If I am running a control vs treatment group experiment where I expose experiment to only 50% of the users and remaining 50 remain in control group, how do I evaluate results? Because the way its configured, experiment assignment only exists for users who were assigned any variant of an experiment. So where do I get the control group values to evaluate the treatment performance against it

Hi Muhammad, good morning 👋 To confirm my understanding here; is 50% of the traffic assigned the control variant, or is it that 50% of the traffic are not assigned the experiment at all, so will have the control rendered but not counted in your analysis? It may help to share a screenshot of your targeting to confirm the above, but if the users were assigned the control, they will be counted in the analysis, and you can cross-compare the variants under the results tab. However, If assignment only logs users who see the treatment and not those who see control, GrowthBook cannot compute a valid control group

I think I got you. I controlled the traffic percentage of experiment to be 50%, assuming the remaining 50% not exposed to experiment as control group.