Media Buyers vs AI
What happens when everyone's got an algorithm?
Everyone in the ad business continues to extol the revolutionary power of AI.
There is a mad rush to use the tech, or say you’re using it, and ideally that you’ve built something special on top of it for your clients - either to run their media campaigns more effectively, or to crank out actual ads more efficiently.
Here’s another sentiment I’ve heard lately - it’s not going to be cool with clients if and when media agencies or top buyers can’t explain how or what’s working in their ad campaigns. A “shrug, it’s just the algorithm response” won’t help marketers feel comfortable keeping their planning and strategy teams around.
Yet I wonder if that’s where things are headed?
I had Jeff Matisoff, Partner at Jellyfish on my Next in Media podcast this week to talk about the comeback of marketing mix modeling, AI, and all sort of macro industry topics. I asked him about the agency race to build something AI-powered that is smart, powerful, and ideally proprietary in a world where only a few tech giants are able to invest the billions needed to build large language models and AI search bots.
“We're 800 people,” he said. “We're not going to build an LLM to beat Gemini. Like that's not going to happen. I don't think Omnicom is going to build one.”
Instead, the race will be to iterate on top of AI platforms, a la ad tech.
“It's just going to be the same way with programmatic, right? Like everyone tried to have their DSP and their bidding solution, and now there are three or four [in the market] Cool. Guess what? It's going to be the same way with AI.”
Maximize your YouTube advertising with VuePlanner. As a member of the YouTube Measurement Program for Brand Suitability and Contextual Targeting, VuePlanner enables you to buy with confidence, clarity, and precision. Using advanced technology and AI-powered optimization, VuePlanner offers custom-curated contextual collections, exclusive content strategies, and transparent reporting for measurable, impactful results. Take control of your campaign performance—partner with VuePlanner now.
The programmatic analogy doesn’t sound so bad. After all, each media agency doesn’t have its own DSP, nor each seller its own SSP - yet lots of business gets done, and thousands of people are employed in programmatic advertising.
But when it comes to AI, the talk is increasingly that media planners won’t need to touch media campaigns in the near future. Analyst Jack Meyer says by 2030, “80% or more of all media planning and buying will be done without human intervention.” At CES in January, GroupM CEO Brian Lesser basically concurred. And of course, Sir Martin Sorrell has been calling media agencies toast for several years.
For its part, Jellyfish has leaned on AI tech to build its own MMM, while acquiring a Gen AI platform geared for ad creatives called Pencil.
Matisoff says that his team will always have to play a big role in steering the machines - and that’s likely to be a big part of agencies’ long term value. “We still want a human in the loop, even if we're working alongside Performance Max or an Advantage Plus to at least be monitoring. Like what is it actually doing? Can you understand it?”
That understanding and insight seems crucial as agencies look to reinvent their roles and maintain their relevancy. However, what happens over time when everyone has their own algorithm, everyone is building on the same platforms? There is already a risk for AI/algo commoditization, say some experts.
Then consider these comments from Procter and Gamble CFO Andre Schulten during a recent earnings call (thank you Madison and Wall).
“We're using automated media buying to increase media reach with greater precision… The Consumer 360 proprietary data platform…enables brands to use target audience algorithms to reach the widest range of consumers where they are the most receptive to messages, serving ads at the right frequency each week all year around.”
“We are reinventing the agency model again. We've reduced the number of agencies retained by more than half from 6,000 10 years ago, delivering $100 million in average annual savings.”
“We are implementing in-house media operations, achieving annual media savings of up to $500 million through data and analytics capabilities applied to planning, negotiation, scheduling, and buying, and we now have in-house operations in nearly 100% of North America.”
That sounds like agency loyalty is going to be very hard to come by in the coming years, especially if more brands think they can do it all themselves.
One other aspect of this would-be AI media buying revolution that I’d think would worry agencies and media companies is the technology’s early bias toward performance advertising; Google calls its product Performance Max after all (if you think transparency and brand safety are issues now…). How is this not going to skew an industry - one that is already embracing performance metrics more than ever - toward becoming 100% about outcomes?
Matisoff thinks the machines will eventually figure out branding.
“Right now [AI buying tools] are hammers for nails, but there are other AI tools that Google and Amazon and Meta are releasing that are not just heavy bottom funnel. So I think that they're experimenting with other ways. But branding means different things to every brand and the KPIs are different. And so it's a, it's a broader struggle.
“I would bet that in six months, nine months, definitely a year, we'll be using branding tools with the platforms in the same way that we use the managing performance.”