Dashboard Design Prompt Templates

AI prompt templates for designing dashboards. Create effective business intelligence visualizations.

Overview

A good dashboard answers questions at a glance. A bad one creates more confusion than clarity. These prompts help you design dashboards that people actually use, whether you're building in Tableau, Power BI, Looker, or a custom solution. Focus on what decisions your dashboard should support, not just what data you can show.

Best Practices

1

Start with the questions your audience needs answered, not the data you have

2

Limit to 5-7 visualizations per dashboard. More than that overwhelms users

3

Put the most important information in the top-left (where eyes go first)

4

Include context: comparisons to targets, prior periods, or benchmarks

5

Test with actual users. Watch them try to answer questions using your dashboard

Prompt Templates

1. Dashboard Requirements Gathering

Help me define requirements for a dashboard.

Business function: [TEAM_OR_DEPARTMENT]
Primary users: [WHO_WILL_USE_IT]
Decisions to support: [WHAT_DECISIONS]
Key questions it should answer:
[LIST_QUESTIONS]

Available data sources: [DATA_SOURCES]
Refresh frequency needed: [REAL_TIME/HOURLY/DAILY]
Business function: E-commerce operations. Primary users: Operations manager and fulfillment team leads. Decisions to support: Daily staffing levels, identifying shipping delays, escalating inventory issues. Key questions: 1) How many orders need to ship today? 2) Are we on track to meet SLAs? 3) Which products are running low? 4) Are there any carrier issues? Available data: Order management system, inventory database, shipping carrier APIs. Refresh frequency: Every 15 minutes during business hours.
**Dashboard Requirements: E-commerce Operations**

**Purpose statement**: Enable operations team to manage daily fulfillment, identify issues before they impact customers, and make staffing decisions.

**User workflow**:
1. Morning: Check order volume, set staffing for the day
2. Throughout day: Monitor SLA compliance, spot delays
3. As needed: Investigate specific issues, escalate problems

**Required views**:

**Section 1: Today's Snapshot (Top row, KPI cards)**
- Orders to ship today (count)
- Orders shipped so far (count + % of target)
- Orders at risk of missing SLA (count, red if > 0)
- Average fulfillment time today vs. target

**Section 2: Order Pipeline (Middle left)**
- Stacked bar or funnel: Orders by status (Received → Processing → Packed → Shipped → Delivered)
- Color code by SLA status (on track = green, at risk = yellow, breached = red)

**Section 3: SLA Tracker (Middle right)**
- Line chart: Rolling 7-day SLA compliance rate
- Current day's projected compliance based on current pace
- Threshold line at target (e.g. 95%)

**Section 4: Inventory Alerts (Bottom left)**
- Table: Products below reorder point
- Columns: SKU, name, current stock, days of supply, reorder status
- Sorted by days of supply ascending (most urgent first)

**Section 5: Carrier Performance (Bottom right)**
- Table: Carrier, packages in transit, average delivery time, issues today
- Highlight any carrier with delivery issues

**Filters**:
- Date (default: today)
- Warehouse location (if multiple)
- Carrier (for drilling down)

**Alerts to configure**:
- Slack/email when SLA compliance drops below 90%
- Alert when any SKU hits zero inventory
- Alert when carrier shows unusual delay patterns

**Data requirements**:
- Order data: Real-time from OMS
- Inventory: Sync every 15 minutes
- Carrier tracking: API refresh every 15 minutes

**Success criteria**: Operations manager can identify and respond to fulfillment issues within 30 minutes of occurrence.

Starting a new dashboard project

  • Interview users about their workflow, not just what data they want to see
  • Ask 'what would you do differently if you knew X?' to identify truly actionable metrics

2. Dashboard Layout Design

Design a layout for a dashboard with these components:

Components needed:
[LIST_OF_CHARTS_AND_METRICS]

Screen size: [DIMENSIONS_OR_DEVICE]
Tool: [TABLEAU/POWER_BI/LOOKER/CUSTOM]
User context: [HOW_THEY'LL_VIEW_IT]

Provide a layout with positioning rationale.
Components: 4 KPI cards (revenue, orders, conversion rate, avg order value), revenue trend line chart, top products bar chart, traffic sources pie chart, geographic heat map. Screen size: 1920x1080 desktop monitor. Tool: Tableau. User context: Executive daily check-in, usually 2-3 minutes.
**Layout Design: Executive Daily Dashboard**

```
+--------------------------------------------------+
| [Filters: Date Range | Region] [Logo] | <- 60px header
+--------------------------------------------------+
| Revenue | Orders | Conv Rate | Avg Order | <- 120px KPI row
| $1.2M | 4,523 | 3.2% | $265 |
| ▲ 12% | ▲ 8% | ▼ 0.2% | ▲ 5% |
+--------------------------------------------------+
| | |
| Revenue Trend | Top Products | <- 400px
| (Line chart, 30 days) | (Horizontal bar) |
| | |
+--------------------------------------------------+
| | |
| Geographic Heat Map | Traffic Sources | <- 300px
| (By region/country) | (Donut chart) |
| | |
+--------------------------------------------------+
```

**Positioning rationale**:

1. **KPIs at top**: First thing executives see. Shows health at a glance. Include comparison (▲▼) so they don't have to calculate change.

2. **Revenue trend prominent**: Largest chart, left side. Eyes naturally go here first after KPIs. 30-day view shows recent performance without noise.

3. **Top products top-right**: Answers 'what's selling?' which is a common follow-up question after seeing revenue.

4. **Map bottom-left**: Geographic view for companies with regional presence. Less prominent because it's reference info, not daily action item.

5. **Traffic sources bottom-right**: Shows acquisition mix. Useful context but not primary focus for operations.

**Tableau-specific tips**:
- Use a vertical container for KPIs so they resize proportionally
- Set dashboard to 'Automatic' sizing with min dimensions of 1200x800
- Use floating objects sparingly (only for logo and maybe filters)
- Add parameter for date range toggle (Last 7 / 30 / 90 days)

**Color guidance**:
- Revenue: Blue (positive association)
- Conversion rate: Use conditional formatting (green > target, red < target)
- Keep it to 3-4 colors max to avoid visual noise

When you know what to show but need to figure out how to arrange it

  • The Z-pattern: Users scan top-left → top-right → bottom-left → bottom-right. Put your story in that order
  • White space is valuable. Crowded dashboards feel overwhelming even if each chart is good

3. Dashboard Review and Critique

Review this dashboard design and suggest improvements.

Current dashboard description:
[DESCRIBE_WHATS_ON_IT]

Purpose: [WHAT_ITS_SUPPOSED_TO_DO]
User feedback: [WHAT_USERS_SAY]
My concerns: [WHAT_YOU_THINK_IS_WRONG]

Provide specific, actionable recommendations.
Current dashboard: 12 charts on one page - 3 pie charts (traffic sources, device type, browser), 4 line charts (sessions, pageviews, bounce rate, avg session duration), 2 bar charts (top pages, top landing pages), 3 KPI cards (users, sessions, conversion rate), plus a data table with 20 rows of detailed metrics. Purpose: Help marketing understand website performance. User feedback: 'It's overwhelming', 'I can never find what I need', 'I just export to Excel anyway'. My concerns: Too much data, no clear story.
**Dashboard Critique: Website Performance**

**Core problem**: This is a data dump, not a dashboard. It shows everything possible instead of answering specific questions.

**Issues identified**:

1. **Too many charts (12 + table)**: Violates the 5-7 rule. Users can't process this much information at once.

2. **Three pie charts**: Pie charts are hard to read, and you don't need three. Device type and browser are rarely actionable.

3. **Redundant metrics**: Sessions and pageviews tell similar stories. Bounce rate and avg session duration are inversely related.

4. **No comparisons**: Raw numbers without context (vs. last week? vs. target?) are hard to interpret.

5. **Data table with 20 rows**: If users export to Excel, the dashboard isn't serving them.

**Recommended redesign**:

**Version 1: Executive View (Default)**
```
+----------------------------------------+
| Sessions | Conversion | Revenue | vs LW |
+----------------------------------------+
| Sessions trend (14 days) with comparison|
+--------------------+-------------------+
| Top landing pages | Traffic sources |
| (5 rows) | (bar chart) |
+--------------------+-------------------+
```

4 components total. Answers: How's traffic? Is it converting? Where's it coming from?

**Version 2: Deep Dive (Separate tab)**
Keep the detailed metrics for analysts who need them, but don't make executives wade through it.

**Specific changes**:

1. **Merge traffic sources pie into single horizontal bar**: Easier to compare, takes less space.

2. **Kill device/browser charts**: Unless you're debugging tech issues, these aren't actionable. Add as filters instead.

3. **Replace data table with top 5 list**: If they need full export, add a download button, but don't show 20 rows on the dashboard.

4. **Add comparisons everywhere**: 'Sessions: 45K (▲ 12% vs last week)' instead of just 'Sessions: 45K'.

5. **Add one insight text box**: 'Traffic up 12% driven by organic search. Conversion rate stable.' Human context helps.

**Test the redesign**: Show users the new version and ask them to answer specific questions. Time how long it takes compared to the old version.

When an existing dashboard isn't working well

  • If users export to Excel, your dashboard is failing at its job. Figure out why
  • The best dashboards have opinions. They highlight what matters, not everything that exists

4. KPI Card and Metric Design

Design the KPI cards/metrics section for my dashboard.

Metrics to display: [LIST_OF_METRICS]
Comparisons needed: [VS_WHAT]
Audience: [WHO]
Action threshold: [WHEN_SHOULD_THEY_ACT]

How should each metric be formatted and what context should accompany it?
Metrics: Monthly recurring revenue (MRR), customer count, churn rate, net revenue retention, average revenue per user. Comparisons: vs. last month, vs. same month last year, vs. plan. Audience: SaaS executive team. Action threshold: Churn above 5% is a problem, NRR below 100% is a problem.
**KPI Card Design: SaaS Executive Metrics**

**Layout**: 5 cards in a horizontal row, each card structured identically.

**Card template**:
```
+-------------------+
| MRR | <- Label (gray, small)
| $2.4M | <- Value (large, bold)
| ▲ 4.2% MoM | <- Primary comparison (with arrow)
| +18% YoY | 102% | <- Secondary comparisons
| plan |
+-------------------+
```

**Individual card specifications**:

**1. MRR (Monthly Recurring Revenue)**
- Format: Currency, abbreviated ($2.4M not $2,400,000)
- Primary comparison: vs. last month (MoM %)
- Secondary: vs. same month last year (YoY %), vs. plan (%)
- Color: Green if above plan, red if below
- Threshold: No specific alert (tracked for growth)

**2. Customer Count**
- Format: Number with comma (12,450)
- Primary comparison: Net change this month (+125)
- Secondary: New customers, churned customers breakdown
- Color: Standard (no threshold)
- Tooltip: Show gross adds vs. churn to explain net change

**3. Churn Rate**
- Format: Percentage with 1 decimal (4.2%)
- Primary comparison: vs. last month (absolute change: +0.3%)
- Secondary: vs. 12-month average
- Color: **Red if > 5%**, yellow if 4-5%, green if < 4%
- Alert: Requires action if red

**4. Net Revenue Retention (NRR)**
- Format: Percentage (108%)
- Primary comparison: vs. last month
- Secondary: vs. plan, vs. industry benchmark (typically 100-120%)
- Color: **Red if < 100%**, green if > 100%
- Explanation tooltip: 'Revenue from existing customers including expansions and churn'

**5. ARPU (Average Revenue Per User)**
- Format: Currency ($192)
- Primary comparison: vs. last month
- Secondary: vs. same month last year
- Color: Standard (track for trends, not threshold)
- Segmentation available: Click to see ARPU by plan tier

**Design principles applied**:

1. **Consistent format**: Same structure for all cards so eyes know where to look
2. **Color with meaning**: Only use red/green for metrics with clear good/bad thresholds
3. **Multiple comparisons**: One number without context is meaningless
4. **Abbreviations**: Executives don't need to see all digits
5. **Actionable thresholds**: Churn and NRR have clear 'pay attention' signals

**Interactivity**:
- Click any card to drill down to trend chart
- Hover for detailed breakdown and calculation explanation

Designing the summary metrics at the top of any dashboard

  • Every number needs context. A standalone metric like 'Revenue: $1.2M' is useless without comparison
  • Limit color coding to metrics with clear thresholds. If everything is colored, nothing stands out

Common Mistakes to Avoid

Designing for the data you have instead of the questions users need answered

Including too many charts because each one seemed useful in isolation

Using color inconsistently. Red should always mean 'problem' if it means that anywhere

Frequently Asked Questions

A good dashboard answers questions at a glance. A bad one creates more confusion than clarity. These prompts help you design dashboards that people actually use, whether you're building in Tableau, Power BI, Looker, or a custom solution. Focus on what decisions your dashboard should support, not just what data you can show.

Related Templates

Have your own prompt to optimize?