Like the ocean tide on a windy day, one wave isn’t done glazing the sandy shore before another comes crashing in with fresh seawater.
Such seems to be the cadence with internet search advertising. Marketers next month will bid adieu to Google’s Expanded Text Ads—introduced in 2016 mainly to cater to the explosion in the use of mobile devices—to yield to Responsive Search Ads. June 30 is the sunset date for the creation or changing of any Google Expanded Text Ads.
In a practical sense, the transition is neither sudden nor earth-shattering. Google announced the sunsetting in August 2021, and the vast majority of advertisers already are using Responsive Search Ads. But, subtly, advertisers no longer having a choice marks a significant step in the shift toward machine learning for world’s biggest internet advertising platform.
Across the board, the industry is using these digital tactics to make advertisements as relevant as they can be for their audiences. This is just working smarter to serve up the most relevant content.
Andrea Ness, Media Manager for ddm
Out with the new, in with the newer
Responsive Search Ads (RSAs) are just two years younger than Expanded Text Ads (ETAs). Google rolled out RSAs in 2018. Both types of internet advertisements were made to boost click-through rates versus standard text search advertisements.
The soon-retiring ETAs boasted a higher wordcount than the text ads that preceded them. Bigger is better – it was as simple as that. Marketers wanting to reach an audience would simply stay within the character counts for headlines and text, and write ads that hit the right key words and appealed to all users on all devices.
The complete creative control was familiar, and the longer ads had ETAs out-performing their predecessors with a higher click-through rate.
RSAs are a different story but with the same ending.
To create an RSA, the user types in multiple headlines and descriptions. Google runs all of these headlines and descriptions in different combinations. Through running the ads and collecting statistics on click rates, Google identifies the most impactful characteristics of the text until it has honed the ad to its most effective iteration.
The person who drafts the ad works a little harder. But Google works a whole lot smarter. The end-goal remains the same: More clicks and more conversions.
Yielding creative control
If a marketing team produces a billboard, an e-mail, or a magazine ad, the client knows exactly what to expect when they see the content live. The copy is finalized. It’s given a blessing from the marketing decision maker, and it goes out to its universe to make its impact.
Those in the digital marketing realm don’t operate that way anymore.
“We can’t create all the copy that would align with everyone’s unique searches,” says Ness. “So we’re letting algorithms do the work.”
In the history of marketing, we’re still in the nascent stage of dynamic, machine-assisted content creation. Never before the last few years have content creators been able to write basic copy and have a machine real-time test it in thousands of increments until it resonates with specific audiences.
We’re just getting used to seeing ads that we didn’t actually create word-for-word, and ads that are never technically finalized.
What you don’t know… markets better?
It appears to be working, too.
Though click and conversion rates vary greatly depending on the campaign type and industry, Google reports that advertisers who replaced ETAs with RSAs in similar situations and using the same assets “see an average of 20 percent more conversions at a similar cost per conversion.”
“They’re showing up more in searches,” Ness adds. “It does lead to more clicks and conversions.”
The proliferation of responsive advertisements dovetails with Google’s ability to use machine learning to target audiences. As we discussed in our recent post on Google Analytics 4, the collection of individual-level data is being curtailed. The saving grace is that we have tools to build, learn about, and hone our audiences based on machine learning, even if we know less about them to begin with.
RTAs fit nicely into this world. We may not know the identity of the individuals we want to target, but we’re learning what we need to know about them as they search, and our content is automatically being adjusted to cater to them.
It’s a clear sign that machine learning will have an ever bigger role in shaping that figurative digital marketing shoreline.