MTA shows observable click paths. MMM examines portfolio-level patterns across sales, spend, promotions, and time. A disagreement is a prompt to reconcile scope and form a testable hypothesis.
Key moments
- 0:00 When MTA and an MMM disagree, neither is necessarily broken: they record and grade different things.
- 0:45 What MTA can show about an observed customer journey, and why it cannot say which touch caused the sale or credit ads people only view.
- 2:11 What an MMM adds by putting spend, sales, promotions, and seasonality on one timeline, including channels people do not click.
- 3:38 Examples of why the two disagree, including promotions and view-only impressions.
- 4:59 Before comparing them: align channel names, the revenue definition, and promotions.
- 6:25 When they still disagree, form a hypothesis and settle it with a lift test.
- 7:48 The difference between ROAS and incremental ROAS.
- 8:20 Summary: how to use MTA and an MMM together.
Read the related guide: Why MTA and MMM Don’t Agree—and Why That’s Useful