Yesterday I was reminded of a paradox in PPC management. As you expand your PPC efforts beyond a single campaign with a single ad group you too have encountered the paradox, though in varying degrees. This paradox is caused by the collision of two commonly accepted best practices: testing and tightly-themed ad groups.
Testing v. Granularization
Anyone worth their salt in PPC knows that you should always be testing. You test two to three versions of ad copy to get a better CTR. You test two different landing pages to improve conversion rate. You also know that these tests require a certain amount of clicks or conversions before they achieve validity.
Another best practice (and Google pushes this a lot) is to create small, tightly themed ad groups. By having only a handful of keywords in your ad group you can write extremely targeted ad copy and send traffic to a more targeted landing page. This practice improves the chain of relevancy (and your QS) while optimizing the user experience. The paradox, however, lies at the convergence of these concepts in practice.
The genesis for this post was a Powerpoint sent to me by one of our Google reps. The recommendations were solidly based on best practices like the two mentioned above, but the part that got me was when the slide said that by making more ad groups that were small and tightly themed I would have “easier campaign management.” On the surface that seems accurate; the ad groups have fewer keywords and the copy would be more specific, but what about tests?
To reach validity tests need to achieve a certain number of actions or responses. This particular client already has their campaign divided into nearly 100 campaigns with hundreds of ad groups. They spend a lot of money, but even with the existing account structure tests can take weeks or months to achieve validity. And now Google wants me to INCREASE the number of ad groups, further spreading out my traffic into more tests? This could potentially slow our test iterations to a virtual standstill.
Perhaps I’m tripping out about nothing, but I’m curious what all of you think. How do you balance testing and granularization?
PS I would especially appreciate someone from AdWords to chime in here, so consider this a challenge Google.