What It Takes to Have an MMM You’ll Actually Use

Episode 27
Listen on Spotify

Everyone wants one number from an MMM, but Alex Nagourney of Sonos and Ed Staples of Smalls agree it’s risky to treat it as the truth. Alex spent years at Netflix and a small startup without an MMM, then joined Sonos, where one is part of how the whole company plans. Ed has bought MMMs at many companies and still thinks most teams misuse them.

In this episode, Haus’ Olivia Kory and Andrew Bubnar get into why most MMM problems aren’t about the math. They’re about how people use the numbers. Ed makes the case that marketers should manage the inputs, not just read the outputs, and Alex explains what it takes to get finance and other teams comfortable with a number built on assumptions.

In this episode:

  • Hot takes on ad volume, and whether spending millions a day is really a different problem than spending tens of thousands
  • Why Ed would buy an MMM over building one, and why Alex’s team never seriously considered building
  • How marketing and data science should split the work
  • Evaluation criteria for choosing an MMM partner, including the “vibe check”
  • Why marketers need to think like data architects
  • Why FP&A wants a single number, and why Ed says not to treat it as truth
  • What you need to trust an MMM’s outputs

About Haus:

Haus is the causal marketing measurement platform built for performance-driven brands and enterprises. From GeoLift incrementality testing to Causal MMM, Haus helps marketing and finance leaders understand what's actually driving growth — so they can make smarter business decisions.

🔗 Learn more about Haus: https://www.haus.io

📅 Book a demo: https://www.haus.io/demo

💡 Follow Haus on LinkedIn: https://www.linkedin.com/company/haus-analytics

Latest Open Haus episodes

See all

Cultivating an Experimental Mindset featuring Squarespace’s Josh Groth

Cultivating an Experimental Mindset featuring Squarespace’s Josh Groth

Squarespace Head of Marketing Josh Groth on why “Why?” is the best question a marketer can ask, and how it sparks incrementality tests that move the business.

Can A Model Predict Incrementality?

Can A Model Predict Incrementality?

Stanford Professor Brett Gordon joins the podcast to discuss his influential research around predicting incremental impact of marketing campaigns using experiments.

What To Do After A Bad Incrementality Test Result

What To Do After A Bad Incrementality Test Result

Haus Measurement Strategists Dean Gordon and Ike Armstrong sit down to talk about next steps after that difficult moment when you get a test result that “punches you in the chin.”