Publicis wins PepsiCo global media without a pitch

Publicis just took PepsiCo's global media account, about $1.7 billion in annual spend, away from OMD after 25 years. No pitch. No formal review. On the July earnings call, Publicis CEO Arthur Sadoun told analysts he deliberately skips reviews he expects to be decided on price, winning instead on C-suite relationships and the Epsilon first-party data platform. He ran the same play on Microsoft, Paramount, and LVMH in the past year.

That pattern is reshaping how large media gets bought across the holdco market.

Why Publicis keeps winning without a pitch

When a client's customer data sits inside Epsilon, pulling the media account means ripping out the data activation layer too. That switching cost is real enough that procurement never opens the review, which means rivals never get the at-bat. The agency that holds the client's data infrastructure is becoming very hard to fire.

That pressure runs in the opposite direction for WPP and Dentsu, both mid-restructure around traditional agency models with no comparable first-party data asset. Omnicom's acquisition of IPG was partly a scale defense against this dynamic, and PepsiCo walking anyway says scale alone doesn't hold a client when a rival is wired into their martech stack.

The limits

The $1.7 billion is gross billings, not revenue. Net of losing Coca-Cola North America to category conflict, the real gain is closer to $295 million. That Coke departure names the hard ceiling on consolidation: blue-chip food and beverage clients will not give up category exclusivity, no matter how good the stack. You can't buy for two rivals in the same category, and this deal shows exactly where that line sits.

The no-pitch model also runs on relationships, and relationships are fragile. One new CFO at PepsiCo can reverse the decision. Whether this is a structural shift or a favorable cycle dressed in impressive gross-billing numbers depends on whether WPP, Dentsu, and the merged Omnicom-IPG can build credible first-party data offerings in the next 18 months. If they can, Sadoun's no-pitch rate mean-reverts. If they can't, brands keep consolidating with whoever holds their data and the pitch-driven agency model slowly loses ground.

Who takes the immediate hit

OMD's planning and investment desks absorb the loss first. PepsiCo likely represents 8 to 12 percent of a major office's billings, so headcount restructuring is coming inside 90 days. The second-order problem is worse for Omnicom: this lands exactly when it needs stable optics for the IPG integration.

For identity, measurement, and clean-room vendors selling into these agencies, the win reprices the buyer map. When a holdco wins media because it owns the client's first-party data, that holdco's own stack becomes the default rail. Client-specified vendors survive consolidation. Agency-preferred vendors get absorbed into the pipeline.

Our call: WPP or Dentsu will lose at least one account of $500 million-plus in annual billings to Publicis or the merged Omnicom-IPG, without a formal competitive review, by the February or March 2027 earnings calls. Confidence: medium. The mechanism is real; the timing and specific account are not. We're wrong if every large account change in that window runs through a competitive pitch, or if WPP and Dentsu hold their major accounts intact. The next large account loss at either holdco will tell us whether the no-pitch pattern is structural or still a cycle.

Publicis just took PepsiCo's global media, about $1.7 billion in annual spend, away from OMD after 25 years. No pitch. No formal review. Arthur Sadoun told analysts on the July earnings call that Publicis now skips reviews it expects to be decided on price, and wins on C-suite relationships and its data stack instead. Same move it ran on Microsoft, Paramount, and LVMH in the past year.

What's actually being decided here isn't one account. It's whether the way big media gets bought is changing under the whole holdco business, and what that means for anyone who sells data, identity, or measurement into these agencies. This is easy to undo for any single client (a new CFO can reverse it), but hard to undo as a market pattern once procurement stops running the process.

The Market Analyst

Two holding companies are drifting apart in how the market values them. Publicis owns Epsilon, a first-party data business, and sells media on top of it. WPP and Dentsu still sell mostly traditional agency services and are mid-restructure. When a client consolidates with whoever runs their data plumbing, the slide-deck agencies lose the room before the pitch starts. For an informed outsider: the agency that holds the client's customer data is becoming impossible to fire without ripping out infrastructure. Omnicom bought IPG partly for scale defense, and PepsiCo walking says scale alone doesn't hold a client when a rival is wired into their martech. Long the data-owning holdco. Short the pure media trader.

The Skeptic

The $1.7 billion is gross billings, what flows through, not what Publicis keeps. The real revenue is a slice of that, and the net gain after losing Coca-Cola North America to category conflict is roughly $295 million. Sadoun saying he "chooses not to pitch" on price reviews is partly a nice frame for losing the cheap, commoditized accounts. And the Coke departure says plainly what blue-chip food and beverage clients will not give up: category exclusivity, which puts a hard ceiling on how much any one holdco can consolidate. No-pitch wins run on relationships. Relationships are fragile. One CFO change at PepsiCo and the streak reverses. This looks like a moat. It may just be a good year.

The Operator

The pain lands Tuesday morning at OMD's planning and investment desks. PepsiCo is probably 8 to 12 percent of a major office's billings, so headcount restructuring inside 90 days while transition teams stand up at Publicis. The second-order hit is worse for Omnicom: this lands right as it needs stability optics for the IPG integration. Media directors at WPP and Dentsu now face an ugly internal question. Why didn't your relationship map catch the PepsiCo signal a year out? No-pitch wins compress the window an incumbent gets to defend itself. If procurement never opens a review, you never get the at-bat.

The Customer / End User (the CMO and CFO on the client side)

The buyer here isn't renewing a media contract. They're deciding not to run a bake-off at all. That only happens when switching costs feel real. Pull media out of Publicis and you're also pulling your data activation out of Epsilon, and that hurts. But be clear about what the client is actually asking for: fewer procurement cycles and one throat to choke on data and media together. They are not asking for the best CPM. That's exactly why Coca-Cola conflict still bites. Category exclusivity is the one thing a blue-chip client won't trade away, no matter how good the stack.

The CFO (of a data or identity vendor selling into these agencies)

If you sell identity, measurement, or clean-room tech, this reprices your buyer map. When a holdco wins media because it owns the client's first-party data, the holdco's own stack (Epsilon here) becomes the default rail, and your product gets squeezed to a feature inside their pipeline. Client-specified vendors survive consolidation. Agency-preferred vendors get absorbed. So the question for your renewal isn't your tech. It's whether you're specified by the brand or resold by the agency.

Where the council splits

Two real disagreements. First, moat or streak. The Market Analyst and the strategist read three big wins in a year as a structural shift toward data-owning agencies. The Skeptic reads a favorable cycle dressed in gross-billing numbers, cappable by category conflict. Both can point at the same PepsiCo win.

Second, how far this actually goes. Everyone agrees data ownership bypasses procurement. Nobody has a clean answer to the Coke problem: if winning PepsiCo costs you Coca-Cola, the data moat doesn't beat the oldest rule in the business, which is that you can't buy for two rivals in the same category. Consolidation has a hard ceiling and this deal shows exactly where it sits.

What it hinges on

Whether competitors can build a comparable data wedge before the relationship advantage compounds. If WPP, Dentsu, and the merged Omnicom stand up credible first-party data offers in the next 18 months, Sadoun's no-pitch rate mean-reverts and this was a cycle. If they can't, brands keep consolidating with whoever holds their data, and the deck-driven agency slowly loses the right to compete. The council leans toward this being structural on the data point, but capped hard by category conflict. Both things are true at once.

The next major WPP or Dentsu account loss will clarify which force is winning. If it goes to a data-owning rival without a review, the pattern is real. If it goes to a cheaper trader after a pitch, this is still a cycle.

Prediction: WPP or Dentsu will lose at least one account of $500 million-plus in annual billings to Publicis or the merged Omnicom-IPG without a formal competitive review by the 2026 full-year earnings calls in February or March 2027.

Confidence: Medium. The mechanism is real, but the timing and which specific account moves is hard to pin.

Why: Publicis has now taken Microsoft, Paramount, LVMH APAC, and PepsiCo without pitches, and Sadoun told analysts on the July earnings call he is deliberately skipping price-driven reviews in favor of C-suite relationships plus a data stack the client can't easily rip out. That only works because clients with their first-party data sitting in Epsilon face real switching costs, so procurement never opens the door for a rival to compete on price. The same mechanic runs in reverse on WPP and Dentsu, who are still restructuring around traditional agency models and have no comparable data asset to lock a client in. The opposite outcome, WPP or Dentsu holding every large account through a clean open review, requires those clients to voluntarily reopen a process the winning side has an incentive to keep closed, which is the less likely path given four no-pitch wins already on the board.

Revisit by 2027-03-31: We're right if a $500M-plus account moves from WPP or Dentsu to Publicis or Omnicom-IPG with no formal review disclosed by the February or March 2027 earnings calls. We're wrong if every large account change in that window runs through a competitive pitch, or WPP and Dentsu hold their major accounts intact.