views:

18

answers:

0

Hey all. I'm attempting to run a multi-variant test on a page, and just have a quick question on validation. Due to the size of the site, and the volume of visitors, we are going to segment the test to only a portion of our users (though the test will be run on the same page that ALL visitors go to). Obviously there is no way to set this in GWO, so we are hoping to employ some tricks. Here is what I'm thinking.

Say we only want to show this test to visitors who like cake (let's pretend we have a magic formula to determine which users like cake, and which don't). I was thinking of employing some if / then conditionals. Reason being, as stated earlier, we only want to show the experiment to a handful of users, and thus only track conversions from a handful of users (we are showing the same test page / conversion page to all users. This becomes an issue since as we all know, Google does not like cake, and thus their tracker will not see and validate our scripts. So two issues here:

  1. How would I get GWO to validate a script like this? Can I simply use the offline validation option, and upload a static page that has the same tags in it that I will use on the live site, so that Google can see everything is in place and tagged properly?

  2. Can I use some simple if / then statements to segment who I show the test to. As I said before, I'm testing (in this example) whether somebody likes cake - If you don't like cake, you won't see the experiment - you will simply see the normal site. If you do like cake, the experiment will be run, and you will see variations.

Obviously with this in place, I would need to do checks across the board. Would something on the page given the following logic work? Or what kind of issues can you see me running into?

On the test page:

if (userlikescake) {
     output javascript control script (and at end of file do tracking script)
}

then employ similar logic on the page sections, following something like:

if (userlikescake) {
    output javascript page sections scripts
} else {
    output standard website section
}

And then on the conversion page, output something similar to what I employed on the test page to ensure the conversion page only counts a conversion if the visitor was shown the experiment script.

Sorry for the long post, and multiple question - Could really use some help. Please let me know if any more information is required.