Is Multivariate Testing SEO Friendy?

This is probably the most common question asked when discussing Multivariate testing.

Well, today the Google Website Optimizer test group confirmed what I have been telling my clients for a while…phew!

The Google Website Optimizer, and other high quality testing tools, such as those from Vertster, leave the original content visable in the HTML source code of your page at all times. As a result the original source code is visible to the search engine crawlers. This means.. “there should be no major impact on search engine rankings.” (Google)
The only change that will possibly occur is when you actually implement the ‘optimised’ changes back into your original page. And this has the same effect as any change to your content would potentially have.

So to sum up, use a good tool for your test and you should be fine.

2 Responses So Far

1. 1

   By Ophir Prusak on 31 March, 2007 at 2:28 am

   I’ve been using Google Optimizer since Nov 2006. I just posted some of my thoughts and tips on how to get the most out of it.

   Here’s the Link:
   Google Optimizer Tips

2. 2

   By Alasdair Bailey on 29 June, 2007 at 6:24 pm

   Hi Jason,

   We work with a number of clients in the UK and Europe with our own multivariate testing platform and have found no detriment to their SEO. I think Google et al must be used to companies serving different variations on home pages because split and A/B tests have been going on pretty much since the start of the web in some form or another.

   Others have been using it for dayparting or seasonal variation – an establisted site may find that different content gives different conversions by time of day or season.

   Kind regards,
   Alasdair