The Costless Commitment

Strategy is choosing what you don't do. A commitment is only a commitment if it costs you something. WPP's new trust principles fail both tests.

WPP quietly published their five trust principles on 22 June, notably the first day of the Cannes festival amongst the many distractions. It's hard to tell why they did this, and why they chose then. My assumption is that CEO Cindy Rose, staring down the barrel of a week of comparative weakness, wanted to spin the strategy they are following due to scarcity, into something that looks more like choice.

Ten days ago, in The Two Juries, I asked which jury is the better futurist, between the one on the Croisette handing WPP its Lions, or the shareholders that have cut the company's value by more than half.

The trust principles are Rose's submission to the forward jury. And the silence that greeted them is not an accident of timing. It is the verdict.

1. Strategy is the art of choosing what you don't do among alternatives genuinely available to you.

That second clause matters. Declining a path you couldn't take anyway isn't strategy. It's description.

2. A commitment is only a commitment if it costs you something.

A promise to keep doing what you already do, at a price you were already paying, is not a commitment. It's a caption.

Hold those two definitions. We're going to run all five principles through a single test: could you profitably do otherwise? If the answer is yes and you're renouncing it anyway, that's a commitment, and it deserves to be believed. If the answer is no, you haven't made a promise. You've made a virtue of your constraints.

Let me make WPP's case before I make mine. Client anxiety about data lock-in and the resulting higher switching costs is real, and the closed-stack model has a genuine trust problem that perceived 'openness' rightly exploits. Nor is this pure vapour: the data bunkers behind these principles are InfoSum's patented federated infrastructure and are an actual technological bet. And the alternative carries its own risk: Publicis's LiveRamp acquisition hasn't closed and brings integration and regulatory questions into 2027. If the agentic era rewards federation over ownership, WPP was pushed onto the right hill.

That said, let's look at the 5 published principles:

Principle one: clients own their data.

WPP's own copy concedes the game as this is, by its own account, how the company has operated for over a decade. Codifying your existing behaviour costs nothing. There is no Epsilon on WPP's balance sheet to renounce, no proprietary identity spine whose monetisation is being sacrificed. The promise is real. The cost is zero.

Principle two: open by design, no lock-in.

The challenger's gospel. Openness is what you preach when lock-in isn't available to you. Android preached it against iOS for a decade (and it's why you happily pay a privacy premium for Apple). If Publicis published this sentence, it would be seismic: it would cost them the entire logic of Epsilon and of LiveRamp. From WPP, it renounces an option the company doesn't hold.

Principle three: consumers are people, not IDs.

The most revealing sentence in the document, because it has an addressee. Growth through data collaboration rather than reliance on personal identifiers is not an abstract philosophy. It is a direct description of the difference between InfoSum's federated model and the identity-graph model Publicis has spent years and billions assembling. To WPP's credit, this principle is backed by real infrastructure. But note what it is: the smaller bet, dressed as the nobler one.

Principle four: AI augments human creativity.

Every agency on earth says this. Moving on.

Principle five: accountability.

Here is where I expected the document to earn its title because this is the one principle that could carry genuine cost. Tie remuneration to client business outcomes: accept downside, share risk, put the fee where the promise is. Read the actual sentence. WPP is "increasingly tying our remuneration to client business outcomes when they wish to do so."

Increasingly. When they wish to do so. The one commitment in the document that would have been expensive has been hedged until it is optional on both sides. Four costless principles and one costly principle carefully drained of its cost. That's the tell.

The oldest unsolved problem in marketing services is attribution. You cannot tie an agency's pay to outcomes unless both parties can agree the agency caused them. Deterministic, ID-based stacks (the model these principles position against) solve exactly that problem. Client and agency each hold their version of the same identified person; when that person's behaviour changes after exposure to the work, there is a shared receipt. Both sides can hang money on it, and increasingly they do.

A probabilistic model produces no such receipt. Strip out deterministic identity and what remains is the headline weather: net revenue movements, brand-equity tracking, net promoter scores. All real, all important but none of them things either party will tie a fee to, because neither side can isolate the agency's contribution from pricing, distribution, weather, or luck.

So WPP cannot unhedge principle five without the very stack that principles one, two and three explain away. It is one capability gap, dressed five different ways. The document renounces deterministic identity as a matter of principle, then discovers (three principles later) that it can no longer price its own accountability.

Now for the timing, because the timing is the story.

In early June, The Coca-Cola Company put its global media, data science and technology business into review. It's a closed review. A two-horse race between the incumbent, WPP Open X, and Publicis. The process is run by MediaSense, began this month, and concludes with a decision expected in late 2026. North America is excluded (Publicis already took that media assignment from WPP in 2025), as are Japan and Korea (Dentsu). Global creative and PR are out of scope and stay with Open X.

Hold the sequence in your head. The review becomes public in early June. The trust principles are published on 22 June, the Monday of Cannes, the week Coca-Cola's executives were on the Croisette, where Publicis was demonstrating its stack to selected clients behind closed doors. A document about why you shouldn't want identity graphs, published in the exact window the industry's biggest advertiser began deciding between an identity-graph model and a federated one.

Read the principles again with that date attached. They are not a philosophy. They are a pitch pre-buttal. And a pre-buttal under the banner of trust, in order to appeal to anxiety, because that is the lever that is left.

And the stakes are brutally asymmetric. Publicis winning Coke would be a prize. WPP losing Coke would be a wound of a different order, because anyone who understands agency economics knows that being a global creative partner without the media assignment is a path to trouble, and a faster one than WPP is already on. The relationship WPP won in 2021 with a dedicated unit of around 5,000 people built to serve it is the scaffolding a creative-only remit cannot support. Why creative-without-media is structurally unsustainable is a newsletter of its own.

There is a scene in Mad Men (season five, an episode called "Dark Shadows") that I think about more than I should. Ginsberg, the brilliant junior copywriter, corners Don Draper in the elevator after Don has pitched his own idea to the client instead of Ginsberg's better one. "I feel bad for you," Ginsberg says. Don, stepping out: "I don't think about you at all."

Devastating. And a lie. Don had spent the entire episode thinking about Ginsberg. He deliberately left Ginsberg's work in the taxi on the way to the pitch. The line was a performance of indifference by a man rattled enough to sabotage a junior.

Which is precisely why the scene is useful here. Attention never shows up in what companies say. It shows up in where their actions and their money point. Publicis's actions (the LiveRamp acquisition, the closed-door demonstrations of Marcel AI at Cannes, the platform economics of its investor story) point at how they are competing with a future version of Google, Meta and Amazon, the players building client-ready tools to skip the agency layer altogether. WPP's document points at Publicis. Coca-Cola's review points at both.

You can reconstruct this industry's real hierarchy from the direction of everyone's attention. Nobody writes principles about the people below them.

I only read about these principles from WPP yesterday, by accident. And I read way too much to have missed any resulting coverage of it. So the market ignored the trust principles for the same reason the forward jury marks WPP down: it prices actions, not adjectives. The one sentence in that document that would have made genuine news (eg: we will tie our remuneration to your outcomes, as standard, starting now) is the one sentence the stack cannot support. Everything else was already true, already free, already priced in.

Cindy Rose's real submission to the forward jury arrives in August, with WPP's H1 results with the first hard answer to the question The Two Juries left open. And the trust verdict that actually matters arrives in late 2026, from Atlanta, and it will not be delivered in a manifesto.

Trust, it turns out, works exactly like strategy. It's not what you publish. It's what you're willing to pay.

So let me ask the question WPP's fifth principle ducks: would you tie your agency's remuneration to outcomes neither of you can measure? And if not then what, exactly, is being trusted?

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 6 July 2026 in the Brandflow newsletter on LinkedIn.