Quantcast’s global chief technology officer has a message for advertisers experimenting with AI-generated creative – sometimes, it’s best to keep things as cheeky, cheesy and human as you can.
Travelling from the US to Sydney this week, Quantcast CTO Christopher Zimmerman spoke with B&T about the role AI is playing in advertising, why fully AI-generated ads can leave consumers feeling uncomfortable, and what makes creative actually work.
While Zimmerman is based in the US and doesn’t necessarily get to see every Australian campaign firsthand, he has plenty of visibility into the advertising running across Quantcast’s platform, alongside the major campaigns coming out of the US.
The Demand-Side Platform (DSP) helps advertisers plan, measure and optimise their campaigns, but doesn’t create the advertising for them. Its platform also includes an AI-powered Creative Grade by Q tool, which evaluates creative assets against industry best practices and assigns them a letter grade, from A through to F.
Speaking with B&T following a recent Quantcast event in Sydney, Zimmerman said one thing that had stood out when looking at the creative assets being uploaded to the platform was how little AI appeared to be involved.
He said he was seeing “barely any” AI-generated creative among the uploads, and believes advertisers still relying on fully AI-generated ads may be getting it wrong.
“I find the stuff that’s generated right now by AI a bit creepy, and obvious. So I don’t even think customers would be comfortable with us automating it,” he said.
For Zimmerman, the problem comes down to something resembling the uncanny valley: advertising that looks almost real, but not quite.
“I think it’s like that uncanny valley. You can just subtly, almost subconsciously know it’s not real, and there’s this discomfort in it feeling real, but you knowing in the back of your head it’s not, and like almost like it’s playing a trick on you.”
And when asked which AI-generated ads make him most uncomfortable, he didn’t hesitate.
“The ones that creep me out the most are the food ones,” he said. “With a lot of the food delivery services in the US you’re seeing a lot of places now generating the photos with AI and the food just looks wrong.”
A lot of these food delivery services require the restaurant to actually photograph each of the entrees, and so when you are browsing their menu on the app, you can see all the different dishes.”
“The best places are going to have the meals plated and professionally photographed – but now you’re seeing a lot of these places generate those photos with AI.”
And it’s not just the US where this is happening.
AI-generated food imagery has become a widespread frustration on Australian delivery apps like Uber Eats and DoorDash, often resulting in table spreads that misrepresent real meals.
And in October this year, Yahoo News Australia reported on the rise of local Australian takeaway shops using AI-generated food spreads on shopfront posters and menus to save on photography costs.
What actually makes an ad work?
But while AI-generated imagery might be grabbing advertisers’ attention, Zimmerman said the industry shouldn’t lose sight of what makes advertising effective in the first place.
He said Quantcast had found that poor campaign performance was often less about the technology being used and more about the creative itself.
“What we realised was, most of the time when our customers’ performance was poor, it wasn’t our technology. It was something about how they had set up the campaign, and often something about the creative,” he said.
Zimmerman believes successful advertising needs to make an emotional connection with its audience, rather than simply showcase a product or service.
“What works best is ads that touch you emotionally in some way. They can elicit joy, potentially sadness, maybe even make you laugh a little – but they touch you emotionally in some way. Because that is what connects you to the brand,” he said.
“If you can somehow create an ad that not only showcases what’s special about you, and then connect it back to that person, that customer, potential customer emotionally, you’re gonna have – that’s going be a successful awareness-brand-style ad.”
And sometimes, that emotional connection doesn’t need to come from an elaborate or particularly polished campaign.
Zimmerman said even “cheeky” and “cheesy” advertising can work, particularly when it helps a brand stand out from the competition.
“Even cheesy ads, because they’ll stand out from everyone else. But it also has to kind of line up with the culture of the company and its perception, right? Because a brand is almost a personality.”
It’s a sentiment that has also surfaced in recent conversations about the role of emotion in Australian advertising.
In a recent interview with B&T, Ogury ANZ country director Chris Manson discussed the importance of understanding the emotional motivations behind consumer purchasing decisions, particularly in categories such as luxury retail.
And while the advertising industry continues to debate how much AI-generated creative consumers will accept, research suggests audiences may not be quite as concerned about its presence as marketers think.
Research conducted by Ideally and commissioned by B&T in March found most Australians couldn’t distinguish between AI-generated advertising and human-made creative.
The study also found two in five Australians said knowing an ad was AI-generated would not change their perception of the brand behind it.
