In paid search you are always, ALWAYS testing. Ad copy is one of the key items that you are always testing in constant chase of increasing click through rate and conversions. As a basic rule of good PPC management, multiple ads should be used within ad groups. You may test how one ad uses a price, where to use the keyword variation in your ad, perhaps how your location plays, the use of an exclamation point, or landing pages.
For this, you need multiple ads in one ad group to be served evenly, and then measure both click through rate, and possibly conversions, depending on the scope of the campaign. Serve rates should be very close to each other.
In the “old days”, you had three options for Ad Rotation Settings, rotate evenly, optimize for conversions, or optimize for clicks. The optimize function was driven back into a formula that dictated if you tested two ads, whatever ad happened to get a few early clicks would get served more. So you ended up with an ad that may not be the best one getting the served the most. For Google, it is designed to automatically be serving the ads that get the most click, <cough, make them the most revenue>. The message to advertisers was – set up the campaign, and let us pick the ad winner for you based on clicks. If you have multiple ads per ad group, the optimize function meant that the serve rates are going to have a huge disparity, thus making it impossible to test performance.
The optimize setting was always on as a default. We always unchecked this and checked rotate evenly. This way you could set up ad groups that would properly test the click through rate of multiple ad copy to measure performance. The only true way to do this is to have the ads be served up evenly.
Back in May Google AdWords released an update which essentially eliminated the ability to always have ads rotate evenly. In what was a clear monetization play, the rotate ads option was removed. All AdWords advertisers could get was an option to have the ads rotate evenly for 30 days, then it they would switch back into an optimized mode. It essentially took the ability for us to properly test ads over time away. The PPC community revolted. Even a petition was created at change.org, which I happily signed.
Surprisingly, Google listened (kind of). They rolled out an update that gave advertisers the option to have ads rotate evenly for 90 days, and then they would switch into an optimized mode. This was still not an optimal change, but it was viewed as a good compromise. Then in June, they created an option for advertisers to fill out a form requesting that ads in their account be served evenly at all times. However as you can imagine, this form was hard to find, and only 1% of account utilized it. In other words, no one knew about it or saw it.
From my standpoint, we rarely have ads that last for 90 days without a refinement, so this should be satisfactory from a testing standpoint. However for typical end users it was still probably very confusing. To compound matters, for a short time Google was showing a dialogue box that “warned” users that selecting this option would hurt campaigns as the best ads wouldn’t show. In my opinion, this was overly skewed towards what they wanted you to do. I believe they have since stopped showing this box, as I couldn’t recreate it for an image for this post. The 90 day option appeared to be a good compromise for the us PPCers that were at Google’s gate with torches, although many were still not happy.
Most recently in early October, the almighty Google relented. The latest release gives advertisers 4 options for ads to be served. Optimize for clicks, optimize for conversions, rotate evenly for 90 days then optimization will kick in, and rotate indefinitely. YES, that is the one you want for testing!
What is most interesting here is that you can see in the ! box below the options, Google still does not want you to use the rotate option. Do not be swayed. For proper ad testing, your serve rates must be as equal as possible. How can we test ad performance if one ad gets served 90% versus 10%. It obviously will generate a skewed result.
I understand Google’s move to maximize revenue through ad copy. However, in the name of quality and relevance I don’t understand why there is such a push from Google to use the auto optimize option. Why isn’t the rotate ads evenly option pushed? Why wouldn’t Google want the best ad copy possible served to searchers. Don’t we want the most relevant and qualified ads served in order to generate a quality user experience with AdWords? Ad testing is one of the best methods to generate optimized campaigns.
Be sure to check the Rotate Ads Indefinitely option when creating your AdWords campaign. Then create at least two ads per ad groups. Launch, measure, refine, and repeat! Good luck!