Overview

Root-Cause Analysis diagnoses why a metric moved. Shadow decomposes the change with a MECE framework — starting wherever your question points — isolates the driver with evidence, and explains it conversationally, not as a data dump.

What it does

• Structured decomposition — breaks revenue, ROAS, CPA, CVR, traffic or spend into their contributing parts.

• Starts at your question — "why did revenue drop?" or "what happened to CPC?" — and works the tree from there.

• Finds the driver — isolates the real cause with evidence, not correlation.

• Compares the right windows — most recent 30 days vs prior 30 by default, or your dates.

• Plain-language answer — the why, explained conversationally.
  
  

How to use it

1. Connect Shopify, Meta and Google — the core decomposition needs them.

2. Ask a "why" question — name the metric and, if you want, the period.

3. Shadow decomposes the change and tests each branch against the data.

4. Get the root cause — with the evidence and what to do next.

Run it in-app or from Claude / ChatGPT via the Shadow MCP.

What it needs

Shopify, Meta and Google connected for a full decomposition. With nothing connected, Shadow walks the methodology using the numbers you describe.

What you get

A diagnostic answer: the metric change decomposed, the root cause identified with evidence, and a conversational explanation of why it happened.

Not the right skill when…

• You want a period summary → use Performance Report, instead

• You want a full business audit → use Growth Audit, instead

• You want conversion-page diagnosis → use CRO Audit, instead