I’ve just belatedly picked up this post. Some very interesting and insightful points on MVT and cloaking. I work for an MVT practitioner- GlobalMaxer- and we have our own javascript based technology not affiliated or associated with GWO.

We commonly get asked from client side SEO teams what affect MVT has on rankings and SEO strategy, and usually explain that e.g. Google can’t read javascript, the variations are noindexed etc etc. However, there is a wider debate about the ethics of the really advanced knowledge acquired by MVT.

Your post throws up some interesting questions about where the ethical line is:

(1) when we do AB testing, quite often clients submit totally different designs for comparison. In no way are these “in the spirit of the original page”. However, as some of them convert better, they’re clearly providing a better landing experience for the user. In some cases these have seasonal or tactical offers, and you wouldn’t want these to be permanent changes to the design.

(2) What is the overall purpose of MVT/Landing page testing? Is it to find a new page design that performs better, in order to make this a permanent change? Or is it rather to take a fluid and dynamic approach to web design in a process of continual refinement and optimisation? How can this strategy co-habit with your SEO strategy, based on a defined set of keywords, metadata and link anchor text? Do MVT practitioners only segment tests via Direct Load visitors into tests?

I think luckily for MVT practitioners, this is currently rather a Google blind spot. But as the market evolves and the engines get better at reading javascript, we will need to engage more with them on getting better guidelines on the ethics of MVT.

This is a review of Jim Stone’s Split Test Accelerator. I’ve used Jim’s software for over a year to help improve an ongoing affiliate marketing campaign. Using his software I was able to increase my conversion rate (click-thrus) from 30% to over …