It's even worst than that. Most of the time the validity of the t-test that they are running is questionable. They are technically running an online t-test and as soon as they find significance they stop. This is fundamentally wrong, and not conclusive at all.
A few years ago when I was still working there, involved in the experimentation tooling among other things, we largely excised that particular behavior. What are you basing your assertion on?
Comments
It's even worst than that. Most of the time the validity of the t-test that they are running is questionable. They are technically running an online t-test and as soon as they find significance they stop. This is fundamentally wrong, and not conclusive at all.
A few years ago when I was still working there, involved in the experimentation tooling among other things, we largely excised that particular behavior. What are you basing your assertion on?
Cheating on a hypothesis test... that's terrible.