Hacker News new | comments | show | ask | jobs | submit login

Yup, this is exactly why I wrote: You should NOT follow me on Twitter http://visualwebsiteoptimizer.com/split-testing-blog/you-sho...

A/B Testing case studies are good for two reasons:

a) Help create curiosity on what the heck this A/B thing is (good for my product) b) Gives you ideas to test on your website

But it is very dangerous to implement results without testing for your case. Very, very dangerous.

Dangerous? How about a best practice? If you don't have time or inclination to test, the first strategy you should use is one that worked for someone else.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact