One of the biggest recent changes to Google Ads has been the introduction of Responsive Search Ads (RSAs). To make sure they’re the right choice for your brand or client, it’s important to test rigorously.
Paid search practitioners are always keen to explore new formats that might have a profound impact on account performance but it’s vital to measure potential ROI before jumping in.
This article discusses our initial findings of this ad format, what we at Blueclaw are testing as an agency, and our thoughts on the future of the Responsive Search Ad format.
So – what do we have to work with?
What are RSA variants?
Responsive Search Ads allow up to 15 headline variants and four description line variants to be created. These are the building blocks of your ads.
From here, Google then optimises and tests each variant using machine learnings, showing a variant most likely to get a user to click through.
For extra control, advertisers can “pin” headlines and description lines in a particular order (useful for advertisers who have disclaimers they have to show for example) although this reduces Google’s ability to optimise and test.
All this is in contrast to existing Expanded Text Ads functionality, where users specify the headline and description order. The character limits are now the same for both Responsive Search Ads and Expanded Text Ads – more on this later.
The selling point of Responsive Search Ads is the notion that combining custom elements to dynamically generate ads will result in ads that should be top performers, drawing on your best headlines and descriptions that are proven to drive clicks.
For some, the support and suggestions of the Google algorithm will be a welcome resource, while for others the thought of trusting Google automation to get things right for their campaigns – particularly in regulated markets – will be a big adjustment.
Creating a Responsive Search Ads testing plan
To develop our own understanding before recommending the use of Responsive Search Ads to clients, we initially wanted to understand how the Responsive Search Ads performed versus existing Expanded Text Ads, using a large, UK-based consumer electronics ecommerce client as a (willing) test subject.
We had a secondary objective of understanding whether pinning headlines and descriptions impacts our click-through-rate (CTR) and conversion rate (CVR) metrics.
To do this, we tested two Responsive Search Ad and Expanded Text Ad variants across non-brand and brand campaigns over an initial 30-day period of our highest impression volume accounts to gain statistically robust data.
First round test insights
Our first test was based on 939k impressions. The data shows that Google will almost always show ads which are not pinned if given the option to do so – though pinning was suggested during the beta preview.
Initially, as documented in our Blueclaw blog post of November 2018, Responsive Search Ads were significantly outperformed by existing Expanded Text Ad formats in terms of click-through rate (CTR) and conversion rate (CVR).
CTR and CVR were almost identical across exact match campaigns but there were significant differences between broad match performance.
An unexpected finding
We expected the dynamic ads to have improved performance on broad match keywords, with the dynamic nature of the ad variations more suited to learning the optimum ads to display based on a wider spread of potential broad match queries.
However, this was not the case and there are no obvious reasons as to why.
In fact, in this round of testing, there were only a small number of campaigns in which the Responsive Search Ads performed better than the Expanded Text Ads.
Aside from the technology being comparatively new, we suspect the main reason responsive ads had a much lower CTR is the potential for ads to serve headlines that draw on the description templates that are closest to the search term used by the user.
USPs versus automation
We recognise that for many sectors, providing headlines that search engine users can immediately recognise as relating to their query is a commonly known and well-established best practice ad copy tactic.
However, we suspected that there were more important aspects for search engine users than finding ads with headlines that mirror their query.
This is particularly in the case with our consumer electronics client who has a lot of information about products, price, trustworthiness and availability that can be surfaced in ad headlines.
Our next test attempted to verify this hypothesis by testing one Responsive Search Ad variant with a pinned Headline One against two Expanded Text Ad variants based on our initial test findings.
Second round test insights
The second test was based on 1.33m impressions – an increase from the initial assessment as we broadened the keywords we used to test Responsive Search Ads. The results from our second test showed Responsive Search Ads significantly outperformed Expanded Text Ad variants in terms of CTR and CVR.
The key difference in this second round of testing was using the RSA functionality to pin Headline One across all ads tested. This was across all match types as well as brand and non-brand keywords.
Google may argue all headlines should be tailored to user’s search terms, so there should be no need to pin a headline.
The Blueclaw view is that for many industries, it’s more important to include business USPs within headlines as a priority, so it’s not always possible to include search term copy also due to character restraints.
Either way, Google will be providing us with suggested Responsive Search Ad copy so we will be able to draw insights from their approach as time goes on.
Watch this space
With responsive ads accounting for 42% of our impressions and Google extending the number of description lines available on Expanded Text Ads, the underpinning thinking and functionality behind Google’s Responsive Search Ads beta is clearly here to stay.
It’s important to acknowledge this is a beta and there will inevitably be refinements to the technology which should be reflected in improved performance. Google also state that Responsive Search Ads should be seen as a supplementary to Expanded Text Ads, rather than outperforming them. This was presumably in response to advertisers struggling with Responsive Search Ad performance which anecdotally suggests our findings may be broadly representative of the wider search advertising industry.
Several advertisers (including myself) saw initially lower CTRs when expanded ads were rolled out as standard more than two years ago – but as functionality matures, results often increase.
What was new becomes part of the standard paid search toolkit. There is also a clear link-up to Google’s next big focus – automated bidding strategies. The two will complement each other and lead paid search towards the direction of machine learning and algorithms.
Google will only make changes to their platform if they can guarantee it grows their revenues, so I expect CTRs to increase over time and for this ad format to become the standard.
The future of Responsive Search Ads
Since launch, Responsive Search Ads have undergone further additions with multilingual capability extending beyond English to Danish, Dutch, Italian, Japanese, Norwegian, Portuguese, Russian, Swedish and Turkish.
Beyond language, Google has also simplified that ad setup process with machine learning used to guide auto-generated suggestions for headlines and descriptions, coupled with real-time feedback on Google’s assessment of ad quality.
This will be another interesting aspect of functionality to test in order to weigh up the value of unique ads, crafted by brands and ad copy that is increasingly led by general ‘best practice’.
Who will be using Responsive Search Ads in future?
We anticipate that in-house marketers will make more use of Responsive Search Ads as a time-saver, though we recommend that testing is done to ensure that the time saved is actually coupled with improved results.
As for paid search specialists like my team, we’ll continue to explore Responsive Search Ads on a case by case, industry by industry, client by client basis to ensure that we’re getting more value for clients.
Agency practitioners need to be totally confident in the work they carry out for clients, and even potentially exciting functionality such as Responsive Search Ads does open up the possibility of automated actions taking place. The guiding hand of an expert paid search strategist is essential – these tools are no substitute for truly understanding a client’s market.