Create Board

create-board

When to use this skill: When users want to create a new analytics board, set up monitoring dashboards, or organize multiple reports into a cohesive view.

Instructions

Overview

This skill helps you create a new Mixpanel board with one or more reports. Boards are collections of reports (Insights, Funnels, Flows, Retention) organized into sections for monitoring key metrics.

Step-by-step workflow

1. Understand the request

   • What metrics or questions should this board answer?
   • Who is the intended audience (PM, executive, data team)?
   • What cadence will this be reviewed (daily, weekly, monthly)?

2. Determine reports needed

   • For each metric or question, identify the appropriate report type:
     • Insights: Trends, totals, averages over time
     • Funnels: Conversion flows between steps
     • Flows: User pathways and drop-off points
     • Retention: Cohort retention curves
   • Aim for 3-8 reports per board (enough to be comprehensive, not overwhelming)

3. Organize into sections

   • Group related reports into logical sections:
     • "Acquisition" (signup metrics, channel performance)
     • "Activation" (onboarding, first actions)
     • "Engagement" (DAU/MAU, feature usage)
     • "Retention" (cohort curves, churn signals)
     • "Revenue" (conversion, ARPU)
   • Each section should tell part of the story

4. Create the board

   Two-step process required:

   Step 4a: Create reports and collect query_ids

   • For each report you want on the board, run:

     Run-Query(
       project_id=X,
       report_type='insights',  # or 'funnels', 'flows', 'retention'
       report={...report definition...},
       skip_results=true    ← Important: skips execution, just saves the query
     )
     

   • This returns a query_id (e.g., "abc123")
   • Collect all query_ids you need for the dashboard

   Step 4b: Create the dashboard

   • Use Create-Dashboard with collected query_ids:

     Create-Dashboard(
       project_id=X,
       title="Dashboard Name",
       rows=[
         {
           "contents": [
             {"type": "text", "html_content": "<h2>Growth Metrics</h2>"}
           ]
         },
         {
           "contents": [
             {"type": "report", "query_id": "abc123", "name": "DAU Trend"},
             {"type": "report", "query_id": "def456", "name": "Conversion Funnel"}
           ]
         },
         {
           "contents": [
             {"type": "text", "html_content": "<h2>Retention</h2>"}
           ]
         },
         {
           "contents": [
             {"type": "report", "query_id": "ghi789", "name": "30-Day Retention"}
           ]
         }
       ]
     )
     

   Dashboard structure:

   • Max 30 rows per dashboard
   • Each row: 1-4 items (reports or text cards)
   • Use text cards (type: "text") as section headers
   • Reports need query_id + name (description optional)
   • Text cards support HTML: h1, h2, h3, p, ul, li, strong, em

5. Confirm with the user

   • Create-Dashboard returns a dashboard URL
   • Share the URL with the user
   • Explain what each section measures and why it matters
   • Suggest review cadence based on the metrics

Report selection guide

For monitoring product health:

• Insights: DAU/MAU ratio, session frequency
• Retention: D1/D7/D30 curves
• Funnels: Core activation flow

For growth analysis:

• Insights: New user signups by channel
• Funnels: Signup → activation → conversion
• Flows: Drop-off points in onboarding

For feature launches:

• Insights: Feature adoption rate over time
• Funnels: Feature discovery → first use → repeat use
• Retention: Retention of users who used the feature

Example interaction

User: "Create a weekly executive dashboard for me"

Assistant:

1. Understands this needs high-level business metrics
2. Determines key sections:
   • Growth (new users, activation rate)
   • Engagement (DAU, sessions per user)
   • Retention (cohort curves)
   • Revenue (conversion rate, upgrades)
3. Creates 5-6 reports using Run-Query with skip_results=true
4. Collects query_ids from each response
5. Calls Create-Dashboard with rows organized by section
6. Returns: "I've created your Executive Weekly Dashboard with 6 reports across 4 sections. It covers growth, engagement, retention, and revenue. View it here: [board_url]. I recommend reviewing this every Monday morning."

Common mistakes to avoid

• Too many reports: Keep boards focused (3-8 reports). More than 10 becomes hard to digest.
• Mixed audiences: Don't combine executive summaries with deep data dives. Create separate boards.
• No context: Always explain what each section measures and why it matters.
• Static snapshots: Boards should update automatically. Avoid date filters locked to specific ranges unless it's for historical analysis.
• Forgetting skip_results=true: Always use this when creating reports for dashboards, otherwise you'll execute queries unnecessarily.

Success criteria

• Board has a clear purpose and audience
• Reports are organized into logical sections
• Each report answers a specific question
• User understands what the board measures and when to check it