Sir Martin Sorrell on Learning to Live With AI Ad Monopolies
Sir Martin Sorrell on Learning to Live With AI Ad Monopolies
Plus Google's trial, and who's going to win the election?
Few executives have a better macro view of the ad world than S4 Capital Executive Chairman and founder Sir Martin Sorrell.
For an industry that has been eyeing a certain antitrust trial, perhaps anticipating several more, on the latest Next in Media podcast episode, I asked Sir Martin, how’s business? More to the point, how’s competition looking in digital advertising? Will all this regulation fervor spark a redistribution of ad dollars?
“It’s really a tale of two cities,” he said. “There's a digital city, which is vibrant. And there's a traditional media city, which is under huge pressure.”
Here’s how he broke down the trillion or so dollars S4 spends for clients:
“700 billion or so of that, maybe a little bit less, is digital and 300 billion, a little bit more is traditional
Digital will be up this year by 10 or 15%
Traditional has been down by 0 to 5%, 5 to 10%. If you didn't have live sports like Warner Bros. Discovery, you might be down as much as 10 or 15%.”
Where is that $700 billion going? Google, Meta and Amazon in the US, and Bytedance, Alibaba and Tencent in the East.
“Those six platforms are the dominant ones,” he said.
It’s not easy to see what changes that dynamic in the near future, as Amazon in particular is on a trajectory to reach $100 billion over the “next three or four or five years.”
And that’s before AI really takes hold in a huge way, which Sorrell sees as inevitable. The question is whether that makes those six platforms all but unstoppable, and turns brands and agencies into bystanders.
Sorrell been very outspoken about his belief that AI is going transform advertising - and not just creative thanks to generative AI, but media buying, planning and optimization.
Shouldn’t we be worried that these companies are not only dominating digital ad revenue, but may be the only ones who'll be able to compete in AI media investment tech?
Should the big media agencies or even brands try to compete on this front?
“No, I don't think so,” Sorell said.
“Let's just think of that trillion dollars that was spent, let's assume it was one client for a minute…that trillion dollars in media planning and buying is something that is much better executed by an algorithm as it is investment management,” he said.
He likened the future of media spending to that of hedge funds.
“I'm sure that BlackRock and other fund managers, Vanguard, etc., they allocate these things to algorithms…it's not 25-year old media planners and buyers,” he said. “I mean, the simple fact is why is it that media planning and buying operations have 10,000 people? It's because the process is semi-automated or manual and they should be automated and, and driven by, by machinery.”
“The outputs which these platforms produce are so sophisticated and good that the people, those planners and those buyers, at 25 or whatever, will be able to make much, much better informed decisions.”
Right, but won’t we be just handing over our keys, and our budgets, to a few black box AI companies?
“It's exactly like making a decision between equities, bonds, properties, alternative investments, exactly the same thing. And instead of making those decisions on whim, gut, insight, relationship, whatever,” he said.
So are media agencies screwed? “My view is the role of the agencies will become much more aligned with the platforms,” he said. “Wo we are validators for the platforms. We forge a very close relationships with the platforms for that reason.”
What about the election? Kamala Harris has a war chest of $1 billion, and could very well lose, which might indicate that advertising isn’t that its cracked up to be.
“You know, it's obviously a 50-50 election,” Sorrell said. “You know, if you push me against the wall, who do you think would win? I would say probably Trump. Trump seems to me to be the, not the favorite, but if you look at the betting markets and the stock market.”
“People focus a lot on the personalities. And some people like President Trump's personality, some people don't. And the reverse is also true.”
If you’re not inspired now, I don’t know what to tell you.
Why Brands and Agencies Need to Shift Their Measurement Approach - With and Without the Walled Gardens
This interview is part of a multipart series with my sponsor Epsilon. Today I’m talking with Esme Robinson, Director, Platform Solutions at Epsilon, about how measurement is changing in digital ads - in light of ongoing (some would say monopolistic) consolidation
Next in Media: What should we make of this shift from measurement to modeled measurement in media. Not everyone likes the idea of supposedly giving up precision targeting…
Robinson: It’s not perfect, but it's better that what we’ve been using. The issue we are facing is, who can brands trust on this? To trust any one thing because that one thing is never going to be perfect. Why should we trust it? So I think it takes a combination of different types of models of measurement. It takes working with partners and vendors that you trust to be able to give you the reporting that you need and not just rely on an output, but be able to be as transparent as possible about what's going into it, why we're getting the results that are coming out of it.
Next in Media: While we’re talking about modeling in measurement, we’ve also seen renaissance of old school marketing mix modeling - but it’s being pushed by platforms like Meta and Google. Is that a positive for the industry?
Robinson: Well, We've got the options. It's positive to have choice, positive to have different ways of doing things. I think the things where it's maybe less positive are kind of what I was going back to before in terms of certain [companies] dominating how things should be done, doing it their way, the fact that it's people running these measurement models that are also including delivery through their own channels. That's, think, when it becomes a bit trickier.
Next in Media: Ideally, should there be a bunch of third parties emerge that do this kind of thing really well? Or do the brands and agencies need to build their own models so that they have a little bit more control?
Robinson: I hope it doesn't sound like a copout to say both because I think it needs both. I think you need to have independent measurement vendors that can be unbiased. But to make that worthwhile in quality, you also need real results, real delivery and performance metrics from the vendors, the agencies that you're working with.
Next in Media: Another thing that's coming up a lot is identity resolution. Why is this so important, considering that many new players like Retail Media Networks have great identity data (albeit, it’s fragmented)…
Robinson: Yeah, exactly. Just because you're targeting first-party data, how do brands know that that's purely what they're reaching? I think particularly in this world of audience segments for retail media…when it's predefined, when you're, you know, working with the vendor and you have kind of zero input or transparency into what makes up the audience, again, it comes back this idea of trust. And with identity resolution, particularly, I think the important word is ‘omni-channel.’ If you're trying to deliver to people on site, look at how they purchase in store, there's no way of connecting that without solid identity resolution. And I think people might misunderstand or underestimate the complexity of trying to connect in-store conversions to impressions in the open web. Like that's no mean feat. That's not simple at all.
And so working with partners who are able to do that can make a big difference.
Next in Media: I have to ask you about AI. There's so much talk about how it's changing the industry or may change the industry. How might it enhance relationships with brand between brands and consumers - or displace that completely?
Robinson: Yeah, and funnily enough, I actually think this is so linked to the measurement idea. I think that's the perfect example of how that's the case. When vendors or these walled gardens act as intermediaries, when you're putting the trust in them, both for who to deliver to, what KPIS to deliver them on, trust my results, it becomes kind of this disjointed, disconnected relationship. The way I can enhance it is knowing what goes into it, understanding what the outcomes are going to be of that model, of that machine learning.
Next in Media: What's your biggest challenge to try and bring a lot of this stuff to fruition?
Robinson: I think cultural and structural organization of setups, I've kind of touched on it before, but the idea that relationships between a partner and the marketing head of CRM, or whether that's of paid media, might be really strong and there might be a really good understanding of the business goals and challenges and then how that partner is delivering on it. But then trying to get different levels of the organization to understand, to buy in, can be really difficult.
Lastly, here’s an interview I conducted during Advertising Week with Mike Treon, Head of CTV & Video Strategy at PMG, talking about what needs to happen to get more small to mid-sized businesses advertising on TV. It’s part of a multipart series I’m working on with my friends at Blockgraph. Check it outL: