Arthur Sadoun, freshly installed as CEO of Publicis Groupe, made two decisions that week that would define the next decade. First, he announced Publicis would skip the awards circus entirely in 2018, redirecting roughly €50 million toward building Marcel, an AI-powered internal platform, in partnership with Microsoft. The industry laughed. Second, and less visibly, he accelerated the Power of One he had architected with the indomitable Maurice Levy before taking the reins himself. Doubling down on the cultural integration that would transform Publicis from a holding company into an operating company. From silos into a platform. The industry shrugged.
Eight years later, nobody is laughing. Last week, Publicis reported another set of formidable results: full-year organic growth of 5.6%, Q4 acceleration to 5.9%, an industry-leading operating margin of 18.2%, and free cash flow surging 10.6% to €2bn. For the sixth consecutive year, Publicis topped the new business rankings. These numbers are not incremental improvements. They are the compound returns of strategic moats that took a decade and some $14bn in data and technology investment to build.
I spent 13 years at Publicis, seven of them on the management committee. I watched these moats being dug and fortified, and I can tell you they are real, substantial, and extremely difficult for competitors to replicate. There are five that matter.
First, culture. Publicis operates as a company, not a holding company. That distinction sounds semantic until you try to get a creative agency, a media agency, and a technology consultancy to share a client relationship without bloodshed. This way of working took five years of grinding cultural transformation, breaking down P&Ls, restructuring incentives, and replacing leaders who wouldn’t adapt. Most competitors are still federations of warring fiefdoms dressed up with shared letterheads.
Second, a genuine technology edge. Roughly one in six Publicis employees is an engineer. That ratio would be unremarkable at a technology company but is extraordinary for an organisation that most investors still categorise as an advertising group.
Third, Epsilon’s first-party data asset, with some four billion consumer profiles globally, provides a critical advantage in a world where platforms are actively working to disintermediate agencies from the data layer. Fourth, Marcel and CoreAI are real AI platforms delivering operational value, not vapourware dressed up for earnings calls. And fifth, media scale: even after the Omnicom-IPG merger, which amounts to crude arithmetic while two cultures learn to coexist, Publicis maintains the number one position in practice.
These moats explain why Publicis is less affected by what I have been calling the Great Compression, naming the simultaneous squeeze of declining budgets, platform disintermediation, AI automation, talent attrition, and margin pressure that is reshaping the entire marketing services industry. Other holding companies are scrambling to articulate their AI strategy. Publicis has been executing one for eight years.
Clients can see the difference. Six consecutive years atop the new business rankings is not luck or a cyclical advantage. It is clients voting with their budgets for a partner they trust to navigate transformation alongside them. When a CMO consolidates a multi-billion-dollar account into Publicis, they are not buying a cheaper media plan. They are buying the integrated capability of data, technology, creative, and media working as a single system. They know Publicis is different, better, and increasingly indispensable.
The problem is that investors do not see the same gap to the market that clients do.
When Publicis reported those excellent results last week, the stock dropped nearly 9% (albeit in a soft week for companies deemed to face AI risk). This was not a reaction to weakness; it was investors lazily applying the same categorical discount they apply to every company with “agency” in its analyst coverage. The marketing services sector trades at roughly 10.6 times forward earnings, against the S&P 500’s 20.5 times. That is a 48% discount applied uniformly to an industry where the distance between best and worst has never been greater.
Publicis is, by a wide margin, the best house on the street. It has renovated every room, modernised the plumbing, and added an extension. The problem is the street. No amount of interior renovation will change the postcode.
Last week, Publicis unveiled its updated strategic positioning: the ambition to become clients’ Most Valuable Partner. I understand the intent, and I imagine that it plays well with CMOs and procurement teams, which is where Arthur directs every restless, focused, waking minute. But “MVP” is generic. Accenture could claim it. Salesforce could claim it. Its caterer may even claim it. It does nothing to escape the gravitational pull of the agency category in investors’ minds. I preferred the earlier framing of “win in the platform world” because at least it named an enemy and staked out territory that traditional agencies could not claim.
For Publicis to get the investor recognition that its performance demands, three things need to happen.
First, escape the AI-as-efficiency trap. Right now, the dominant narrative around AI in marketing services is faster execution at lower cost. And clients, reasonably, expect those savings to be passed back. Publicis needs to reframe its AI story toward the bluer ocean of intelligent creativity and human insight. Epsilon is extraordinary for downstream precision, but how does that data asset improve upstream idea generation? How does it unlock cultural insight that informs better creative and braver innovation? Future acquisitions should target capabilities that move Publicis up the value chain, not further along the execution layer that platforms are building to own.
Second, rebalance leadership energy. Arthur is a force of nature with clients, and it shows in the results. But Publicis is a communications company that somehow under-communicates its own story externally. Underperforming competitors seem to consume a disproportionate share of industry oxygen with restructuring announcements and AI press releases, while the top performer simply gets on with the work. I wish that were enough. That quiet confidence is admirable, but it is not enough when investors are making allocation decisions based on category perception rather than fundamental analysis.
Third, put as much energy, love, and creativity into positioning as they do for clients. Most Valuable Partner is a promise any large professional services firm could make. Publicis needs a positioning that is specific, falsifiable, and exclusive…one that makes clear it is no longer competing in the same category as the companies dragging down its multiple, but if necessary, is now in a category of one. The company needs to articulate which addressable markets it is entering that carry different valuations, and why its data, technology, and cultural integration make it the only credible player to do so.
None of this is criticism; I have deep admiration for the company and its leaders. This is the constructive nudge of someone who spent over a decade watching Publicis transform itself from the inside, who respects what Arthur and his team have achieved enormously, and who believes the company’s clients, people, and investors deserve the recognition that the numbers have long since earned. The strategic work is done. The moats are built. The results are undeniable.
What remains is a perception problem masquerading as a valuation problem. And Publicis, of all companies, should know that perception is its business.
The renovation is complete. The house is immaculate. It is time to move neighborhood.
