| 5:46 pm on Dec 29, 2009 (gmt 0)|
I could be wrong about this but... assuming you're using Google Website Optimizer... I can't see why Google would see this as a problem, especially considering they are giving you a tool to do the A/B testing.
| 5:57 pm on Dec 29, 2009 (gmt 0)|
Thanks, chadhenry. I am not using Google Website Optimizer. I use proprietary tools for optimization.
| 6:08 pm on Dec 29, 2009 (gmt 0)|
Okay, well no matter what tool you use I can't see it being a problem. If you still feel it might be a problem you could always add nofollow to one or both of the pages until you are done testing. You could also add one or both pages to your robots.txt file and disallow the spiders.
| 7:43 pm on Dec 29, 2009 (gmt 0)|
Sorry, but can you explain your basis for saying no problem? Landing page is a major factor in QS, so I don't understand how hiding the page from a robot wouldn't kill QS.
| 8:51 pm on Dec 29, 2009 (gmt 0)|
Per Google, disallowing AdBot is not recommended.
| 7:09 am on Dec 31, 2009 (gmt 0)|
It really depends on what you are doing. If the pages are very different and look like you are trying to game the system then yes. If The pages are not that different then no. It can hurt your QS if one page is better than the other but it would not be a penalty it would just be how the system works.
| 2:59 pm on Dec 31, 2009 (gmt 0)|
Why not use the Website Optimizer?
|That is, destination URL is a page that randomly redirects to one of two landing pages where I measure conversion. |
Or why not create two indentical ads and point one ad to landingpage 1 and the other ad to landingpage 2?
| 8:05 pm on Dec 31, 2009 (gmt 0)|
|Why not use the Website Optimizer? |
As a general rule, I keep conversion/ROI data away from companies that can use it to drive up my costs...and that definitely includes Google
|Or why not create two indentical ads and point one ad to landingpage 1 and the other ad to landingpage 2? |
My landing pages are set at the keyword level, not the ad level, for tracking (querystring) reasons. However, I suppose I could set up an adgroup just for testing, with a few select keywords, and do the two-ad scheme you described. Good idea. Thanks!