Hacker News Reader

Yeah exactly, it was some sort of seasonality effect that, or just plain randomness.

But you showed an example which was significant. That confuses your point somewhat.

It’s also not clear why you think “most” startups are not doing statistical significance tests on their a/b test results