Pattern Recognition: The Future of Ad Holding Companies?

AI is fraught with peril yet can empower agencies to regain their core value in building and marketing brands by relating the client's own knowledge and data to their own experience and capabilities, to balance business, societal and ecological goals.

Advertising holding companies should be called Integrated Marketing Services companies (or orchestrators). As audiences and media fragmented, they developed to offer a giddy array of specialisms. Their relative decline in revenue, stock price, and (most tellingly) importance to business decision-makers, is all well-documented.

Management consultants usurped business strategy from agencies. In 1992 my independent UK agency won what was described as a large client assignment, I was mortified to discover the real client was McKinsey, and our fees were less than 20% of their consultancy team’s.

Technology platforms usurped data and measurement and extracted rent from marketing and media execution. In 2013 I was mortified when WPP convened its strategy meeting in Mountain View CA, and we were bussed around Apple, Facebook, and Google.

It was easy to criticize holding companies, as I did until I accepted the inevitable and joined WPP. I learned the obvious: the orchestration of diverse capabilities and best practices was critical to deliver to large global brands and corporations. The holding companies only succeeded when they integrated and orchestrated.

The agencies have adapted to the shifting media landscape as well as most of their corporate clients (which is just OK).

The question now is what is their future in the age of AI?

Joe Mandese ran a ‘quantitative’ count for his MediaPost newsletter analyzing the mentions of AI in holding co investor updates. I noted that he may have had his tongue in his cheek here in ascribing leadership to WPP followed by Stagwell.

WPP, my alma mater, has in recent years followed a failing strategy, resulting in the worst performance among its peers. Stagwell is a cobbled-together minor player, also a poor performer. It would be tempting but simplistic to conclude AI will be bad for agencies.

Publicis and Omnicom seem the true leaders, Publicis after a recent, impressive high profile executive briefing. Publicis was well-positioned for AI in data and digital after its acquisitions of Epsilon and Sapient. Omnicom has invested in its own data platform and is executing an eCommerce strategy with the acquisition of Freewheel.

The CTO of WPP, Stephan Pretorius, gave a strong presentation on AI and the Radical Transformation of Marketing, and WPP’s progress, at an Nvidia event.

Yet Stephan’s response to audience questions confirmed the confusion in the industry.

Asked about AI’s potential to accelerate greed and inequality, he did not have an answer. This is not a criticism: 99% of the industry ignores the impact of marketing on society and ecology. Mandese noted no references to ethics in recent company presentations.

Asked about the content explosion fueled by AI, Pretorius noted WPP’s tools improved relevance, personalization, and engagement. This is a criticism. Brands are not built 1:1, they connect to culture and communities that share values and experiences i.e., mass media.

Asked about new and transformative workflows, he recognized the opportunity, but other than promising a new conversational and dynamic approach to the CMS, he had no answer.

The serious point is this: brands and their agencies were captured by the digital platforms, adtech and martech (according to Scott Brinker there are apparently 14,106 viable martech tools), Advertising with honorable exceptions offered no strategic response as to how brands should be sustained and how innovation can create new value in an unregulated media ecosystem with increasing concentration of power and data.

Over the last decade. with rare exceptions, the technology platforms extracted billions of dollars of brand value from brands and agencies, degraded the online customer experience, and caused devasting societal harm. Social observers and economists document the process. Marketers must understand the analysis and reporting of Shoshana Duboff, Cory Doctorow , Yanis Varoufakis, Tristan Harris, and Jonathan Haidt.

Will it be the same with AI?

To be clear, I subscribe to their views, that the risks of AI are terrifying: deeper inequality, more misinformation, an explosion in synthetic content (even more of it toxic), further degradation of the online experience, and for brands and agencies, further domination and concentration by monopolistic platforms suppressing innovation and competition.

As Cory Doctorow wrote just today: AI is a WMD.

I am presently drafting a paper for the Journal of Current Issues & Research in Advertising to call for research into ‘LLM’s, Brands and Meaning in Advertising.’ I have a point of view, informed by my experience as a statistical modeler. As a practitioner, not an academic, I’ve been scouring both academic and industry sources for insights to test against my own.

In this discovery I found stimulating perspective, again in Joe Mandese’s MediaPost in an interview with Shelly Palmer, Another Canary in the Coal Mine.

Shelly Palmer suggests a positive plausible future for Integrated Marketing Services holding companies.

The data on which LLM’s are trained does not include the proprietary, private, and encrypted data within enterprises. The holding companies have the scale, relationships, trust, and objectivity to access that client-specific history and knowledge and build brand-specific models.

(Pretorius showed an impressive case of an AI trained by WPP on Mondelez brand knowledge and data outperformed a generic chatbot interface.)

The data on which LLM’s are trained is searching for short-term signals rather than longer-term trends that inform business strategies. The holding companies are well-positioned to relate marketplace trends to the client’s own data and use time series to develop strategies and model their potential.

Throughout my career I was motivated by the power of insights and data to connect the client’s business to the trends in the marketplace and the needs of their audiences. This is how we made a difference.

The best definition is that Marketing is responsible for everything outside an enterprise’s physical walls and digital domains. I noted what psychologist Daniel Goleman identified as the critical ability: "pattern recognition, the big-picture thinking that allows leaders to pick out the meaningful trends from a welter of information around them and to think strategically far into the future.”

Pattern Recognition is also my favorite novel by science fiction’s William Gibson, with a plot set in motion by a branding project from a marketing agency.

Marketing in the age of AI remains an exciting, dynamic, and challenging bcareer whether you are an artist or scientist, creator or technologist, and strategist or statistical modeler.

Stewart Pearson

Stewart believes in Consilience, the unity of knowledge across disciplines. He has lived, worked, and traveled globally in Europe, Asia, and the U.S. He settled in the Evergreen State and Seattle. After studying Statistics and Marxist Economics in the U.K. he had four decades of experience in marketing and advertising focused on building client brands directly and globally.?He was Global Chief Client Officer and Vice-Chairman of Wunderman when it was the fastest-growing major agency in WPP. David Ogilvy once sent him a telex from India and Lester Wunderman told him stories of Picasso from the village in France where both of Stewart’s heroes had lived. Stewart is on LinkedIn and Twitter, and at [email protected].