Ah, Google. They have so many options for advertising, even just underneath the umbrella of “search ads.” Obviously, you have your basic search ads, but what about the other options that take search ads and pump them up a bit with new features?
Well, today we’re here to talk about RSAs: Responsive Search Ads.
A quick look at what they are and how they can spice things up
Basically, RSAs are a search ad variant that combines Google’s machine learning and your own copy. This works by letting you plug in multiple titles and descriptions. From there, Google will rotate combinations of three headlines max of 15) and two descriptions (max of 4) depending on their relevance to a given search query.
Example of Headline Options for an RSA (see final ad below)
Over time, Google learns which titles and descriptions work best for which searches, prioritizing those combinations.
Example of an RSA (using headline options above)
The general gist of RSAs’ benefits are probably pretty intuitive from that description alone; for starters, they allow you to easily test multiple title and description variants within a single ad group. Because RSAs are constantly rotating copy combinations, you can over time develop a clear picture of which combinations, themes, etc. are performing most effectively. And by allowing advertisers to compete in more ad auctions, RSAs naturally increase exposure.
What do we recommend?
In our experience, many types of dynamic ads have their ups and downs — this includes RSAs. They’re not going to magically rocket your campaign’s ROAS and effectiveness up a million percent, but they are useful when used right.
Some perform exceptionally well over more basic expanded text ads, but in other cases they can just end up being more expensive without much to show for it.
The first rule to making sure your RSAs don’t fall into the latter category is to adhere to the best practice of running three ad variants per group. When we use RSAs at Fujisan, we’ll usually run one alongside two expanded text ads; we’ve found that to be a pretty reliable starting point.
Now that we’ve done enough downplaying of RSAs, we do of course have to talk about their benefits. The greatest of these is that with RSAs you have the ability to inherently test your ads’ performances and through extension determine the best ad content and variations. Obviously, you can (and should!) be testing all your ads and other aspects of online marketing, but the nature of how RSAs rotate content and the insights you gain from that means that just by running them, you’re effectively running tests all the time.
Of course, part of understanding these tests is not overwhelming the ability to interpret your data by starting off with a bunch of new variables. For this reason, we highly recommend incorporating headlines from expanded ads into your responsive ads. That way, you’ll have some controlled variables that give the resulting data some context, rather than relying solely on new copy that would then give you minimal understanding as to its performance relative to your baseline.
Of course, you’ll have to wait until you’ve accrued a statistically significant amount of data — we recommend 10,000 impressions as a safe place to start, although you can always wait for more. At Fujisan, once data’s at a place we deem enough to work with, we like to toss it into a pivot table. In our experience, that’s the ideal medium for breaking down RSA performance.
And that’s that!
Once you have this foundation down, you can take whichever path you want with RSAs. It all comes down to the scientific approach of determining how they’re best utilized for your business, plus having the creativity to come up with new copy content that gives your campaign more potential pathways to succeed.
