Hi everyone! Lately I've been looking at Growthboo...
# announcements
l
Hi everyone! Lately I've been looking at Growthbook, and I'm really interested in doing a POC with it within our company. One question that I have is regarding the way Growthbook needs to be configured against its datasource. Am I correct that the backend only really requires read-only rights to the datasource (e.g. Snowflake in our case), as long as we have the necessary data to perform the experiments and pageviews query?
f
Hi! Yes, we only need readonly access
and the queries we run only return aggregate data (number of users, mean, std dev for each variation)
l
I see! Thank you for the quick answer! I also see that there's an
experiment_id
and
variation_id
required in the experiments query. Isn't that something that Growthbook would generate? So then we would need some way to store these two variables in our datawarehouse, correct?
f
Yes, typically you would have a table in your db that maps a user to which experiment/variation they saw.
So at the point you run the experiment, you would need to log that and store it in Snowflake. You can do this with Segment, Snowplow, Fivetran, Jitsu, or your own custom event tracking system.
🙌 1
l
That sounds reasonable.
I will discuss with our data people. Thanks bunches! Will let you know how it goes 🙂
👍 1