Recent searches
Search options
A/B testing is often misunderstood by bad product teams. You can’t A/B test your way to a good product or find product market fit.
It’s most useful as an optimization technique to drive more usage or nudge the user behavior you want after it’s clear users like using your product
@carnage4life While it’s literally true that there are some things you truly can’t a/b test (marketing spend in different media markets) and econometric approaches like causal inference are more well suited…
…I’d also argue there are a lot of test practitioners who lack the imagination on how to build a a/b-test-based learning agenda that helps a PM learn what the customer actually wants beyond the nudges.
(It’s fine for nudging, too. My argument is simply that if we can develop advanced medicine with a/b testing, I think it’s fair to suggest that the tools are only as good as good as the user in some cases — to your point about bad product teams.)