cut to 5-7 top-level KPIs
No comparisons	Numbers without context mean nothing	Add vs. target, vs. last period, vs. benchmark
Equal-size everything	No visual hierarchy = no attention guidance	Make the most important metrics 3x larger
Pie charts for comparison	Humans are bad at comparing angles	Use bar charts — angles are deceptive, lengths are clear
Updated too infrequently	Stale data erodes trust and relevance	Match update frequency to decision frequency

✅ Quick Check: Why should every metric on a dashboard include at least one comparison? Because a standalone number is meaningless. “Revenue: $1.2M” tells you nothing. “Revenue: $1.2M (up 15% vs. last quarter, 5% above target)” tells a complete story. Comparisons provide the context that transforms a number into an insight.

Key Takeaways

Design dashboards for decisions, not data — ask “what decisions does this dashboard support?” before choosing any metrics, and if a metric doesn’t connect to a decision, it doesn’t belong on the dashboard
Follow the three-layer narrative: What’s happening (top KPIs with status indicators), Why it’s happening (segment breakdowns and trend analysis), and What to Do (recommended actions and responsible owners)
Use visual hierarchy based on the F-pattern: top-left gets the most important metric (largest), top row gets strategic KPIs, middle gets explanatory visualizations, bottom gets detailed drill-down data
Every metric needs at least one comparison (vs. target, previous period, benchmark, or segment) — a number without context is just a number, not an insight
Avoid the five common mistakes: too many metrics, no comparisons, equal-sized everything, pie charts for comparison, and stale data that erodes trust

Up Next: You’ll add AI to your analytics toolkit — learning to use natural language data queries, automated anomaly detection, and predictive analysis to surface insights that manual analysis would miss.

Knowledge Check

1. You're building a dashboard for the VP of Sales. She currently checks 15 different reports. She asks you to 'put everything in one place.' What's the right approach?

Consolidate all 15 reports into one comprehensive dashboard with tabs for each area Ask her: 'What are the 3 decisions you make most often?' then build the dashboard around those decisions. A dashboard designed around 15 reports becomes a data graveyard — all the numbers, none of the insight. The right approach: (1) IDENTIFY DECISIONS: 'Every Monday I decide which deals to prioritize. Every month I decide where to allocate the team. Every quarter I decide whether our pipeline supports the forecast.' (2) DESIGN FOR THOSE DECISIONS: Monday view — deal pipeline sorted by probability × value, with close-date and last-activity flags. Monthly view — team performance by rep with capacity indicators. Quarterly view — pipeline coverage ratio and forecast confidence. (3) EVERYTHING ELSE: Move to a diagnostic layer she accesses only when the primary dashboard flags a problem. The principle: dashboards should be decision-centric, not data-centric. Start with the most important 5 reports and add more based on feedback

2. Your dashboard shows that customer satisfaction (NPS) dropped from 45 to 32 this month. The dashboard displays this number clearly with a red indicator. Is this dashboard design effective?

Yes — the red indicator clearly flags the problem for immediate attention Partially — it shows WHAT happened (NPS dropped) but not WHY or WHAT TO DO. An effective dashboard follows the three-layer narrative: (1) WHAT: NPS dropped from 45 to 32. Red indicator. ✓ Done. (2) WHY: Break down NPS by customer segment, product line, or support channel. Which group drove the decline? Show the trend over the last 6 months — is this a sudden drop or a gradual decline? Add a drill-down to recent detractor comments. (3) WHAT TO DO: If the decline is concentrated in one segment, the action is targeted. If it's broad, the action is different. Link to the responsible team's action items. The red indicator gets attention. The why-layer prevents panic. The what-to-do layer prevents the meeting where everyone stares at the red number and asks 'so what do we do?' No — NPS is a lagging indicator and shouldn't be on the primary dashboard

3. You're designing a dashboard and have 20 metrics to display. Your instinct is to arrange them in a grid — 5 rows of 4 metrics each. A colleague suggests a different layout. What's the better approach?

The grid layout is clean and organized — it works for most dashboards Use visual hierarchy instead of a grid — the 3-5 most important metrics should be large and prominently placed at the top-left, with supporting metrics smaller and below. Research on how people scan screens shows an F-pattern: they look at the top-left first, scan across, then scan down the left side. Dashboard design implications: (1) TOP-LEFT: Your North Star or most critical KPI — large, with trend and comparison. This is the first thing anyone sees. (2) TOP ROW: 3-5 strategic KPIs with sparkline trends. These answer 'how are we doing overall?' (3) MIDDLE SECTION: Charts and visualizations that explain the WHY behind the top metrics. (4) BOTTOM: Detailed tables and diagnostic data for drill-down. The grid layout treats all 20 metrics as equally important — which means none of them feel important. Visual hierarchy tells the viewer where to look first, what matters most, and how to go deeper when needed. Use tabs or pages — put the top 5 metrics on page 1 and the rest on subsequent pages

Answer all questions to check

Complete the quiz above first

Related Skills

Data Interpreter