It looks like the optional "Name" field in A/B Exp...
# give-feedback
It looks like the optional "Name" field in A/B Experiment Features for variations is not persisted through the GBExperiment or GBExperimentResult objects in Kotlin. Are there any thoughts on adding this as a property within the Kotlin SDK and/or in other SDKs to assist with internal analysis? Thanks!
Yeah, we only expose the numeric index of the variations (
) in the SDKs. I can definitely see how names or other identifiers could be useful though. I think we can come up with a way to do this that's backwards compatible with existing SDKs.
That would be fantastic, for now I believe that we can build using the variation indices but having a separate identifier would be very useful.
As an expansion, there are also no current methods for variation identifiers for Forced Rules, A/B Experiments, and Percentage Rollouts within the same Feature outside of the raw value payload. This may be a bit larger than the aforementioned bit of feedback though.
do you have a use case in mind for that extra data?
Data team analysis of results based on different identifiers rather than having to list out the treatment itself. Much easier to analyze data with identifiers such as "v2_referrals_low_risk" rather than {"treatment_amount": 100, "quantity": 10, "platform": "email", ...}. For sole A/B experiments at least we can label groups based on variationId, but as soon as Forced Rules or Percentage Rollouts are used in tandem we cannot identify groupings outside of the payload, which often does not detail the group that is being targeted and instead just have the treatments. This is especially prevalent when we have small-medium groups that we would prefer to use forced targeting rules with but still want an identifier to go with the rule.