Ten passes left to attend the Digiday Publishing Summit
The promise of generative AI was to make marketing faster and cheaper, optimizing everything from media operations to creative. But in Silicon Valley’s push to give marketers a one-stop-shop with AI, marketers say AI has not given them solutions for brand building.
The industry might still be a ways off from fulfilling Mark Zuckerberg’s vision of fully automated ads by next year.
AI and automated tools, like Google PMax or Meta Advantage+ have largely been relegated to performance marketing campaigns, according to six marketers Digiday spoke with for this story. That’s in part because there’s enough hard data and attribution in a performance marketing campaign to train the automated tools. Meanwhile the media buying and marketing automated tools available for CTV, digital video and other brand marketing campaigns (a la The Trade Desk’s Kokai platform) hasn’t exactly wooed marketers.
“Generally, it seems like AI is useful when you can tell it exactly what you want to get and then feed a bunch of data into it to make it happen,” said Chris Rigas, vp of media at performance media agency Markacy.
AI excels at data-driven tasks and scaling content generation, Rigas said. The challenge with brand marketing metrics, however, is longer feedback loops, fuzzier outcomes and other intangibles making it more difficult to drill down data points to feed the automated tools. “With the upper funnel channels, telling AI to get you a click or view or an impression is not helpful because those metrics are not really driving business value in any way,” Rigas added.
Even with AI’s black box nature, marketers say automated performance marketing tools have made good on the promise to target audiences likely to click and buy. Targeting audiences for brand awareness in upper funnel campaigns is proving a tougher hurdle to clear.
To some extent, marketers say, brand marketing has always been harder to quantify — measuring things like brand affinity, lift and awareness aren’t as definitive of metrics as performance results like return on ad spend and sales. But for all the promise of AI efficiency in marketing, agencies have had to build out proprietary tools to layer on top of the tech tools currently available to bridge the gap.
For example, Brandtech Group’s Jellyfish digital marketing agency this year launched its Share-of-Model platform, an AI-powered market research tool that scrapes insights from large language models like OpenAI’s ChatGPT, Meta’s Llama and Google’s Gemini. Those insights can then be used to inform other aspects of the campaign to sharpen media buying and audience targeting, according to Karen Bennett, U.S. managing director for Jellyfish.
Jellyfish isn’t alone in its approach. B2B marketing agency Transmission leverages lookalike audiences and digital twinning for more accurate campaign results, and simulates brand impact, said Alex Beddoe, head of biddable media at Transmission. It’s a move to create digital models of customers to help predict how brand campaigns will perform more accurately. “Being able to take our client data and then modeling against that with like a lookalike, that’s where a lot more of our spend is going towards,” Beddoe said, referring to an uptick in client spend on predictive audiences to better contextualize campaigns. He did not provide specific client spend figures.
Meanwhile, Rigas at Markacy says clients have also pulled back from Meta Advantage+ campaigns. Over the past year, the agency has been recommending clients carve out for a video buy to drum up brand awareness within Meta Advantage and Google PMax campaigns, he said.
Platforms like Google, Meta and Amazon promise scale and automation with their AI tools. But brand marketers aren’t fully on board yet. There have long since been concerns about the black box nature of these tools and those concerns have only been heightened as tech innovation continues to move more quickly than marketers can keep up with. In one case, a 20-year growth marketing veteran has advised clients to walk away from PMax.
Still, marketers believe it’s only a matter of time before automated tools become one-stop-marketing shops. But current capabilities and measurement limitations are keeping them at the bottom of the funnel for now.
“It’s still early — but clients are seeing some benefits from leaning into AI tools. But there’s still a lot of work to do,” said Anthony Costanzo, chief analytics officer at Mile Marker independent media agency.
More in Marketing
WTF are AI agents? (video update)
Despite so much use of the A-word, it’s still early for AI agent adoption, meaning marketers should ask what agents are for, how they’re made, what they do, what they might do — and what they can’t do — including potential reputational risks.
‘Some brands will continue to take liberties’: Confessions of an influencer marketer on brands misusing creator content
While most brands do the right thing, there are still the odd few which try to cut corners, or, (more worryingly) think the same rules don’t apply to them.
What Blue Apron’s move to in-house its influencer marketing strategy says about the creator economy
Blue Apron has brought its influencer marketing in-house, part of a broader push to streamline operations following its $103 million acquisition by Wonder Group in 2023.