-
Testing Responsive Search Ads copy effectively in Google Ads?
Posted by on July 2, 2021 at 3:10 pmI’m struggling to work out how to test copy effectively for RSA ads and would hugely appreciate some advice.
I’ve created one RSA ad per ad group which Google has optimised copy and is showing the most relevant permutations.
I would really appreciate thoughts on best practice for quickly, efficiently and methodically testing more ad copy variants.
Thanks,
D
Some further thoughts:
Google shows the impression stats for headlines, but not the actual conversion data.
Using Ad variations, I can only really change one headline at a time, and because of the number of headlines that’s unlikely to effect conversions.
One idea I had was draft and experiments, one 1 campaign with the original headlines, and another that replaces the 4 or so headlines with the current lowest impressions and A/B tested for conversions.
-
3 Replies
-
Generally speaking, we don’t test RSAs. RSAs are provided for Google to come up with the right combination. It’s almost always difficult to get 5 or 6 catchy headlines, let alone 15. So the whole concept of RSAs was invented for Google’s algorithm not for marketers to play with.
I’d personally recommend you A/B test some headlines and include the winning ones in your RSAs. You can pin 2 and let Google’s algorithm figure out the third one. Do the same for descriptions if you’re really keen on spending that much time and effort with ad copies.
-
Back in Sept 2020, the search terms report in Google Ads started only including terms that a *significant number of users* searched for, even if a term received a click.
Advertisers today don’t even have decent visibility into which search queries trigger their ads, *even if they incur a click or conversion.* Google hides a chunk of this information.
That shouldn’t even be legal. You pay for a click, you should know what triggered that ad.
Point is, all the testing you’re doing is for Google’s benefit. You won’t get complete data to know what’s effective no matter what you do.
And you’re paying for it.
-
You don’t get to directly test RSAs. You are supplying Google with the best “parts” so they can build the best ad. RSA performance can only be reviewed after the ad has at least 5000 impressions on top page 1. Google is just telling us to replace the “low” rating with something else.
I don’t see as an issue for not seeing conversion data, since you can always view it from the RSA itself. We can tell that Google do prioritize RSAs over ETAs as it usually receives the bulk of the traffic.
Testing ads will be less prevalent in the future as compared to testing “parts”.
Log in to reply.