Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Killer idea: play the split testing game with your boss and colleagues (visualwebsiteoptimizer.com)
15 points by paraschopra on Dec 16, 2010 | hide | past | favorite | 11 comments


I thought this article was going to be about acting differently with your boss and colleagues to see which gets you the most money/respect/whatever you're chasing.

Needless to say, I was very disappointed.


That's a good idea too but unless you have a twin brother working at the same office I don't see how it is going to be easy.


Does anyone have experience with Visual Website Optimizer and its influence on loading time and user experience? With all these Javascript A/B testing services, I'm worried about content flashes and long loading times. E.g. for the JS redirect, the browser would first load the page, THEN execute a redirect, then load another full page. Or the content might change after DOM completion, delaying rendering, or after rendering, causing flashes.


I have seen one of our large customers (can't name them) do this and it is indeed very interesting. What they do is this: they name one variation of the test each for all the team members in the design and analytics team and then make them live in a split test (irrespective of whether they all agree to specific changes/ideas or not). I observed that they consistently get great results for their split tests.

EDIT: typos and clarity


So, this is multivariate testing, not A/B testing, right?


No, this is A/B testing. Why do you think it is A/B testing?


Ok, maybe I don't understand the difference. Aren't you suggesting testing many different versions of a page, rather than just two?


Yes, in split testing you can test as many different versions as you like. Multivariate testing is different. You can read about the difference in the recent post I wrote: http://visualwebsiteoptimizer.com/split-testing-blog/differe...


From your post:

In A/B testing you split traffic amongst two or more completely different versions of a webpage (landing page, home page, etc.) The variations of your original page can differ in any manner. You can either just change the headline; or you can even change entire design, layout, offer and what not in the variations.

In multivariate test, you identify a few key areas/sections of a page and then create variations for those sections specifically (as opposed to creating variations of whole page in an A/B split test).

I thought multivariate testing was testing lots of different variables (as the name suggests) at once, while A/B testing is just testing two variations of one variable at a time.

Edited after eru's comment below: Multivariate testing is an extreme form of A/B testing.


I guess you should go with the general definition of multivariate testing (http://en.wikipedia.org/wiki/Multivariate_testing) as it is used in statistics.


Yes, I got it mixed up.

In internet marketing, multivariate testing is a process by which more than one component of a website may be tested in a live environment. It can be thought of in simple terms as numerous A/B tests performed on one page at the same time. A/B tests are usually performed to determine the better of two content variations, multivariate testing can theoretically test the effectiveness of limitless combinations.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: