Root Cause Investigation

Overview

Root Cause Investigation runs when a key metric drops or spikes unexpectedly, a stakeholder asks "why did this happen," or a post-mortem needs an evidence-based answer instead of a guess.

What it covers

• Validates the change is real using a z-score against normal variance, and closes the investigation early if it's within range.
• Plots the metric over time to pinpoint whether the change was a sudden step or a gradual drift.
• Decomposes the metric into its parts (for example, revenue as volume times price times mix) to see which component is driving the change.
• Drills down across geography, platform, channel, and segment, ranked by absolute contribution to the change.
• Tests explicit hypotheses against a change log of known events, then documents the primary driver with tiered recommendations.

Part of Data Analytics Skills

This is one of 31 skills in the Data Analytics Skills library, in the Data Analysis & Investigation group.

Installation

1. Clone the repo once: git clone https://github.com/nimrodfisher/data-analytics-skills.git
2. Copy just this skill: cp -r data-analytics-skills/03-data-analysis-investigation/root-cause-investigation ~/.claude/skills/ (or ~/.agents/skills/ for Codex)
3. Ask your agent something like "why did our activation rate drop 12% last week" and it activates on its own.

See the Data Analytics Skills entry for the full library.