Exclusion Groups (Mutually Exclusive Experiments)

What this is for

and when to use mutually exclusive experiments

If experiments could overlap technically or in hypothesis, they would be mutually exclusive.
Let experiments run in parallel without getting in each other's way and don't wait for one to finish.

How to create a group

You can create a new exclusion group for mutually exclusive experiments from the menu.
Under Experiments you find Exclusion Groups.
In a group, add all the experiments that should not run in parallel.
If you want to give one experiment a higher priority than the others, you can distribute the 100% among the experiments in such a way that there is a corresponding probability that a particular experiment will "win" more often and thus be the only one to run from the group.

How it works

All active experiments (contained in the snippet) are used as the basis for the decision.
For the experiment it is checked if the targeting overlaps (regarding Audience and Pages) with other experiments and thus experiments would potentially run in parallel.
If this is the case, an experiment is randomly selected based on the set weighting, which should run for the user. For all other experiments in this group the user is excluded.

If the user has already been a participant bucketed in one of the experiments contained in a group, the user will also be excluded from all other experiments.