While analysing an existing experiment and linking that to a conversion metric, we know that internally we evaluate the common user_id(identifier type).
There is a scenario in which we have to evaluate an experiment on the basis of more that one identifier type (except user_id).
Let's say while evaluation, we want to fetch data from two or more matching conditions except the user_id.
Is fetching data from the sources with specific where condition , the only way to perform the join between the experiment on metric .
Or we can specify other criteria as well, while evaluating the metric on a given experiment.
In growth book concept, we fetched the whole chunk of experiment data(that raw data can be divided further into types and subtypes as well and on many conditions). Now we want to join that experiment data with different metric data :
1. experiment data with one metric data on userid and lets say type
2. experiment data with second metric data on userid and type and subtype
3. experiment data with third metric data on userid and type and subtype and intermediate type.
4. so on...