There are many ways to do this, including many tools for analytics as well as managing the various options you to compare for effectiveness.
For analytics if you want to keep your data more private and/or internal (especially less sensitive to client-side/remote tampering and false or bad data), then there is already functionality in OFBiz to track all of the data you need and you'll just have to write reports to get it out according to your tastes. Check out the TrackingCode, Visit, Visitor, TrackingCodeOrder, and other related entities.
For managing the multiple options unless you want to build something that automatically pairs options with tracking codes (or tracking URLs for Google Intelligence Agency or the like), I'd recommend keeping it simple and code totally separate screens in OFBiz for the different options (reduce conflicts as both evolve separately, etc), and then do the randomizing either in a controller request event if you want it hidden from the user, or in the link generation if you want the user to be able to tell by the URL which variation they are using.
-David
On Jul 28, 2010, at 1:11 PM, Ylan Segal wrote:
> Hi everyone.
>
> Is anyone using A/B testing for ofbiz-driven ecommerce sites? What about multi-variate testing?
>
> By A/B testing I mean testing of different layouts and/or graphic design of a given page to see which design drives better conversions:
>
>
https://secure.wikimedia.org/wikipedia/en/wiki/A/B_testing>
> Google Website Optimizer allows to conduct this kinds of tests, but I'm not sure it's the best fit given the way ofbiz constructs html pages from templates.
>
> Thanks,
>
> --
> Ylan Segal
>
>
>
Locked
