How does it help if a feature misses its targets?

It uses a built-in Root Cause Analysis framework to identify if the failure was a discovery problem, a quality issue, a lack of relevance, or a frequency problem.

What information does the /feature-results skill require?

The skill needs the feature name, the original hypothesis (which it can often pull from your PRD), and actual performance metrics or access to an analytics MCP.

Can this skill help improve my future impact estimates?

Yes, the skill includes a calibration section that compares original estimates to actual results, helping you identify if your team consistently over or under-estimates impact.

Where are the analysis reports stored?

All generated reports are saved to the 'thoughts/shared/pm/analyses/' directory in your workspace for easy retrieval and sharing.

PM Post-Launch Analysis

Name: PM Post-Launch Analysis
Author: coalesce-labs

bycoalesce-labs

•

Collaboration Tools

Analyzes feature performance and documents post-launch learnings through data-driven results and root cause diagnosis.

The feature-results skill provides a structured framework for product managers and engineering teams to evaluate the success of a launch directly within Claude Code. It automates the creation of comprehensive results documents by comparing original hypotheses against actual performance data, performing segment-level analysis, and conducting root cause investigations for missed targets. This skill ensures that every launch contributes to institutional knowledge, offering actionable recommendations on whether to scale, iterate, or sunset features while providing tailored communication templates for various stakeholders.

Key Features

01Multi-dimensional segment breakdown (e.g., user tier, power users, geography)

02Structured Root Cause Analysis (RCA) for discovery, quality, and relevance issues

03Automated documentation generation saved to persistent workspace memory

04Stakeholder communication guidance for executives, engineering, and sales teams

05Hypothesis vs. actual metric calibration to improve future estimation accuracy

069 GitHub stars

Use Cases

01Conducting post-ship reviews for feature launches to capture organizational learnings

02Evaluating A/B test outcomes to determine the next product iteration

03Performing quarterly portfolio analysis to calibrate team impact estimates

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add coalesce-labs/catalyst feature-results

For use in Claude.ai and ChatGPT

Key Features

01Multi-dimensional segment breakdown (e.g., user tier, power users, geography)

02Structured Root Cause Analysis (RCA) for discovery, quality, and relevance issues

03Automated documentation generation saved to persistent workspace memory

04Stakeholder communication guidance for executives, engineering, and sales teams

05Hypothesis vs. actual metric calibration to improve future estimation accuracy

069 GitHub stars

Use Cases

01Conducting post-ship reviews for feature launches to capture organizational learnings

02Evaluating A/B test outcomes to determine the next product iteration

03Performing quarterly portfolio analysis to calibrate team impact estimates