Seven rules for dashboard design, with examples
Seven rules for dashboard design are: start from the decisions users make, show only the metrics that matter, give every number context, choose the right chart for each question, guide the eye with hierarchy, design for action and test with real users. Applied together, they turn crowded screens into dashboards people actually use.
Why rules help
Dashboard design involves many decisions about content, layout, charts and colour. Simple rules help teams make those decisions consistently and explain them to stakeholders. This post presents seven rules, each with a realistic example showing how it changes a dashboard. For the broader process and visual details, see our guides to effective dashboard design and dashboard UI tips.
The seven rules at a glance
| Rule | What it prevents |
|---|---|
| 1. Start from decisions | Dashboards full of data nobody uses |
| 2. Show only what matters | Overload and slow scanning |
| 3. Give every number context | Numbers that cannot be interpreted |
| 4. Choose the right chart | Misread or confusing visuals |
| 5. Guide the eye | Important information being missed |
| 6. Design for action | Insight without follow-through |
| 7. Test with users | Assumptions that do not match real use |
Each rule below includes a short example based on common situations in operations, sales, finance and product teams. The examples are illustrative rather than case studies, but they reflect patterns seen across many dashboard projects.
Rule 1: start from decisions
Before choosing metrics, ask what decisions the dashboard should support. Example: a retail operations manager originally received a dashboard with every available store metric. Interviews showed her main daily decision was where to send extra staff. The redesigned dashboard opens with store footfall against staffing levels, highlighting stores that are understaffed for expected demand.
Rule 2: show only what matters
Limit the main view to the few metrics that drive decisions, often five to nine, with detail available on demand. Example: a SaaS executive dashboard with twenty-four charts was reduced to six key metrics: new revenue, churn, active customers, activation rate, support backlog and net promoter score. Detailed breakdowns moved to linked pages.
Rule 3: give every number context
A number alone rarely tells people whether things are good or bad. Example: "Orders: 1,240" became "Orders: 1,240, up 8% on last week, target 1,200". A small trend line showed whether the increase was part of a pattern or a single spike.
- Target: How the value compares with the goal.
- Previous period: Change since last week, month or year.
- Trend: A small sparkline showing direction.
- Benchmark: How the value compares with peers or averages.
Rule 4: choose the right chart
| Question | Better chart | Example |
|---|---|---|
| How has this changed over time? | Line chart | Monthly revenue trend |
| How do categories compare? | Bar chart | Sales by region |
| Is this on target? | Bullet chart or KPI card | Quarterly target progress |
| What makes up the total? | Stacked bar | Revenue by product line |
Example: a pie chart with eleven product categories was replaced by a sorted horizontal bar chart, making it immediately clear which three categories drove most revenue.
Rule 5: guide the eye
Place the most important information where people look first, usually the top-left in left-to-right layouts, and use size and colour sparingly to signal importance. Example: in a logistics dashboard, late deliveries were moved from the bottom of the page to the top, shown in a prominent card with an alert colour, while routine metrics were placed lower in a neutral style. Our post on choosing UI colours covers using colour with purpose.
Rule 6: design for action
Dashboards should help people act, not just observe. Example: a sales dashboard showed three deals stuck in negotiation for over thirty days. Each row linked directly to the deal record with a button to schedule a follow-up, turning insight into action in one click.
- Link metrics to the records or reports behind them
- Add clear next steps for exceptions and alerts
- Allow filtering to the user’s own team or region
- Send notifications for thresholds that need attention
Rule 7: test with users
- Set realistic tasks: For example, "Which region is furthest behind target?"
- Observe: Watch how quickly and accurately people answer.
- Note confusion: Record where they hesitate or misread charts.
- Refine: Adjust layout, charts and labels.
- Track use: Monitor which parts of the dashboard are used after launch.
Example: testing a finance dashboard revealed that users misread a stacked area chart. Replacing it with separate line charts cut the time to answer key questions significantly. Our guide to running a usability test explains how to plan sessions like this.
Mistakes these rules prevent
- Adding every metric a stakeholder requests
- Showing numbers without targets or comparisons
- Using decorative charts that are hard to read
- Hiding urgent issues among routine metrics
- Launching without testing with the people who will use it
Revisiting these rules every few months keeps dashboards focused as teams, goals and data change.
These rules apply directly to Power BI; our guide to SLA dashboards in Power BI shows them on a service report.
Frequently asked questions
What is the most important dashboard design rule?
Start from the decisions users need to make. Every other choice follows from that.
How many metrics should a dashboard show?
Typically five to nine key metrics on the main view, with detail available on demand.
Are pie charts bad for dashboards?
They work for very simple part-to-whole comparisons, but bar charts are usually clearer, especially with many categories.
How do we know if a dashboard works?
Test with users on real questions and track which parts are used after launch.
Who should be involved in designing a dashboard?
The people who will use it, a designer, a data or analytics specialist and someone who understands the business decisions it supports.
Should dashboards be interactive?
Usually yes, with simple filters and drill-down, while keeping the default view clear for quick answers.
How often should dashboards be reviewed?
Every few months, or whenever goals or teams change, to remove unused metrics and add new ones that matter.