SKILL.md

--- frontmatter

name: product-health-dashboard-designer
description: Use when defining product analytics requirements. Use after product live. Produces KPI definitions, dashboard specifications, alert thresholds, and measurement methodology.

Product Health Dashboard Designer

Overview

Design dashboards and metrics that tell the story of product health. Define KPIs, specify visualizations, and establish alerting that enables proactive product management.

Core principle: Measure what matters, not what's easy. A good dashboard enables decision-making, not just data display.

When to Use

•Product launching to production
•Redesigning existing analytics
•Onboarding new team to product metrics
•Quarterly health check methodology

Output Format

yaml

product_health_dashboard:
  product: "[Product name]"
  version: "[YYYY-MM-DD]"
  audience: "[Who uses this dashboard]"
  
  north_star:
    metric: "[Primary success metric]"
    definition: "[Exactly how it's calculated]"
    data_source: "[Where data comes from]"
    current: "[Current value]"
    target: "[Goal value]"
    refresh_rate: "[Real-time | Hourly | Daily]"
  
  kpis:
    - category: "[Acquisition | Activation | Engagement | Retention | Revenue]"
      metrics:
        - name: "[Metric name]"
          definition: "[Calculation formula]"
          data_source: "[Where from]"
          visualization: "[Number | Chart type]"
          healthy_range: "[X to Y]"
          warning_threshold: "[Trigger for concern]"
          critical_threshold: "[Trigger for action]"
          trend_period: "[Days/weeks to show]"
  
  dashboard_layout:
    sections:
      - section: "[Section name]"
        purpose: "[What decisions this enables]"
        components:
          - type: "[Big number | Line chart | Bar chart | Table]"
            metric: "[Metric name]"
            size: "[Full | Half | Quarter]"
            comparisons: ["[vs last period | vs target]"]
  
  alerts:
    - metric: "[Metric name]"
      condition: "[When to alert]"
      severity: "[Critical | Warning | Info]"
      channel: "[Slack | Email | PagerDuty]"
      recipients: ["[Team/person]"]
      runbook: "[Link to response procedure]"
  
  segments:
    - segment: "[User segment]"
      filter: "[How to identify]"
      rationale: "[Why this matters separately]"
  
  data_requirements:
    events: ["[Event name: description]"]
    properties: ["[Property: what it captures]"]
    implementation_notes: "[Technical requirements]"
  
  review_cadence:
    daily: "[What to check daily]"
    weekly: "[Weekly review focus]"
    monthly: "[Monthly deep dive]"

KPI Framework: AARRR

Acquisition

Metric	Definition	Healthy
New signups	Unique accounts created	Growing
Signup conversion	Signups / Landing page visitors	>2-5%
Cost per acquisition	Total spend / Signups	Decreasing
Channel attribution	Signups by source	Diversified

Activation

Metric	Definition	Healthy
Onboarding completion	Users completing setup / Signups	>60%
Time to first value	Time from signup to key action	Decreasing
Activation rate	Users performing key action / Signups	>40%

Engagement

Metric	Definition	Healthy
DAU/MAU	Daily active / Monthly active	>25%
Session frequency	Sessions per user per week	Stable/growing
Feature adoption	Users of feature X / Active users	Per feature target
Session duration	Average time in product	Appropriate for use case

Retention

Metric	Definition	Healthy
D1/D7/D30 retention	Users returning after N days	D1>40%, D7>20%, D30>10%
Churn rate	Users lost / Total users	<5% monthly
Cohort retention	Retention curves by signup cohort	Flattening curve

Revenue

Metric	Definition	Healthy
MRR	Monthly recurring revenue	Growing
ARPU	Revenue / Active users	Stable/growing
LTV	Lifetime value per customer	>3x CAC
Expansion revenue	Upgrades / Total revenue	>20%

Dashboard Design Principles

Layout Hierarchy

code

┌────────────────────────────────────────────────────────────┐
│ NORTH STAR METRIC                                    BIG   │
│ [Primary success metric with trend]                        │
├────────────────────────────────────────────────────────────┤
│ KEY HEALTH INDICATORS                                      │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐        │
│ │ Metric 1     │ │ Metric 2     │ │ Metric 3     │        │
│ │ [Trend chart]│ │ [Trend chart]│ │ [Trend chart]│        │
│ └──────────────┘ └──────────────┘ └──────────────┘        │
├────────────────────────────────────────────────────────────┤
│ DETAILED VIEWS                                             │
│ [Segment breakdowns, cohort analysis, feature metrics]     │
└────────────────────────────────────────────────────────────┘

Visualization Selection

Data Type	Recommended	Avoid
Single value + trend	Big number + sparkline	Pie chart
Time series	Line chart	Bar chart for >7 periods
Category comparison	Horizontal bar	Pie chart
Distribution	Histogram	Line chart
Funnel steps	Funnel visualization	Line chart

Alert Design

Threshold Setting

yaml

alert_thresholds:
  approach: "Statistical"
  method: "Mean ± 2 standard deviations over 30 days"
  
  example:
    metric: "Daily signups"
    mean: 100
    std_dev: 15
    warning_low: 70   # Mean - 2σ
    warning_high: 130 # Mean + 2σ
    critical_low: 55  # Mean - 3σ

Severity Levels

Severity	Criteria	Response
Critical	Business-impacting now	Immediate action, page on-call
Warning	Concerning trend	Investigate same day
Info	Notable but not urgent	Review in next meeting

Alert Hygiene

yaml

alert_principles:
  - "Every alert should be actionable"
  - "If nobody acts, remove the alert"
  - "Review alert fatigue monthly"
  - "Each critical alert needs a runbook"

Segmentation Strategy

Common Segments

Segment Type	Examples
User lifecycle	New (<7d), Active, Dormant, Churned
Plan tier	Free, Pro, Enterprise
Use case	By primary feature used
Size	SMB, Mid-market, Enterprise
Cohort	By signup month

Segment Dashboard

yaml

segment_view:
  primary_segment: "User lifecycle"
  default_view: "Active users"
  comparison: "Side-by-side segment comparison"
  
  metrics_per_segment:
    - "Segment size"
    - "Key action rate"
    - "Revenue contribution"
    - "Trend vs previous period"

Implementation Checklist

Before Launch

After Launch

• Dashboard built and tested
• Alerts delivering correctly
• Team trained on interpretation
• Review cadence established
• Runbooks created for critical alerts

Common Mistakes

Mistake	Problem	Fix
Too many metrics	Dashboard overload	Focus on 5-7 key metrics
Vanity metrics	Looks good, not actionable	Tie to decisions
No targets	Can't assess health	Set clear targets
Raw numbers only	Missing context	Add comparisons, trends
Alert flood	Fatigue, ignored	Reduce to truly actionable
No segments	Masked problems	At least lifecycle + tier