On average, how many ads do you create for each group? How often do you erase and create new ads?
Freq---
This approach is based on the Taguchi approach to multi-variate testing, also known as MVT.
You'll likely hear a lot more about this breakthrough ad testing approach in the future, because running these specially designed 18 ads is equivalent to running as many as over 4,300 separate A/B split tests.
Imagine how long it would take you to run over 4,300 separate A/B split tests!
With this scientifically proven approach to testing many variables at once, you only need a relatively small number of responses to achieve significant response increases.
Marty Foley
Why such blatant self promotion. Have you not read the TOS of this website? If you have read the TOS, then atleast you could have tried to use a different user name from the name you have used on your site. Just wait for the Mods to delete your entry.
If my post is otherwise considered self-promotional by a moderator, I apologize, stand corrected, and welcome editing by a moderator.
However, my recommendation of using multi-variate testing with Adwords still stands. Anyone who isn't leaving a lot of money on the table.
Marty Foley
P.S. Thanks for the suggestion to post under a different username to try to get around the TOS, but that's not something I'm comfortable with.
I like the idea of modifying ads every week or so and deleting the ones that don't perform.
What I am trying to do is avoid showing the same thing with each impression so that visitors won't get tired of it, but at the same time, I want them to start to recognize my name and associate it with the service I provide.
So far my groups with 2 ads are showing 50% of the clicks for each ad. The groups with more ads start to show a separation in ad performance, but not by much, maybe I'll give it a few more days before I edit and delete.
The vast majority of your visitors will not be repeat ones, so you needn't worry about them getting tired of your ads.
Fact is, advertisers tire of their own ads long before the public does, and often pull winning ads prematurely in favor of less effective ones.
Aside from "brand recognition," which is ambiguous at best, is one goal of your ads also to generate sales, leads, or some other sort of measurable conversion?
It's far better to determine the success of ads based on profitability, and maximize their return (as well as to keep testing new versions to beat your winning "control").
Marty Foley
I have to disagree a little on the name recognition though, in my market there are very few names that are well known and there is a lot of business from word of mouth. So although I am looking for leads, if I can get my name around enough, it may get me that lead with a big mouth that I'm looking for. ;o)
Since you can't measure "name recognition" from one ad to another, then it doesn't really matter which ad you run, as long as it has your "name" in it. :-)
