Product Analytics: AI Skill for Product Management
Product Analytics is an AI skill that turns your assistant's data reads into decisions: question first, metrics you can act on, segmented results, and an insight-to-action story. No vanity metrics, no correlation dressed as causation.
npx skills add Uxcel-Lab/product-skills --skill pm-analytics
What this skill does
Helps an AI assistant run product analytics that produces a decision, not a dashboard nobody acts on.
- Starts from the decision or question, not the data.
- Picks metrics you can act on over vanity numbers like page views.
- Separates signal from noise with rolling windows and repeated patterns.
- Names the data traps: correlation as causation, survivorship bias, small samples, cherry-picking.
- Segments results by device, source, and behavior, since averages hide the story.
- Pairs numbers with user conversations to explain the why.
- Ends every analysis in an insight, impact, and action story.
- Flags context-dependent choices like method, cohort type, and attribution instead of running everything.
When to use it
Use it when you need an AI assistant to:
- analyze product data or metrics;
- plan what to measure for a feature or launch;
- build a funnel, cohort, segment, or retention analysis;
- explain why a metric moved;
- check an analysis for bias or vanity metrics;
- turn analytics into a recommendation.
Decisions this skill helps you make
| Decision | Options |
|---|---|
| Analysis method | Funnel for conversion, cohort for retention, segment to explain variance, journey for cross-touchpoint behavior |
| Leading vs. lagging indicators | Leading to act early, lagging to confirm outcomes, or paired |
| Tooling | GA for web traffic, Mixpanel or Amplitude for events, a warehouse for depth and history |
| Cohort type | Behavioral, time-based, or acquisition |
| Retention window | N-day, weekly, monthly, or custom cycles matched to usage rhythm |
| Attribution model | Last-click for simple cases, linear or time-decay across channels |
| Qualitative depth | Light checks for clear patterns, interviews when the why carries a big decision |
How the skill works
1
Establish context
The skill identifies the decision to inform, the goal it supports, and the data available.
2
Apply the core
It starts from the question, picks metrics you can act on, segments results, and adds the user's why.
3
Surface trade-offs
It presents the analysis choices that depend on context: method, cohort type, retention window, and attribution.
Try it with
Related skills
1.0.0
MIT
Product Management
Processes
