Looking around what's out there I have a question. First, when you guys split test do you just point your ads at different urls...like one google ad points at page1.html and another points to page2.html? I've seen the programs out there and some make you point your url to them but is this bad, do you think customers would care/notice a different url than yours? This is done a lot in email marketing for tracking and I was curious if it's done the same here.

If your sales letter is compelling enough, I don't think they will care which page they land on. If they start fiddling with their address bar instead of reading your message, you've got bigger problems than worrying about the name of the page they land on.

That being said, it could matter, but more than likely there are are other things that will matter much more that you should focus on, such as your headline and lead.

Some testing software sends you to different pages, and some return the page in one page, usually the index page. When you use multivariate testing, it's usually returned on one page (again, usually the index page), because you're not going to have dozens of different pages with every possible combination of all variables being tested.

But I've tested both ways (started out with split testing on different pages and later moved to multivariate testing on a single page), and I really haven't noticed a difference, certainly nothing worth worrying about.

Perhaps others have different results.

John

I think I will try the whole page split test first. How often do you split test for, do you leave the pages up there for a month or so and then maybe adjust and try again?

Depends on how much relevant traffic your getting. You want enough to make a statistically valid decision.

Perry Marshall's free split testing tool can help you to determine a confidence factor. Although it was designed for Adwords PPC meterics, it will easily apply toward split testing as well.

:::SplitTester.com:::

Simply enter your unique page visitors where it says "number of clicks" and your conversion percentage where it says CTR. The tool will tell you whether you have enough information yet to determine how likely your test subjects are to have different long-term results. A confidence factor of 95% or higher is where I like to be. Anything less than 90% means wait a little longer and try again once you have more data.

John

The different page URLs don't seem to make any difference.

In many cases visitors to your sales page are coming through redirecting affilate links anyway.

For more information about split testing services check these two sites...
http://www.goto-pro.com

and...
100% guaranteed copywriting and split testing service for sales copy

Kindest regards,
Andrew Cavanagh