Apply Priority Formula Experiments strategy customization
Hi,
Regarding Lever Priority experiment, there is no check in the strategy (ApplyPriorityFormulaExperiments) to filter out those actions which do not have any Lever applied. It just applies P C V formula for the control group which is ideally not giving us the correct information as the actions will behave the same with P C L V formula. It would be good if this logic can be implemented within the product to take only those actions for the experiment which has Weight calculated.
Since the associated strategy is available in 8.8, can we modify it to suit our requirement? Please suggest.
Thanks,
Deepna
@DeepnaK1 i see your point, but I don't think it's as simple as excluding Weight = 1 from the control group, because if you do that you'd also need to exclude Weight = 1 from the test group, and we do not persist Weight in IH (there's pyWeight, but I don't think they are the same). This would mean that the test group would be larger than it should be by an unknown amount, and so the reporting would not be accurate.
If the percentage of actions not using levers is roughly the same for the test and control groups, then perhaps their effects will cancel one another out.