Can we try Google LightweightMMM?

After Apple killed user-level tracking in iOS14, Facebook went all in on Marketing Mix Modeling with Robyn, an open source library. Now Google is getting into the game with an *unofficial* library called LightweightMMM, and it couldn’t be more different.


Automated MMM

Marketing Mix Modeling has historically been a manual process...More


Mike Taylor

Built a 50-person growth agency.
💪 Useful 0
😓 Difficult 0
🎉 Fun 0
😴 Boring 0
🚨 Errors 0
😕 Confusing 0
🤓 Interesting 0
Free access for email subscribers.
Python experience recommended.
1. Scenario
Your boss sends you an email. She wants you to check out the new automated marketing mix modeling solution from Google.
Charlotte Cook
at Vexnomics

Google MMM

Hey I saw this Twitter thread on Google getting into the MMM space. It looks similar to Facebook Robyn but this one is Bayesian, which is promising.

Can you check out their solution and see how it works?


This course is a work of fiction. Unless otherwise indicated, all the names, characters, businesses, data, places, events and incidents in this course are either the product of the author's imagination or used in a fictitious manner. Any resemblance to actual persons, living or dead, or actual events is purely coincidental.

2. Brief

Google LightweightMMM is an open-source and automated marketing mix modeling tool built by Google engineer, though it’s not an official project. It uses modern techniques to reduce human bias by automating the modeling process. For example it uses Bayesian MCMC to automatically find the most probable model parameters, which means less manual work needed. The use of a Bayesian MCMC algorithm separates the library philosophically from Facebook’s Robyn, which uses Ridge Regression paired with Nevergrad’s evolutionary algorithm. Through use of Bayesian priors, the model can ensure more plausible results in the final model, so you are less likely to get unrealistic estimations. The library is based on several papers published by Google on Bayesian MMM, for example Bayesian Methods for Media Mix Modeling with Carryover and Shape Effects (Jin, et al.). It allows for geo-level modeling right out of the box, which further distinguishes it from Facebook Robyn, which does not offer this feature at the time of release. Seasonality is modeled automatically, and you can add organic variables in, separate to media spend. After running the model LightweightMMM displays a series of charts for the model showing the impact of each channel, the response curves for each channel (how they behave at each spend level) and the predicted vs actual results over time. You can also optimize your media budget using the parameters from the model to determine what to spend in each channel to maximize your ROI.

3. Tutorial

Hey, let me run you through Google's lightweight. Mmm. Library. Uh, so we're using Google CoLab, which is like a Jupiter notebook.

Lightweight (Bayesian) Marketing Mix Modeling
4. Exercises
5. Certificate

Share This Course