Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Robyn

facebookexperimental.github.io

  • Free plan
  • Windows
  • macOS
  • Linux
  • API
  • Self-hosted

What Robyn is

Robyn is a free, open-source Marketing Mix Modeling package from Meta Marketing Science. It applies machine-learning methods to estimate media-channel efficiency and effectiveness, adstock rates, and saturation curves, with a focus on granular datasets containing many independent variables. Robyn automates hyperparameter optimization using evolutionary algorithms and applies ridge regression to address multicollinearity and overfitting. It uses Prophet to decompose trends, seasonality, and holiday patterns, and can calibrate models against approaches such as geo-based tests, Facebook Lift, and MTA. A budget allocator uses a constrained nonlinear solver to propose reallocations intended to maximize outcomes, while model one-pagers support model comparisons. Robyn does not need personally identifiable or individual-level data and does not depend on cookies or pixel data. A stable R version is available on CRAN, with a development version on GitHub; the Python version is beta and may have translation issues. It is MIT licensed.

Who it suits

Robyn is intended for advertisers with granular data and many independent variables, especially digital and direct-response advertisers. It may suit teams seeking marketing mix analysis and budget planning without individual-level data.

Good to know, and what to watch

  • Models media efficiency, effectiveness, and saturation
  • Automates hyperparameter optimization
  • Can calibrate against geo experiments and Facebook Lift
  • Budget allocator proposes budget reallocations
  • Does not require individual-level data
  • Python version is beta and may have translation issues
  • Python API requires the Robyn R package installed first
  • Paid media variables and spend vectors must match in length and order

The verdict

Robyn combines media modeling, calibration, and budget allocation for granular marketing datasets. The R version is stable, while the Python version is beta and carries stated limitations.

What the maker publishes

Adstock options
Robyn offers geometric, Weibull CDF, and Weibull PDF adstock transformations. source
Availability
Robyn has a stable R version on CRAN and a development version on GitHub; the maker also documents a Python version marked beta. source
Budget allocation
Its budget allocator uses a gradient-based constrained nonlinear solver to maximize outcomes by reallocating budgets. source
Calibration
Robyn can calibrate models against ground-truth methodologies including geo-based tests, Facebook Lift, and MTA. source
Input requirement
Paid media variables and paid media spend vectors must have the same length and media order. source
Integrations
Robyn uses Nevergrad for optimization, Prophet for trend and seasonality decomposition, and glmnet for ridge regression fitting. source
Intended users
The package is built for granular datasets with many independent variables and is described as especially suitable for digital and direct-response advertisers with rich data sources. source
License
The repository states that Robyn is MIT licensed. source
Model calibration
Robyn calibrates marketing mix models using causal experiments such as randomized controlled trials and geo experiments. source
Model comparisons
Robyn generates model one-pagers to support intuitive model comparisons. source
Modeling
Robyn uses machine-learning techniques to estimate media channel efficiency and effectiveness, adstock rates, and saturation curves. source
Optimization
Robyn automates hyperparameter optimization with evolutionary algorithms from Nevergrad and uses ridge regression to regularize multicollinearity and prevent overfitting. source
Optimization
Robyn uses a multi-objective evolutionary algorithm for hyperparameter optimization. source
Privacy
The maker describes Robyn as privacy friendly, requiring no PII or individual-level log data and not depending on cookies or pixel data. source
Privacy design
Robyn does not require personally identifiable information or individual-level data and does not depend on cookies or pixel data. source
Product
Robyn is an experimental, AI/ML-powered, open-source Marketing Mix Modeling package from Meta Marketing Science. source
Product type
Robyn is an experimental, AI/ML-powered, open-source Marketing Mix Modeling package from Meta Marketing Science. source
Python API limitation
The beta Python API requires the Robyn R package to be installed first. source
Python availability
The Python version is a beta rewrite of Robyn's R package and may have translation issues. source
Python limitation
The repository says the Python version is an LLM-translated beta and may encounter bugs. source
R availability
Robyn has a stable version on CRAN and a development version on GitHub. source
Support
The maker points users to a public Robyn MMM Users Facebook Group and GitHub issues. source
Support
Users can join the public Robyn MMM Users Facebook Group, raise GitHub issues, and take Meta's official Robyn blueprint course. source
Target users
Robyn is built for granular datasets with many independent variables and is especially suitable for digital and direct-response advertisers with rich data sources. source

Checked against the maker's pages on 30 Sept 2026.