Integrating Google Website Optimizer with Drupal
Here it is, in plain English. The steps in order to integrate Google Website Optimizer with Drupal:
1. Download the Google Website Optimizer module from http://drupal. org/project/google_website_optimizer and install it just like a normal Drupal module.
2. Visit the following link, http://www.yourDrupalsite.com/admin/ settings/google_website_optimizer, or go to your admin screen and click on Administer | Site Configuration | Google Website Optimizer
3. Now you need to select or create a testing page and a success page on your site. The testing page is the page where you want to test some variations of your site. The success page is the page that your visitor will see when they have successfully clicked past the testing page. For example, maybe it's a landing page about widgets with a link to the contact us page. The key here is to make sure there is a link from the testing page to the conversions page.
4. Once you have testing and success pages selected or created, decide which element of the page you would like to test. For example, maybe you would like to test the headline of the page.
5. Now, go to the Google Website Optimizer site and click Create a new experiment.
6. Click Multivariate Experiment. Agree that you've created the required pages and click Create.
7. Name your experiment. Enter in the full URLs of the Test page and the Conversion page and click Continue.
8. Select You will install and validate the JavaScript tags and click Continue.
9. You should now see the testing JavaScript code that you will need to run the test.
10. Back at the module on your Drupal site, click on Add test.
11. Name the test on your Drupal site the same as the test on Google Website Optimizer. This is just for convenience—you could name it anything you want.
12. Under Testing Page, enter the Drupal path of the page that you're doing the tests on. Under Conversion Page, enter the Drupal path of the page that is your thank you page.
13. Copy and paste the code from Google into the corresponding fields on the module: Control Script, Tracking Script, and Conversion Script.
On your Drupal site, click Create script. You should see some text confirming Test created.
15. Now, you need to tag the sections that you want to test. Navigate to your testing page and put this code immediately before each section you want to test (note that you can copy and paste this code from Google's screen.):
< script >utmx_section("Insert your section name here")< / script >
And paste this text immediately after it:
< / noscript >
Note that in place of Insert your section name here text, you need to give a unique name to each testing section.
16. Back on Google Website Optimizer, click Validate pages. You will see that the Google Website Optimizer JavaScript tags are verified.
17. Still on Google Website Optimizer, click Continue and then click Continue again.
18. Now you will create the variations of your content that you will be testing. Click the Add new variation link. Name the new variation. Replace the default text with your variation code and click on Save.
19. Keep doing this until you have all the variations you want to test. Until you're familiar with this tool, you may want to only use a couple of variations on a single element. When you're done, click Save and Continue.
20. Now, click on Preview this experiment now. A separate window will open.
21. You can now select the different variations of your experiment. As you select each combination, you should see the variations that you've set up. Try out all the different combinations to make sure that none of your code combinations break your layout.
22. When you’re finished testing, close the testing window and click the Launch now button.
23. Your experiment is now running and you have a couple more options:
Total traffic sent through this experiment: It lets you decide if you want all the traffic that visits your testing page to use the testing code. Usually I'll set this to 100% for elements that I haven't tested before. But, if I have a pretty good conversion rate (2 to 3%) then I’ll only run half the traffic through it. It takes longer but I'm not risking losing many conversions by running the test.
Autodisable losing variations will turn off combinations that are performing poorly allowing you a true 'set it and forget it' experiment. There are three options: Conservative, Moderate, and Aggressive. Conservative will only turn off the variation when it's getting way less conversions than the original. Moderate will turn off the variation when it’s significantly lower. Aggressive will turn it off even when it's just a little lower. For most experiments, I like the Moderate setting.
You should begin to see statistics within three hours.