Overview

Incrementality estimates how much of your paid revenue is truly incremental — using geo holdouts, spend-scaling tests and iROAS math. It compares platform-reported ROAS to measured incremental ROAS, flags channels that over-claim, and recommends the next test on the BEATS hierarchy: experiments over models, models over tracking.

What it does

• True ROAS — measured incremental ROAS vs platform-reported.

• Over-claim flags — channels taking credit they didn't earn.

• Revenue at risk — what you'd lose if you cut a channel.

• Test recommendation — the next incrementality test to run.

• Evidence hierarchy — experiments first, tracking last.
  
  

How to use it

1. Connect Meta, Google and TikTok (and share any holdout results).

2. Ask for an incrementality analysis.

3. Shadow triangulates the strongest available evidence.

4. Get true ROAS — and the next test to run.

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

What it needs

Connected ad platforms; geo-holdout or spend-test results sharpen it. MMM outputs, if any, are treated as manual inputs.

What you get

True incremental ROAS by channel, over-claim flags, revenue-at-risk estimates and a recommended next test.

Not the right skill when…

• You want a tactical Meta audit → use Meta Ads Analyzer, instead

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

• You want a single promo's impact → use Event Effect Analysis, instead