Prior Updates

Hey there. I would like to understand please how Growthbook updates its

priors

once the experiment starts. I understand Growthbook starts the experiment with Uninformative Priors. After the 1st day of the test, we have some new data. For 2nd day of test, is it considering as a prior: • the data from day 1? • Or the initial prior ? • Or both with weights? One extra question: Would it be possible to have a session with the DS Team to understand the calculation more deeply? Maybe with a practical example? (I'm currently looking at the open-source code but a session would be very helpful)

oh, hey! sorry we missed your message. The priors are always set using the uninformative priors and don't change over the course of the experiment. We are working on a way to let users define them and or have them set using past data, but for now they are uninformative.

Hey @helpful-application-7107 thanks for the reply 🙂 No worries!