OutFront Forums
     Home    Register     Search      Help      Login    

Follow Us
On Facebook
On Twitter
RSS
Via Email

Recent Posts
Todays Posts
Most Active posts
Posts since last visit
My Recent Posts
Mark posts read

Sponsors
Shopping Cart Software
Ecommerce software integrated into Frontpage, Dreamweaver and Golive templates. No monthly fees and available in ASP and PHP versions.
Website Templates
We also have a wide selection of Dreamweaver, Expression Web and Frontpage templates as well as webmaster tools and CSS layouts.
Frontpage website templates
Creative Website Templates for FrontPage, Dreamweaver, Flash, SwishMax

 

A/B - Multivariate testing

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
Printable Version 

All Forums >> Web Development >> Search Engine Optimization and Web Business >> A/B - Multivariate testing
Page: [1]
 
Scotty

 

Posts: 223
Status: offline

 
A/B - Multivariate testing - 1/25/2009 13:12:51   
We were looking at trying the Google A/B testing, however it appears to be a major hassle to set up when you have a hosted shopping cart.
I've been advised that an easy work around is to not have the test dependent on the final sale (e.g. going to the 'Thankyou' page), rather base it either on time on page or, in the case of the home page, clicking to go deeper into the site (rather than bouncing).
Anyone doing any testing? Ever base it on time on site? Any good/fairly easy programs that are preferred over googles'?
Regards, Scotty
Mojo

 

Posts: 2472
From: Chicago
Status: offline

 
RE: A/B - Multivariate testing - 2/10/2009 15:19:06   
I do extensive split testing across many different industries. Unfortunately, I don't use Google's split test product so I can't help much there, but I can make some recommendations.

You can't split test based upon time on the site. The biggest problem with the time on site stat is the inaccuracy of the data. Especially when doing something so statistically driven such as split testing - you must have accurate data. There is no way to tell if someone went to the bathroom, took out the trash, took a phone call, looked at another Web site with a different browser or was otherwise distracted somehow from the site at hand. In this case, that 4 minute site visit may have only been 2 minutes. It will be impossible to accurately have a split test.

I have tested many other actions besides sales. A commonly requested test is which home page design performs better or which form is the better converting. These things are fairly easy to test. You must have your A and B pages show for ever other visitor. A test is less accurate is you show the control page on Monday and then the B (test) page on Tuesday. That data will likely not hold up well over time.

Joe

_____________________________

Split Testing
GPS Vehicle Tracking
Chicago Marketing

(in reply to Scotty)
Page:   [1]

All Forums >> Web Development >> Search Engine Optimization and Web Business >> A/B - Multivariate testing
Page: [1]
Jump to: 1





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts