Comments on: 3 ways to use an a/b split test (Billy's Optimization Guide Part 1) http://blogs.webtrends.com/optimization/2009/01/26/pro-a-b-split-test-method/ Just another Webtrends Blogs weblog Wed, 23 Dec 2009 20:59:34 -0800 http://wordpress.org/?v=2.8.6 hourly 1 By: ¹ SEO ROI Services: Pay For Professional Search Marketing. http://blogs.webtrends.com/optimization/2009/01/26/pro-a-b-split-test-method/comment-page-1/#comment-63 ¹ SEO ROI Services: Pay For Professional Search Marketing. Thu, 30 Apr 2009 06:56:29 +0000 http://testingblog.widemile.com/?p=298#comment-63 [...] 3 ways to use an a/b split test (Billy’s Optimization Guide Part 1) [...] [...] 3 ways to use an a/b split test (Billy’s Optimization Guide Part 1) [...]

]]> By: Billy http://blogs.webtrends.com/optimization/2009/01/26/pro-a-b-split-test-method/comment-page-1/#comment-62 Billy Fri, 30 Jan 2009 17:43:03 +0000 http://testingblog.widemile.com/?p=298#comment-62 Thanks for the link. A lot of technologies are out there to run a/b split tests. Be mindful of any imperfections in the way it executes/measures data though as you do your analysis. Even in that link, they mention that it isn't completely accurate. Just from a cursory glance though, this is recommending testing in a way that I am discouraging, meaning using a/b split tests for swapping out a headline or button color. These type of content swaps are more efficiently done in a multivariate test. If you're looking for a free solution, I recommend trying Google Website Optimizer rather than doing it through mixpanel, as it has true a/b split testing and full factorial multivariate testing. If you're looking for services and/or something more robust, you should checkout Widemile of course! Thanks for the link. A lot of technologies are out there to run a/b split tests. Be mindful of any imperfections in the way it executes/measures data though as you do your analysis. Even in that link, they mention that it isn’t completely accurate.

Just from a cursory glance though, this is recommending testing in a way that I am discouraging, meaning using a/b split tests for swapping out a headline or button color. These type of content swaps are more efficiently done in a multivariate test.

If you’re looking for a free solution, I recommend trying Google Website Optimizer rather than doing it through mixpanel, as it has true a/b split testing and full factorial multivariate testing. If you’re looking for services and/or something more robust, you should checkout Widemile of course!

]]> By: Richard http://blogs.webtrends.com/optimization/2009/01/26/pro-a-b-split-test-method/comment-page-1/#comment-61 Richard Fri, 30 Jan 2009 04:36:23 +0000 http://testingblog.widemile.com/?p=298#comment-61 Great intro article, we've been doing a little research ourselves We actually read this http://blog.mixpanel.com/2009/01/ab-testing-your-website-with-mixpanel/ for our company's ab testing, results initially seem pretty good. Great intro article, we’ve been doing a little research ourselves

We actually read this http://blog.mixpanel.com/2009/01/ab-testing-your-website-with-mixpanel/ for our company’s ab testing, results initially seem pretty good.

]]>