Datalad tool
Concurrent Test Collision Checker
Running several experiments at once is normally fine, and the reasons it is fine are not the ones people give. This works out which of your tests actually collide, how many users land in both, and which of the overlaps genuinely threaten the result rather than merely looking untidy.
1
What is running
Per cent relative
2
Every pair
Two tests only interfere if they overlap on all three of time, audience and surface. Overlapping on traffic alone is not interference, it is just two tests running.
3
[…] the effect each test is powered for A “what to do about it” verdict per overlap Open the checker Free · runs in your browser · no […]