If a user has already been bucketed into a variation, we keep those people in that variation. If we were to push users to different variations, it would ruin the statistical significance of an experiment.
When you change the distribution to 100%, only new users would be bucketed in the new experiment.
However, if you want to set a winning variation and set it to 100%, then you would see it reflected in the UI.
Comments
0 comments
Please sign in to leave a comment.