Data Dashboard Design
When to Use This Skill
Use this skill when you need to:
- Design a business metrics dashboard layout
- Select the right visualizations for different metric types
- Plan a dashboard hierarchy (executive, operational, team-level)
- Specify requirements for a developer or BI tool configuration
DO NOT use this skill for building dashboards in code, creating actual charts, or configuring specific BI tools. This is for dashboard planning and design.
Core Principle
A DASHBOARD THAT REQUIRES EXPLANATION HAS FAILED — EVERY METRIC SHOULD ANSWER A QUESTION THE VIEWER ALREADY HAS IN 5 SECONDS OR LESS.
Phase 1: Brief
Required Inputs
| Input |
What to Ask |
Default |
| Audience |
"Who will view this dashboard? (founder, team leads, investors, whole team)" |
Founder / CEO |
| Purpose |
"What decisions should this dashboard inform?" |
Weekly business health check |
| Key questions |
"What 3-5 questions should someone answer by looking at this?" |
Must be provided |
| Data sources |
"Where does your data live? (Stripe, GA4, CRM, spreadsheets)" |
Mixed sources |
| Tool |
"What will you build this in? (Google Sheets, Looker, Tableau, Databox, Notion)" |
Google Sheets or Looker Studio |
| Refresh frequency |
"How often should data update? (real-time, daily, weekly)" |
Daily |
GATE: Confirm brief before proceeding.
Phase 2: Design
Dashboard Architecture
- Top-line KPIs — 3-5 big numbers at the top answering "How are we doing overall?"
- Trend charts — time-series showing direction over the last 30/90 days
- Breakdowns — dimensions that explain the top-line numbers (by channel, product, segment)
- Comparison context — vs. last period, vs. target, vs. benchmark
- Action triggers — thresholds that signal when something needs attention
Visualization Selection Rules
| Data Type |
Best Visualization |
Avoid |
| Single KPI |
Big number with trend arrow |
Pie chart |
| Trend over time |
Line chart |
Bar chart for 30+ data points |
| Part of whole |
Stacked bar or donut |
3D pie charts (always) |
| Comparison across categories |
Horizontal bar chart |
Vertical bars with 10+ categories |
| Correlation |
Scatter plot |
Line chart |
| Geographic |
Map / heatmap |
Table with location names |
GATE: Present the dashboard wireframe and wait for approval.
Phase 3: Build
Deliverables
1. Dashboard Specification Document
- Complete layout with sections, metrics, and visualization types
- Data source mapping for each metric
- Filter and drill-down requirements
- Refresh schedule and data latency notes
2. Metric Definitions
- How each metric is calculated (formula)
- Data source and table/field references
- Expected range and alert thresholds
3. Wireframe Layout
- Text-based or sketch layout showing placement of each element
- Reading order: top-left to bottom-right, most important first
- Mobile considerations if viewed on phones
4. Implementation Checklist
Phase 4: Polish
Review Cadence
- Week 1: Verify all numbers match source data
- Week 4: Survey users — is anything missing or confusing?
- Month 3: Remove metrics nobody looks at, add any new questions
Dashboard Hygiene Rules
- No more than 10 metrics on a single view — if you need more, create sub-dashboards
- Every metric has a definition tooltip or footnote
- Color coding is consistent (green = good, red = attention needed)
Example 1: SaaS Founder Dashboard
Top-line: MRR, Active Users, Churn Rate, NPS
Trends: MRR growth (12 months), signup trend (30 days)
Breakdowns: Revenue by plan tier, signups by acquisition channel
Example 2: E-commerce Weekly Dashboard
Top-line: Revenue, Orders, AOV, Conversion Rate
Trends: Daily revenue (30 days), traffic (30 days)
Breakdowns: Revenue by product category, traffic by source, conversion by device
Anti-Patterns
- Dashboard overload — 30 metrics on one screen means nobody reads any of them. Ruthlessly prioritize.
- Vanity metrics — total pageviews without context is meaningless. Always show metrics that connect to revenue or retention.
- No comparison context — a number without a benchmark is just a number. Show vs. last period, vs. target, or vs. industry.
- Rainbow color coding — 8 colors for 8 categories creates visual noise. Use 2-3 colors max with gray for secondary elements.
- Stale data — a dashboard that updates monthly when decisions happen weekly is useless.
Recovery
- User cannot define key questions: Ask "What do you check first thing Monday morning? What number keeps you up at night?" Extract dashboard requirements from the answers.
- Too many data sources: Start with 1-2 sources for the MVP dashboard. Add sources incrementally.
- No BI tool: Google Sheets with charts and a refresh button is a valid V1 dashboard for small businesses.
- Data quality issues: Flag metrics with known quality problems. A metric with an asterisk is better than a missing metric.
1---2name: data-dashboard-design3description: Plans data dashboard layouts with metric selection, visualization types, refresh frequency, and user-focused design. Use when building dashboards that drive decisions.4---56# Data Dashboard Design78## When to Use This Skill910Use this skill when you need to:11- Design a business metrics dashboard layout12- Select the right visualizations for different metric types13- Plan a dashboard hierarchy (executive, operational, team-level)14- Specify requirements for a developer or BI tool configuration1516**DO NOT** use this skill for building dashboards in code, creating actual charts, or configuring specific BI tools. This is for dashboard planning and design.1718---1920## Core Principle2122A DASHBOARD THAT REQUIRES EXPLANATION HAS FAILED — EVERY METRIC SHOULD ANSWER A QUESTION THE VIEWER ALREADY HAS IN 5 SECONDS OR LESS.2324---2526## Phase 1: Brief2728### Required Inputs2930| Input | What to Ask | Default |31|-------|------------|---------|32| **Audience** | "Who will view this dashboard? (founder, team leads, investors, whole team)" | Founder / CEO |33| **Purpose** | "What decisions should this dashboard inform?" | Weekly business health check |34| **Key questions** | "What 3-5 questions should someone answer by looking at this?" | Must be provided |35| **Data sources** | "Where does your data live? (Stripe, GA4, CRM, spreadsheets)" | Mixed sources |36| **Tool** | "What will you build this in? (Google Sheets, Looker, Tableau, Databox, Notion)" | Google Sheets or Looker Studio |37| **Refresh frequency** | "How often should data update? (real-time, daily, weekly)" | Daily |3839**GATE: Confirm brief before proceeding.**4041---4243## Phase 2: Design4445### Dashboard Architecture46471. **Top-line KPIs** — 3-5 big numbers at the top answering "How are we doing overall?"482. **Trend charts** — time-series showing direction over the last 30/90 days493. **Breakdowns** — dimensions that explain the top-line numbers (by channel, product, segment)504. **Comparison context** — vs. last period, vs. target, vs. benchmark515. **Action triggers** — thresholds that signal when something needs attention5253### Visualization Selection Rules5455| Data Type | Best Visualization | Avoid |56|-----------|-------------------|-------|57| Single KPI | Big number with trend arrow | Pie chart |58| Trend over time | Line chart | Bar chart for 30+ data points |59| Part of whole | Stacked bar or donut | 3D pie charts (always) |60| Comparison across categories | Horizontal bar chart | Vertical bars with 10+ categories |61| Correlation | Scatter plot | Line chart |62| Geographic | Map / heatmap | Table with location names |6364**GATE: Present the dashboard wireframe and wait for approval.**6566---6768## Phase 3: Build6970### Deliverables7172**1. Dashboard Specification Document**73- Complete layout with sections, metrics, and visualization types74- Data source mapping for each metric75- Filter and drill-down requirements76- Refresh schedule and data latency notes7778**2. Metric Definitions**79- How each metric is calculated (formula)80- Data source and table/field references81- Expected range and alert thresholds8283**3. Wireframe Layout**84- Text-based or sketch layout showing placement of each element85- Reading order: top-left to bottom-right, most important first86- Mobile considerations if viewed on phones8788**4. Implementation Checklist**89- [ ] Data sources connected and verified90- [ ] Metric calculations validated against source91- [ ] Date filters working correctly92- [ ] Comparison periods accurate93- [ ] Loading time under 5 seconds94- [ ] Mobile layout reviewed9596---9798## Phase 4: Polish99100### Review Cadence101102- Week 1: Verify all numbers match source data103- Week 4: Survey users — is anything missing or confusing?104- Month 3: Remove metrics nobody looks at, add any new questions105106### Dashboard Hygiene Rules107108- No more than 10 metrics on a single view — if you need more, create sub-dashboards109- Every metric has a definition tooltip or footnote110- Color coding is consistent (green = good, red = attention needed)111112---113114## Example 1: SaaS Founder Dashboard115116**Top-line:** MRR, Active Users, Churn Rate, NPS117**Trends:** MRR growth (12 months), signup trend (30 days)118**Breakdowns:** Revenue by plan tier, signups by acquisition channel119120## Example 2: E-commerce Weekly Dashboard121122**Top-line:** Revenue, Orders, AOV, Conversion Rate123**Trends:** Daily revenue (30 days), traffic (30 days)124**Breakdowns:** Revenue by product category, traffic by source, conversion by device125126---127128## Anti-Patterns129130- **Dashboard overload** — 30 metrics on one screen means nobody reads any of them. Ruthlessly prioritize.131- **Vanity metrics** — total pageviews without context is meaningless. Always show metrics that connect to revenue or retention.132- **No comparison context** — a number without a benchmark is just a number. Show vs. last period, vs. target, or vs. industry.133- **Rainbow color coding** — 8 colors for 8 categories creates visual noise. Use 2-3 colors max with gray for secondary elements.134- **Stale data** — a dashboard that updates monthly when decisions happen weekly is useless.135136---137138## Recovery139140- **User cannot define key questions:** Ask "What do you check first thing Monday morning? What number keeps you up at night?" Extract dashboard requirements from the answers.141- **Too many data sources:** Start with 1-2 sources for the MVP dashboard. Add sources incrementally.142- **No BI tool:** Google Sheets with charts and a refresh button is a valid V1 dashboard for small businesses.143- **Data quality issues:** Flag metrics with known quality problems. A metric with an asterisk is better than a missing metric.
reaperinvest