The programmatic advertising industry may have reached its “where’s my flying car?” moment.
The automation of ad buying has advanced the market beyond the days of manually delivered insertion orders, but is programmatic advertising fulfilling its promise or has its pace of progress slowed? That’s a question that GroupM Nexus president of North America JiYoung Kim put to a room of programmatic marketers at the Digiday Programmatic Marketing Summit in Palm Springs, California.
“We’ve kind of took the pedal off the gas when it comes to innovations [in programmatic advertising],” Kim said onstage on May 22. In other words, programmatic advertising has become the predominant means for marketers to reach people digitally, she said, “and we’ve stopped working so hard to understand how do we use this to make advertising actually better.”
A consequence of programmatic advertising’s stalled status is its potential impact on the future of the programmatic workforce. “It used to be so exciting to be in programmatic and digital and social, search — this used to be cool. And if the thing we’re going to hang our hat on is we can reach millions of people, that’s not exciting,” said Kim. At a time when automated tools are making it easier to offload work onto computers, she called for a renewed emphasis on tradecraft.
“You can’t just apply machines because everyone has the same machines. The thing that will make a difference between a better campaign for your client versus your competitor, better allocation versus another, is really the decisions that human beings make,” said Kim.
One area in need of innovation is ad creative. She also cited measurement, but she emphasized “creative probably more because I do feel like we’ve spent less time on it.”
Companies like Meta have rolled out dynamic creative options to formulate ad slots on the fly based on a set of advertiser-provided assets like image and copy variations, but those tools are akin to automated color-by-numbers creative production. What Kim is looking for is more insights — such as the sequence of events that led up to an ad exposure and then followed it — that can be used to inform the creative.
“Imagine being able to predict or pretty much understand what a set of behaviors and a set of qualities mean in a model or in a cohort. That’s a pretty interesting space to play in,” said Kim. “And that’s essentially what ChatGPT and a lot of these AI [tools] are. It’s taking a lot of different potential outputs, and it’s selecting the best combination of responses based on what they have observed over time.”
Cohort-based advertising may not generally be considered all that advanced, or at least not very personalized compared to one-to-one ad targeting. But Kim outlined an advanced version of the cohort-based model that would not simply group people into broad categories like outdoor enthusiasts but would layer in other information to provide more context and would seem to aim for some hybrid between the cohort- and individual-based approaches.
“What do they do before and after they go outdoors? How often do they go? Where do they live? Do they go with people? Do they go by themselves?” said Kim. “All of these things start making my predicted engagement with you a lot more accurate. … It’s not as creepy as it sounds.”
This article has been updated from an earlier version sent exclusively to Digiday+ members that included incorrect information about the amount of money advertisers spend programmatically to reach U.S. adults each day.