Skip to main content


Robyn is an experimental, ML-powered and semi-automatetd Marketing Mix Modeling (MMM) open source package.

A New Generation of Marketing Mix Modeling

Robyn aims to reduce human bias in the modeling process, esp. by automating modelers decisions like adstocking, saturation, trend & seasonality as well as model validation. Moreover, the budget allocator & calibration enable actionability and causality of the results.

[object Object]

Reduces human bias

  • Automated hyperparameter optimization with evolutionary algorithms from Facebook's AI library Nevergrad
  • Ridge regression in order to regularize multi-collinearity and prevent overfitting
  • Facebook's Prophet library to automatically decompose the trend, seasonality and holidays patterns

[object Object]

Aligns with the ground-truth

  • It calibrates models based on ground-truth methodologies (Geo-based, Facebook lift, MTA, etc.)
  • Facebook Nevergrad's multi-objective optimization minimizing the error between MMM prediction and ground-truth

[object Object]

Enables actionable decision making

  • Budget allocator using a gradient-based constrained non-linear solver to maximize the outcome by reallocating budgets
  • Enables frequent modeling outcomes due to stronger automation
  • Allows intuitive model comparisons via automatically generated model one-pagers

[object Object]

Private by Design

  • Privacy friendly, with no requirement for PII or Individual log level data
  • Not dependent on Cookies or Pixel data

Robyn Code Walkthrough Video

Please watch this walkthrough video to understand better how the code works