‘One of our core areas’: Ahead of global agency review, Coca-Cola’s CFO focuses on data matching

He added: “Today, one of our core areas, how do we take the, let’s say, first-party data that we own — it’s proprietary, it’s ours — and how do we marry that with the customer data that these proprietors who are bottlers, given the amount of engagement they have with millions of customers every day. Or how do we work with some of our partners, whether it’s in the U.S. with Publicis or WPP in other parts of the world, to bring that together and then to create a sort of a new intelligence.” 

While it’s hardly a fresh observation, it is a timely one. Coca-Cola is about to kick off a global agency review covering media, data and technology, triggering a tussle between Publicis Groupe and WPP, as Ad Age first reported. It will exclude North America, where the advertiser already works with Publicis as well as Japen and Korea where it works with Dentsu, the company said in an emailed statement. Moreover, it will not include global creative and PR disciplines, which will remain with WPP.

The review, which is being managed by Mediasense, will begin in July, with the decision expected to be announced in the fall. When it lands, it will settle (at least for now) which of the two prevailing holdco views on data a major advertiser buys into. Publicis has spent years building a proprietary data stack across Epsilon, Lotame, LiveRamp and its CoreAI operating system,  and will no doubt pitch itself as a tech company that also does media. WPP, on the other hand, will do the opposite. Rather than owning the data, it acquired InfoSum to connect data sources across the ecosystem without centralizing them, a privacy-safe collaboration model it has since folded into GroupM and WPP Open. 

Put another way, one holdco bets on owning the database, the other on owning the connective tissue. The Coca-Cola review therefore is about as direct a test of those two philosophies as it gets, and one of the more consequential. A Publicis win would be a timely endorsement of its data stack bet, arriving just as it moves to close the LiveRamp acquisition. A WPP win would be a crucial jolt for a holdco that has been searching for consistent momentum after sluggish growth and too many false dawns.

“The Coca-Cola Company is evolving its digital-first marketing operating system for future growth,” reads its statement on the matter. “This includes a shift in mindset from traditional media planning to the emerging ways we need to reach consumers through technology, including agentic tools.”

Whatever it signals about the holdco model, the review could just as easily expose the trap both companies risk walking into. They’ll lead with AI and data because that’s what clients want to hear. But turning combined first-party data into something that actually drives decisions is ultimately a people problem as much as a technology one. The hard question is whether either agency will have the senior talent capable of interpreting what the tools produce.

“To unlock real value from AI, you’ve got to use your own data as a base,” said Robert Webster, founder of AI marketing consultancy TAU. “The real battleground is how you show Coca-Cola how your data, your approach and the AI it plugs into actually matters. Publicis has won a lot of business in Europe, but has been — in my opinion — less AI-focused than perception might suggest. WPP, meanwhile, has very much focused on larger clients, with EssenceMediacom particularly strong on CPG, which again is relevant here. On Europe and the U.K. I’d give WPP a slight edge, on Asia definitely Publicis. But crucially, what a client really wants is to make sure they own their data and their method to execute against it — and neither of them properly delivers that yet. On data ownership, LiveRamp gets Publicis closer.”

That’s the real issue for the largest companies. A Bain & Company survey of 951 companies found the top reason AI programs underperform is that companies still can’t reliably reach their own data after a decade and hundreds of billions spent on modernisation. MIT found the same wall from a different angle: 95% of corporate GenAI pilots stalled, mostly on tools that integrate badly. Through that lens the Coca-Cola review becomes a test of which holdco can actually solve a plumbing problem most enterprises haven’t cracked.