RSAs, or Responsive Search Ads, are the newest ad type by Google (still in beta) to help improve relevancy and ad performance. There is hesitancy for implementation due to some lack of reporting and ability to understand what variation and how they improve performance. So, today, I’d like to share some specific ways to start testing RSAs when there is limited visibility into variants and their results.
In short, the biggest tip is: Use Pins
There are a lot of cases where we deal with brands and brand messaging that test multiple brand phrases in headlines. If we lend all possible duplicate phrasings to automation, they could show up multiple times in one ad. So, for this example, we will use PPC Hero to demonstrate the multiple headline variations that could be used in a brand headline:
- PPC Hero
- PPC Hero by Hanapin Marketing
- PPC Hero Blog
All of these say the same thing, but in typical ad testing, we’d have 3 ads using these different variations separately. With RSAs, they all are compiled together under “headline”.
Plus, we would also have other headline options to test, that would typically go in headline 2:
- PPC Strategies, Tips & News
- Read Trending Articles on PPC
- PPC Advertising Articles
. . .And so forth.
Again, with typical ad testing, we would have multiple ads using variations, but with RSAs, we allow the system to automate this process.
Due to that, in this scenario, we could have all of these headlines show in any order or variation, with no detailed insight into how they showed up on a search result. So, in theory, this could mean an ad shows 2 brand variants in both headline 1 and 2, causing this:
PPC Hero | PPC Hero By Hanapin
Or for our typical ‘headline 2’ messaging, they could show, like the brand:
PPC Strategies, Tips, & News | PPC Advertising Blog
Because all of the options are slotted as headlines with no pins, this gave the idea to use pins in a much more strategically detailed way, creating specific variations of ads we approve.
This uses the power of the RSA to create ad variations through automation but helps the system by selecting pin “themes” we have outlined. We can then re-order the pins by creating another RSAs for deeper ad testing to find out where the brand headline or description headline best resonate. Here is an example of the RSA strategy.
In the diagram example, we use Headline 1 for the brand, as we want it to be the most prevalent. Then, in headline 2, we use our sub-header language. Since there is a headline 3, this can be used for call-to-action tests. (Note: be careful of call-to-action testing if description lines also use calls to action).
As described, we’ve broken out each theme, as to not create any overlap or duplication that is sure to happen during automated testing.
Again, this can then be switched around to test brand, call-to-action, etc. in different headline locations. The same strategy can also be used with description lines 1 and 2, as well.
The moral of this strategy story is; don’t be afraid of testing new strategies and the newest platform integrations. Just be the PPC strategist, you are, and think through the opportunities it creates and how to use new features to your advantage.