Let us be precise about what investors rewarded on Thursday.

WPP did not win the Coca-Cola global media review.

Coca-Cola has not chosen anyone.

The process is unresolved, the incumbent is still the incumbent it was on Tuesday, and not one dollar of revenue moved in WPP's direction. Reported figures put the share move somewhere between 4.5 and 5.6 percent depending on which session and which outlet you read, and Omnicom fell roughly five percent on losing PepsiCo.

So the market marked WPP up for the disappearance of a competitor.

Which is a strange thing to buy. It is not confidence in the proposition. It is not a bet on the relationship. It is a bet that Coca-Cola now has nobody else to call, and that an advertiser with nobody else to call eventually signs something. The trade got repriced on the absence of an alternative, which is the most honest description of what happened this week that anyone has produced, and it came from a screen rather than a spokesperson.

Hold that thought, because everyone else is telling a different story.

The dominant read across the trades is that Publicis had no choice. Marketing-Interactive describes the withdrawal as a consequence of the two companies' longstanding rivalry. Most coverage treats it as conflict management, an unfortunate collision, a thing that happened to Publicis rather than a thing Publicis did.

ADOTAT reported yesterday that this does not survive the timeline, and nothing since has changed that. A capabilities review with no pitch is not a same-day event. It takes weeks minimum and realistically months, because the client is assessing an operating model rather than a deck. Publicis sat in Coca-Cola pitch meetings across the last two weeks of August. Somebody senior watched both processes run and understood exactly what winning one would do to the other.

Inevitable describes the second half. It says nothing about the first.

This publication is happy to be the outlier on that, but it is worth naming the disagreement plainly rather than writing as though it does not exist. If you believe the withdrawal was forced, you have to explain why Publicis stayed in the room in August. Nobody making the inevitability argument has tried.

Yesterday's piece argued that Coca-Cola's account economics are structurally hostile to the agency holding it, sourced to people with direct experience of the scopes. WPP told ADOTAT directly that the account makes money and that describing it otherwise is not accurate.

Since publication, two named industry figures have said versions of the same thing in public.

Gowthaman Ragothaman, commenting on ADOTAT's reporting, framed Publicis as having walked away from a loss-making arrangement that the eventual winner would inherit, and characterized the choice as picking a better construct with PepsiCo and leaving WPP to bleed. He also volunteered, unprompted, that in his reading none of it turns on the data stack question of InfoSum versus LiveRamp.

Jennifer Hohman, a global growth executive, put it as a question the industry has been avoiding: whether everyone can finally admit that Coca-Cola has been less than fair to agencies.

Neither is a leaked document. Both are named people willing to attach themselves publicly to a proposition that a holding company denied to a reporter in writing this week. The claim is no longer resting on people who cannot be named.

One correction of our own, offered before anyone demands it.

Yesterday's piece said there is no recent precedent for a holding company walking out of a live global review at this scale. Wojciech Ras, who runs a media procurement consultancy and spent years at P&G, says he has seen agencies leave mid-pitch before, and describes what it does to the advertiser running the process: choice narrows, no single agency is best at everything in every market, and a two-agency field leaves the client with almost no fallback position worth having.

His account does not contradict the scale point, and ADOTAT has asked him for specifics. But "no precedent" was a stronger claim than the reporting supported, and the accurate version is narrower: not that it never happens, but that it is close to unheard of on a global review of this size and at this stage of the process.

Ras's substantive point is the more useful one anyway. His view is that this is the moment for Coca-Cola to pull out the wish list, particularly on transparency and accountability, because WPP cannot afford another major loss and Omnicom badly needs a bounce-back. A client with two desperate suppliers has a kind of leverage that a client with two confident ones never gets.

Which is a considerably more optimistic reading than the one the share price implies. One of them is wrong.

Omnicom, which said nothing on Wednesday, has now said it is proud of its PepsiCo work. I guess that’s now paste tense. Was proud.

That is not a fight. It is a posture, and a dignified one, and it is what a company says when it has decided the story is not worth contesting. PepsiCo has helped by describing Omnicom as a critical strategic partner across creative, sports and PR briefs, which softens the picture without altering the fact that a global media lead role held for more than twenty years has moved.

Bank of America went further than anyone in the trades. Its analysts said WPP is well placed to recapture the North American portion it lost to Publicis in 2025, on the theory that a reunified global model beats leaving Coca-Cola split between two holding companies.

That is sell-side opinion, not a Coca-Cola intention, and it should be read as such. But it sits interestingly beside something ADOTAT reported yesterday. This publication has been told that giving up Coca-Cola North America was a condition of the PepsiCo appointment. If that is right, then Bank of America is predicting the second half of a trade it does not appear to know about. Publicis and PepsiCo have both been asked directly and neither has answered.

Which brings us to the argument that makes all of this look small, and it comes from one of the few people who analyzed this review publicly back in June.

Robert Webster, CEO of TAUMS, told ADOTAT on the record that the review was never really an agency contest. "Coca-Cola did not launch this review merely to choose between two media agencies," he said. "It was deciding who would help operate the intelligence system connecting media, data and technology."

On what a one-bidder process can produce, he is unsparing. An uncontested retention can still deliver a good contract and even a materially different operating model. "What it cannot do is demonstrate that WPP's proposition defeated the alternative." Coca-Cola retains the ability to use MediaSense and the review to change architecture, commercial terms, audit rights and ownership. But if the supplier stays and the dependency stays, then, in his phrase, "it is an incumbent renewal dressed as transformation."

Then the part that should outlive this news cycle entirely.

Media does not simply spend money, it manufactures context. Webster's list of it: "the prices actually paid, inventory quality, audience and identity matches, creative response, experiments, marginal performance curves, measurement assumptions and the history of every optimisation decision." Run that through an AI system and, in his words, "that accumulated context, combined with AI, becomes the marketing brain."

From which follows the sentence that ought to be on the wall of every procurement department in this industry: "The advertiser should own what its own money taught the machine."

His test for whether that is happening is refreshingly free of consultant fog. Can Coca-Cola retain and move its client-specific memory, its performance curves, its workflows, its decision history, its audit trail? Can another partner run the system without starting over? If not, "Coca-Cola owns its data but still rents its brain."

And the standard, stated as a single line: "The agency should be replaceable without the client suffering a lobotomy."

Webster declares his interest unprompted, which is more than most manage. TAU builds advertiser-controlled intelligence infrastructure. "But it is also the standard I apply to TAU," he says. "A client should be able to replace us and retain the intelligence created with its data, activity and money." A vendor who volunteers the test that would disqualify his own company is a vendor whose test you can probably use.

To WPP's credit, and Webster gives it, the company has publicly committed to the right promise: an intelligence layer that can sit in the client's environment, client control of data and derived insights, architecture built to interoperate rather than trap. He calls that "the correct promise" and updates his June position to say so. He then asks whether Coca-Cola gets it "in the contract and in the plumbing" or only in the deck.

On Publicis holding PepsiCo globally while still holding Coca-Cola in North America, the industry's answer has been separate teams and NDAs, delivered with a shrug.

Webster's point is that the reassurance has expired. "Separate teams and NDAs were designed for the people era," he says. In an AI system you need demonstrable separation across "permissions, data stores, retrieval systems, memory, derived features, model tuning, logs, evaluation data and client-specific outputs."

His conclusion: "The wall has to exist in code and contract, not only on an organisation chart."

Which is exactly right, and slightly funny, because an org chart is a diagram of who is not supposed to speak to whom at lunch. It has no view whatsoever on what a retrieval index is doing at three in the morning.

He is careful, and so is this publication. He has seen no evidence that Publicis is combining confidential Coca-Cola and PepsiCo information. Neither has ADOTAT, and nobody here is suggesting it. He notes that Publicis describes CoreAI as a central group intelligence system built on its own proprietary data, is explicit that this is not evidence of anything improper, and says it makes the boundary between group intelligence and client-specific intelligence "a legitimate and important due-diligence question."

Nobody asked it this week. A holding company pulled both of the world's largest beverage advertisers into overlapping orbit and the industry's collective response was to speculate about hurt feelings.

The most original observation of the week came from a commenter rather than a columnist. Alexander Nempeque, whose background spans Publicis, Coca-Cola and Warner Bros Discovery, pointed out that the capabilities review is the actual story.

If global accounts start moving through capability reviews instead of classic pitches, competitive advantage shifts upstream to operating model, data, technology and the ability to integrate them at scale. The agency stops being judged on what it promises in a room and starts being judged on the system it has already built.

Think about what that does to everyone. It rewards whoever spent the last five years buying infrastructure. It devalues the entire pitch apparatus the industry has organized itself around. And it makes the winner harder to displace, because you cannot out-present a system, you can only out-build one.

Which is why the ownership question Webster is asking is not a governance footnote. If accounts are now won on infrastructure, then exit rights are the only leverage a client has left, and most of them have not noticed yet.

Whether Coca-Cola reopens the field or concludes quietly. The share price is a bet on the second.

Whether Publicis holds Coca-Cola North America. Bank of America thinks WPP takes it. ADOTAT has been told Publicis agreed to give it up. Those two predictions are compatible in a way that should make Publicis uncomfortable.

And whether "One PepsiCo" turns out to be an advertiser-controlled operating system or a more concentrated dependence on one holding company's stack. That is the question that determines whether this week was a transformation or a very well-executed lock-in, and the announcement does not answer it.

Webster's summary needs no MBA. Own the data. Own the method. Own the evidence. Anything else is renting part of the company's brain.

How we reported this: on-record interviews, named public commentary, COMvergence billings data, and the companies' own published statements. WPP disputes the characterization of the Coca-Cola account as unprofitable, telling ADOTAT the account makes money. Dollar figures vary between outlets and are approximate. Unconfirmed material is labelled as such and was put to the companies concerned before publication. All companies named were contacted; refusals are reported as refusals. ADOTAT has no financial relationship with any company named here, none reviewed this piece before publication, and ADOTAT is funded by subscriptions and reader support.

Everything above is what I can prove. ADOTAT+ is where I tell you what WPP said when I put it to them, what the scopes actually look like, what I was told and couldn't stand up, and the four things I still don't know.

Nobody else in this business publishes their own open loops.

There's a reason for that.