Setup guide from Google Website Optimizer on the subject of testing old versus new.
I dont think this fits with current model at dsgi but provides useful articles in interpreting test results going forward.
<div>Google Website Optimiser</div><div>
Content:
Redesigning your website is a major undertaking, and can have a huge effect on
your conversions. If done by an outside agency, it can also cost thousands of dollars,
so you want to be sure that the new design is really working for you. Website Optimizer
can measure this.
Before pulling down your old site and replacing it with the new, run an experiment
pitting the two versions against each other. To do so:
- Upload your new site into a subdirectory (eg. www.example.com/newsite/). You may wish to prevent search engine crawling of this directory, through the use of a robots.txt file. You should also make sure that users are not able to navigate between your two sites (eg. new product detail pages linking to your old site's privacy policy page), or your experiment may fail.
- Create a new A/B experiment in Website Optimizer.
- In the A/B setup, enter your old site's URL in the Original page URL field. Your new site's URL goes into the Page variation URL field.
- In the Conversion page URL field, enter the conversion page for your old site. You'll identify your new site's conversion page later, by adding conversion code to that page.
- When adding the Website Optimizer code to your pages:
- Original page code: add this to your old site's landing page
- Variation pages code: add to your new site's landing page
- Conversion page code: add this to both your new and your old sites' conversion pages.
This experiment setup will serve your old site to one group of visitors, and your new site to the remaining visitors. With enough traffic, you’ll soon be able to measure the improvements you’ve wrought with your site’s redesign.</div>