Member Exclusive   //   June 18, 2024

DTC Briefing: As Meta & Google embrace AI ad products, brands fear losing creative control

This is the latest installment of the DTC Briefing, a weekly Modern Retail+ column about the biggest challenges and trends facing the volatile direct-to-consumer startup world. More from the series →

It’s the year of AI for ad platforms like Meta and Google. 

Nearly every week, it seems like Meta, Google or one of their competitors has an announcement about how they are automating some part of the advertising process with the help of AI. Last week, Pinterest was the latest to announce a new suite of AI-powered features. This includes a tool that produces personalized backgrounds to a mysterious new program called “Performance+” that’s in closed beta for now, but promises to “build goal-based campaigns that help deliver better advertiser performance.” At the beginning of May, Meta also released a series of tools that allow brands to upload images and create AI-generated variations of them. Then there’s the rush from both Meta and Google to integrate AI assistants into search.

If ad platforms like Meta and Google encourage brands to adopt these new AI ad tools, many of them have no choice but to follow suit since that’s where many of them spend the majority of their advertising budget. In 2022, these two platforms combined accounted for 50% of digital advertising spend, according to eMarkter. While most brands are adopting these ad products — and many said they are seeing positive results — it raises questions about just how much of a black box brands can be comfortable with. All of these AI-driven campaign tools require brands to give the platforms wide leeway over where they might serve an ad. Some marketers also expressed concern that these AI-powered tools will push brands to produce a deluge of creative assets that ultimately all look the same. 

“I think AI tools are responding to the same thing marketers have already been responding to, [which is] what drives outsize results in-platform,” Matt Breuer, the senior vice president of portfolio and growth strategy at e-commerce holding company Aestuary said. But, he said, if it gets to the point where these AI tools are encouraging brands to spit out copy and images that all look the same, “then your ability as an advertiser to stand out starts to be really diminished.”

It’s also important to note the backdrop against which these tools are being released. Amid inflation, more DTC startups are struggling to grow their revenue, especially in the e-commerce channel. In the latest round of quarterly earnings reports from e-commerce companies, most DTC startups ranging from Solo Brands to Warby Parker reported negative or near-flat e-commerce growth. By investing in these AI tools that are supposed to make certain parts of the advertising process easier, platforms like Meta and Google. hope that these tools will help more brands run more successful campaigns. And, in turn, spend more money on their platforms. 

As Sean Frank, the CEO of Ridge Wallet put it to me in an email, “more and more of the solutions to problems are AI tools.” 

“If things are going great, then [there’s] no need to use AI,” he said. “But they are building and offering solutions for people that have a harder time scaling.” 

The marketers I spoke with said that they would prefer that the platforms, most specifically Meta, spend more time fixing bugs related to the ads manager. And already, some marketers are seeing some bugs with these new AI tools.

Still, Katya Constantine, founder of the agency DigishopGirl Media, said that many of the AI tools that have been announced are simply an “evolution” of what Meta, Google and more have already been investing in for the past couple of years. 

All of these tools are meant to automate certain parts of the advertising process that a media buyer and their team might have had to previously do manually. That includes creating a series of images for the ad campaign, deciding where to place an ad, and what type of user to show an ad to. 

Many of the newest tools are being compared to Google’s PMax, one of the first end-to-end automated advertising solutions that was initially released in 2021. PMax, Meta’s Advantage+ Shopping Campaigns and Pinterest’s soon-to-be-released Performance+ all make a similar pitch: proprietary machine learning models will predict who the highest-value customers might be for any given brand, and serve ads effectively to them. 

It’s worth noting that for many platforms like Meta, the push to invest in its own machine-learning models was largely driven by Apple’s iOS 14 update, which required users to opt in before apps could serve targeted ads to them based on their past browsing history. In turn, Meta has spent the last two-plus years rebuilding its entire ad platform using AI, so that it can now serve ads based on the historical data it has. Meta’s ASC, which it released in 2022, hit a $10 billion run rate in 2023

Most brands have adopted PMax or ASC in some form. Duane Brown, the founder of the agency Take Some Risk, said that ASC doesn’t work as well in smaller markets outside the U.S. Constantine said it is better suited for marketing to new customers.

Meanwhile, Constantine and Brown both say that all of their clients have adopted PMax, largely because Google has depreciated Smart Shopping, its predecessor to PMax. Brown, however, said that he usually advises clients to take a “’yes, and’ approach to PMax” — that is, running PMax in conjunction with search campaigns or standard shopping campaigns. 

With PMax, brands are encouraged to use broad match, which means that ads show up on searches that may relate to a specific keyword. Brands can add exclusions of keywords they don’t want to show up against, but, overall, they are encouraged to take a broad approach to give Google’s algorithms more jurisdiction over where they may place an ad. 

Brown then often encourages clients to still run a standard search campaign, “to be able to make sure we are still showing up for the right searches.” At the same time, “you don’t want an ad account that’s just for brand searches,” Brown said, and PMax can be useful in finding a broad range of search terms that people might input when researching a product decision to run ads against.

It speaks to the inherent tension many marketers have with this new generation of AI tools. They essentially ask marketers to trust Meta and Google to figure out where to place ads and to find the right customers, which, to be fair, is what these ad platforms’ initial promise is to begin with. 

But, marketers are thinking about what sort of data and reporting they need to feel confident in these opaque AI ad programs.

With Google’s PMax, Breuer said he would like to see more granularity around what search terms Google decides to run ads against. He believes that Meta’s ASC — while it also chases broad placements — gives granular enough reporting that if, say, he sees that performance has changed significantly and Meta has suddenly placed a ton of ads into its Stories placement, he can make subsequent decisions based on that data. 

What also remains to be seen is just how much new AI tools that automate the creative generation process will change what type of creative resonates. 

Constantine believes these tools are welcome because brands often have to test out a large volume of creative assets to find success on platforms like Meta. “I think where there’s a little bit of tension is with brands, who have a really strong creative-first approach or a brand-specific voice and vision,… [becoming] comfortable [with] AI-generated creative,” Constantine said. “I think this is going to be a little bit more painful for brands that have a harder time adapting to this methodology.” 

Breuer said that he thinks how brands decide when to accept Meta and Google’s AI-led recommendations will be “the central tension with this marketing shift.”

“No one wants to sit on the sidelines and say, like, ‘all the new ad products, I’m just not going to use that and get left behind,’” he said. However, he went on, “At some point in time — or in certain situations — I think the right answer is going to be to opt out or just potentially set yourself apart from the ways these tools tend to make things feel and sound the same.” 

What I’m reading 

  • Care/of is shutting down, close to four years after Bayer bought a majority stake in the supplement and vitamin company. Care/of said on Instagram that “we unfortunately no longer have funding to operate in the way we have been” but did not provide further details. 
  • Kim Kardashian’s Skims is opening its first permanent brick-and-mortar location in Georgetown, D.C. later this week, and will follow that up with four more store openings later this year, Women’s Wear Daily reports
  • Rihanna’s Fenty Beauty is expanding into hair care

What we’ve covered 

  • How brands like Fulton, August, Haven’s Kitchen and more are experimenting with guerrilla marketing tactics, from wheatpastings to cross-country road trips documented on social media. 
  • An interview with DayDayCook about how it plant to grow its portfolio of Asian food brands after acquiring Omsom
  • Last year, TikTok ignited a cottage cheese renaissance, after recipes for everything from cottage cheese flatbread to cottage cheese ice cream went viral. Now, brands like Good Culture are trying to take advantage of the renewed interest by investing more in brand-building efforts.