furthermore im wondering about the pageviews. does this truly need to be pageviews? the documentation states it is used to predict how long an experiment will take, but wouldn’t we need to have data on the target metrics to do so? (rather than pageviews)

it doesn't technically need to be page views, the naming here isn't ideal

but how should we handle it if we have multiple metrics we use in testing then for this query? the example is hit based (counting rows to get number of pageviews), but can we also retrieve the actual metrics here or how should I go about this?

it's just for estimation, and not required at all. Do you have ideas that have multiple goal metrics?

ok.. no we will have 1 target metric (sometimes calculated metric based on multiple other metrics) It’s probably because I’m in the mindset of what I’m currently doing in Jupyter Notebook. For the MDE and run time calculation I’m looking at the actual test metric.. 😉

sounds good!

I guess I need to create the test in growthbook, and then populate the data with it’s test ID ?

if you have the main data source connected, you can click on import past experiments, and we make that easy if you already have the data there

cool!