metric-diagnosis

What problem does it solve?

This skill helps analysts systematically investigate unexpected metric changes by applying structured root-cause analysis across defined data dimensions and factors to identify true causes rather than quick conclusions.

Core Features & Use Cases

• 4D segmentation: People, Geography, Technology, and Time to isolate affected segments.
• Intrinsic vs Extrinsic analysis: Distinguish internal changes (code, experiments) from external events (market, campaigns).
• Hypothesis testing: Generate and test hypotheses with a structured table and data-driven validation.
• North Star alignment: Assess impact on top-line metrics and strategic goals.

Quick Start

1. Provide the metric name, timeframe, and context (e.g., WAU dropped 12% in the last 7 days).
2. Include any data-quality concerns and available data segments (segments by user type, geography, platform, time).
3. Request a phased diagnosis covering Phase 1: data quality, Phase 2: hypotheses, Phase 3: hypothesis testing, Phase 4: North Star impact, and Phase 5: recommended actions.