Skip to content

Comment on Experiments at Airbnbparent

Comments

Yes but in your A/B test you shouldn't be picking first half vs second half. Each visitor should be randomly assigned, so it should mitigate the problem you mentioned.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.