The webpage discusses a Bayesian approach to experimentation, focusing on interpreting and extrapolating experimental results, mainly in a tech environment aiming to maximize user retention. It addresses challenges like the inference problem, the extrapolation problem, the explore-exploit problem, and a culture problem within tech companies around misuse of experiments. The author suggests providing decision-makers with benchmark statistics to help them estimate true effects of different policies, and discusses a model of experimentation dealing with observed and true effects along with the noise in experiments 1 .
Comments
The webpage discusses a Bayesian approach to experimentation, focusing on interpreting and extrapolating experimental results, mainly in a tech environment aiming to maximize user retention. It addresses challenges like the inference problem, the extrapolation problem, the explore-exploit problem, and a culture problem within tech companies around misuse of experiments. The author suggests providing decision-makers with benchmark statistics to help them estimate true effects of different policies, and discusses a model of experimentation dealing with observed and true effects along with the noise in experiments 1 .