September 4, 2019
How You Should Actually Be Testing and Evaluating Google’s Responsive Search Ads
With an ever-increasing emphasis on automation, Google seems to be updating ad types with more regularity than ever. Most recently, Google rolled out the Responsive Search Ad (RSA) unit, allowing advertisers to include up to 15 headlines and 4 descriptions. See our blog post from earlier this month on how to properly test RSAs in your campaigns.
However, it seems like just yesterday that Google rolled out the new Expanded Text Ad (ETA) unit, encouraging advertisers to include a third headline and second description. As Google fully transitioned to these ETAs, many advertisers maintained their tried-and-true ad copy and opted not to include the new fields. Understandable as that may be, the inclusion of up to 120 additional characters in a single text ad could improve engagement and CTR as ads take up more real estate. Additionally, when used properly, additional characters should, in theory, allow for improved ad relevance, which can significantly impact quality score and average CPC.
This begs the question – what are the actual benefits to fully expanding your search ads? Is bigger better? We performed a test to find out.
A four-week A/B test was established in two separate ecommerce accounts. Within each account, we tested the inclusion of Headline 3 and Description 2 in two campaigns: the primary branded search campaign, and a high-volume non-brand search campaign. All the campaigns had three or four active search ads in each ad group, each only including Headlines 1 and 2 and a single description field.
After duplicating the existing ad units, the same Headline 3 and Description 2 were applied across all the new units. After switching each campaign’s ad serving setting to “rotate indefinitely,” we were ready to A/B test our new fully expanded ads against the existing ad units.
In both accounts, the brand campaign saw the most noteworthy improvement in performance from the fully expanded test units. At relatively equal volume, the expanded test units saw an impressive 32% improvement in average CPC alongside a 7% improvement in CTR and a 26% improvement CVR. A jump in CTR from 32% to 34% may not be all that exceptional, but at the upper reaches of what CTR you’d expect an ad to bring in, that’s actually a really meaningful jump. However, a 33% drop in average CPC behind brand terms is hugely impactful for any account, no matter how you look at it. For ecommerce accounts with high-traffic brand search campaigns, an improvement like this in CPC can result in a lowered overall investment and a significantly improved topline ROI.
Cumulative data shows similarly significant improvement in performance with the expanded test units in our non-brand campaigns. With a 3% reduction, average CPC didn’t see quite the same improvement as it did within the brand campaigns. However, the expanded test units saw serious improvement in conversion rate, which jumped from 1.3% to 2.4% on a 46% increase in conversion volume.
While one could argue that improvements in post-click metrics such as CVR are likely unassociated with copy changes, it could also be that the larger ad format worked to attract additional click volume from a high-intent user base. In non-brand search efforts, it is especially critical to nuance messaging towards your desired audience. In scenarios with high non-brand competition, a third headline can go a long way in conveying a pivotal call-to-action or product detail to a potential converter faced with crowded search results.
While seeing such dramatic improvements in average CPC and CVR is by no means guaranteed with the inclusion of a third headline and second description, this data goes to show the importance of performing frequent tests. Branded campaigns can often be neglected in terms of testing, as strong performance is taken for granted from established legacy keywords. However, advertisers should constantly seek ways to further improve brand campaign efficiency, since a small shift in CTR or CPC can dramatically impact account-wide performance trends. In this case, our test only took minutes to set up and resulted in the valuable discovery that the leveraging of additional fields in expanded ads can considerably reduce average CPCs for high-traffic brand keywords.
For more of our expertise on testing in your account, contact our team.