Files
gh-coalesce-labs-catalyst-p…/commands/segment_analysis.md
2025-11-29 18:14:34 +08:00

3.2 KiB

description, category, tools, model, version
description category tools model version
Analyze user segments and cohorts for targeted insights analytics Task, TodoWrite inherit 1.0.0

Segment Analysis

Deep-dive into specific user segments, cohorts, or customer groups using PostHog data.

Usage

/segment-analysis <segment-description>

Examples:
  /segment-analysis "users from paid plans vs free plans"
  /segment-analysis "power users who use feature X daily"
  /segment-analysis "users who churned in last 30 days"
  /segment-analysis "cohort: signed up in Q4 2024"

What This Analyzes

User Segments

  • By plan type (free, pro, enterprise)
  • By geography (country, region)
  • By acquisition source (organic, paid, referral)
  • By behavior (power users, casual users, at-risk)

Cohort Analysis

  • By signup date (monthly, weekly cohorts)
  • By first feature used
  • By activation milestone reached
  • By engagement level

Comparison Analysis

  • Segment A vs Segment B
  • Before/after feature launch
  • Treatment vs control (A/B tests)
  • Time period comparisons

Example Analyses

Plan Comparison

/segment-analysis "Compare engagement patterns between free and paid users: session frequency, feature usage, retention"

Power User Identification

/segment-analysis "Identify our power users: who are they, what features do they use, what's their profile?"

Churn Analysis

/segment-analysis "Analyze users who churned: what were their last actions, which features didn't they use?"

Geographic Performance

/segment-analysis "Compare conversion rates and engagement across our top 5 countries"

Cohort Retention

/segment-analysis "Show retention curves for each monthly signup cohort in 2024"

Output Format

Analysis typically includes:

  • Segment characteristics (size, demographics, behavior)
  • Key metrics for each segment
  • Comparative insights between segments
  • Behavior patterns unique to segment
  • Recommendations for targeting or improvement

Segmentation Criteria

You can segment by:

  • Demographics: Country, language, device type
  • Behavior: Feature usage, session frequency, engagement score
  • Business: Plan type, payment history, LTV
  • Temporal: Signup date, last active, tenure
  • Custom: Any event or property in PostHog

Advanced Analysis

Multi-dimensional Segmentation

/segment-analysis "Power users (5+ sessions/week) from enterprise plans who use feature X"

Funnel by Segment

/segment-analysis "Compare signup to activation funnel for organic vs paid traffic"

Retention by Segment

/segment-analysis "30-day retention by initial feature used"

Tips for Better Analysis

  1. Be specific - Define your segment clearly
  2. Ask for comparisons - "vs" between segments reveals insights
  3. Look for patterns - What makes segments different?
  4. Consider time - Trends over time matter
  5. Combine criteria - Multi-dimensional segments can be revealing

Context Cost

Plugin uses ~40k tokens. Disable when analysis is complete:

/plugin disable catalyst-analytics

See also: /analyze-user-behavior, /product-metrics