1
This is why you don't look into A/B test outcomes early and don't stop a test when it's "trending" well for you.
Back to archive
Thread
2 tweets