On Thursday, July 30, WPP Media launched a video series about the partners it loves.

That is the actual premise. Not a product announcement, not a research drop. A recurring franchise in which the world's third-largest holding company sits down with the companies it admires and admires them, on camera, at length.

I did not go looking for it. It came to me.

About a dozen people forwarded it to me inside five minutes, which is not a thing that happens with partnership marketing, and nearly every one of them attached the same three letters.

WTF?

No context. No argument. No thread. Just the link and the reaction, over and over, from people who buy and sell media for a living and are not easily startled by anything.

So I watched it.

Two chairs. Two men. A tasteful backdrop that cost more than your quarterly retainer. The lighting is soft in the way lighting is soft when somebody has been paid to make it that way.

Brian Lesser, CEO of WPP Media, opens by explaining that this is a brand-new series in which WPP will highlight its partners and the wonderful work they do together. He then introduces the honored first guest.

Jeff Green. Founder and CEO of The Trade Desk.

Green says he is honored to be first. Lesser says the honor is his. They are, at this point, roughly nine seconds into the video and have already honored each other twice. They get the platitudes out, give each other a massage, and hold hands.

Then, reader, they reminisce about how they met.

I am not editorializing. There is a coffee shop. It is 2009. The Trade Desk does not exist yet and has raised no money, and a young, still-religious Jeff Green just off his latest attempt to convert the heathens, is explaining to a young Brian Lesser why the world needs another DSP. Lesser confesses, on camera, that he did not believe him. There were already so many DSPs.

Which is a lovely detail, except that in 2009 there were not so many DSPs. There were barely any. The term was still finding its feet, and you could count the companies on your fingers with some left over. The objection he remembers having is one people started making around 2013. The meet-cute has been to the gym.

But it lands, because it is supposed to.

Green accepts the confession with the grace of a man who has waited sixteen years and a public listing to hear it. It is a meet-cute about demand-side platforms.

Somewhere out there is a person who watched this on a phone, on a train, and had to look up from it.

From there, the montage. He always believed in agencies. His northstar has not moved in sixteen years. Lesser says he is grateful. Lesser says he trusts The Trade Desk. Lesser says WPP has never had an issue with how they charge, a sentence we are going to return to at some length.

Nobody asks about the fees. Nobody asks about February. Nobody asks anything, because this is not an interview.

It is nineteen minutes of two men agreeing in a room where somebody adjusted the lighting.

Near the end, Lesser says he cannot think of a better partner to kick things off with. Green says he is honored. Again.

Third time. I counted. So much honoring each other.

Somewhere, a brand safety analyst is being made redundant.

One of the dozen went further than WTF.

He works at one of the other holding companies. I am not going to say which, because he would like to remain employed through the weekend. He watched the whole thing and then pointed out something I already half knew but had not assembled properly.

By this spring, there was barely a holding company left with a kind word for The Trade Desk.

Publicis pulled them from its recommended list after an audit. Dentsu stepped away from OpenPath. WPP stepped away from OpenPath. Omnicom went poking at the fees. Five of the largest buyers of programmatic media on the planet, and the score was four unhappy and one about to film a video.

Then he asked the question I have not been able to put back down. Buddy video, or hostage video?

I don't know. But it was... curious, no?

Rewind five months. To a part they’d both like you to forget.

On February 19, Adweek reported that Dentsu and WPP had quietly exited The Trade Desk's OpenPath, the direct-to-publisher buying rail. The reason, per that reporting and the pile-on that followed: hidden fees, and a lack of clarity about where client ads were actually running. Dentsu, which had used OpenPath since launch, reportedly switched it off entirely.

WPP's exit appears to have been more gradual, with later analysis suggesting it had pulled back earlier and never used the product in some markets at all.

Quietly. No press release. No statement of principle. A door closing softly at the end of a long hallway while everyone pretended not to hear it.

The Trade Desk pushed back publicly. Then CMO Ian Colley said on LinkedIn that TTD doesn't push spend to OpenPath.

Then March, which was a bloodbath.

Publicis yanked TTD off its recommended DSP list, telling clients its audit had found fee-stacking, clients auto-enrolled in tools they had not approved, and a master agreement violated. A leaked email went to Publicis clients advising them to stop using the platform.

The Trade Desk denied all of it. They sorta had no choice.

For the avoidance of doubt, and I want to be careful here: failed an audit is Publicis's characterization, the audit was run by a firm Publicis engaged, TTD disputes the findings, and no independent adjudicator has ruled on any of it. Publicis asserts. FirmDecisions concluded. The Trade Desk denies. Pick your priors.

Jeff Green, who does not have an inside voice according to friends and family, published an op-ed on The Trade Desk's own news property accusing the trade press of preferring controversy to truth, and accusing holding companies of manipulating programmatic supply chains to skim undisclosed margin.

Which, and this is what makes the story load-bearing rather than gossipy, may well be correct. Principal media is precisely that practice. It is the subject of live litigation against WPP. The ANA has circled it for years. Green was not inventing a slur to deflect one. He was pointing at something real, from a company that had just been accused of something structurally similar.

Then Omnicom went poking at the fees too.

Three of the Big Five in open or semi-open warfare with the largest independent DSP on earth, all accusing each other of the same category of sin. This was not a disagreement about roadmap priorities.

This was a knife fight about who gets the margin. And every party to it had an interest in the fight ending before anyone looked too hard at their own books.

WPP's Q1 call. An analyst asks CFO Joanne Wilson about the friction. Wilson responds by examining her nails.

She says WPP works with a number of DSPs and SSPs. That decisions get made project by project, client by client.

That sometimes a DSP works well for a client and sometimes it does not. She says transparency twice, the way you say a word when you would like it to do the work of a sentence.

Notice the grammar, because it is doing something specific. Every phrase names a category instead of a quantity. How many DSPs? A number. How much with each? Project by project. It is the linguistic equivalent of turning your body slightly away from somebody at a party.

And then she drops the line anyone who has sat in a competitive briefing will recognize instantly as a knife: The Trade Desk operates in the open internet, which tends to be a smaller segment of the overall advertising market. WPP has elsewhere called that same open internet the long tail, which is not a neutral phrase either.

The long tail is where you put things you do not consider important.

Hold that word. In part two it collides with a number and does not survive the impact.

WPP Media then turns up in a glowing Trade Desk case study about Unilever, Walmart Connect data and streaming TV, in which the word partnership appears with the frequency of a nervous tic.

Digiday notes that WPP's latest partnership is set to push the envelope into programmatic buying. Because that’s what the press release said.

Then, on July 30, the video, in which Lesser reminds Green that they expanded their commercial relationship about a year ago, and that Green called it the most structured partnership The Trade Desk had ever built with a global holding company.

Read that again slowly. The most structured partnership The Trade Desk has ever built with a global holding company.

That is not "a number of DSPs." That is not "project by project, client by client." That is not a smaller segment.

That is a wedding.

And the line about never having an issue with how The Trade Desk charges is, to the syllable, the exact thing Publicis had spent March alleging. One of those two companies is wrong. Neither has been asked which.

Watch what Green actually says, because he is extremely good at this and he is telling you precisely what he wants.

The Trade Desk represents buyers. It owns no inventory. It is not an ad network. It is objective, and objectivity is not a virtue it selected but a structural condition it accepted, because you cannot represent both sides of a trade and be trusted by either.

Every clause of that is true. That is what makes it work.

Then he says: think of us less as a destination and more as a portfolio.

Then he says the platform will fix attribution with a coming product called Open Measurement, because last-click has plagued the open internet for twenty years.

Then he says measurement and audience have to be inseparably connected.

Now run it end to end. The platform executes the buy. The platform shapes the supply path. The platform builds the identity graph, on terms it recently rewrote, moving identity partners from volume-based pay to an incrementality standard the platform itself defines. And the platform now proposes to define what "worked" means, bound inseparably to the audience it also constructs.

Buyer. Optimizer. Scorekeeper. One box. Sold as objectivity.

A lovely arrangement. Also, functionally, grading your own homework, in a class you designed, at a school you built, using a rubric to be published at a later date.

And notice what has quietly happened to the argument. We own no inventory was genuinely powerful in 2009, when the thing worth owning was inventory and the conflict everyone feared was an ad network sitting on both sides of a trade. Sixteen years later the valuable real estate is not the impression. It is the decision about which impression, the identity that makes the decision possible, and the scorecard that says whether the decision was right. None of those require owning a single ad slot.

We are objective because we own no inventory is true the way I own no printing press is true. It answers a question nobody is asking anymore.

I am not saying it is a conspiracy. I am saying it is a business model, an extremely good one, and the man in the other chair is nodding along.

Here is the context WPP would rather you did not lay on top of the hugging.

Cindy Rose took over in September 2025, the first woman to run a major holding company, inheriting a business the market had started describing out loud as buyable. By December, WPP was thrown out of the FTSE 100 after nearly thirty years.

By March, the shares hit their lowest level since 1996.

Since then: Coca-Cola's US media, gone to Publicis. IBM, over after a relationship measured in decades. Adidas, roughly $560m in billings and eight years at WPP, lost to Omnicom, which also handles Nike outside the US and has evidently decided that particular conflict is a them problem. Roughly 4,000 roles cut in a six-month stretch, with several hundred more due by year end.

To be fair, and this belongs in the same breath: WPP Media is also winning. It topped JP Morgan's new-business rankings for Q4 2025 and led COMvergence's global rankings for Q1 2026 with around $1.5bn in net new billings, including Jaguar Land Rover at a reported $475m, Estée Lauder, Wendy's US and SC Johnson. A company can lose Adidas and win Jaguar Land Rover in the same year. Both are true, and anyone telling you only one of them is selling you something.

Then the 2025 results. Revenue down 8.1% to £13.55bn. Reported operating profit down 71.2% to £382m after a £641m goodwill impairment and a further £114m written off property. A loss for the year of £172m, of which £215m was attributable to WPP shareholders, hence the reported diluted loss of 20p a share against earnings of 49.4p the year before. Operating cash flow halved to £724m. Adjusted free cash flow fell from £738m to £202m. Dividend cut 62%.

And the detail that tells you most with the least rhetoric: WPP made pre-tax profit of £131m and paid a tax charge of £303m. An effective rate of 231%, because impairments are generally not deductible. The goodwill writedown hit twice, once through operating profit and again through a tax bill the company could not relieve. A small profit became a loss on the way through the tax line.

A goodwill impairment is a company writing down, in public, that things it bought are worth less than it paid. WPP's own results release names WPP Media as the biggest single drag on margin.

There is also a former GroupM chief investment officer in China serving life imprisonment for commercial bribery tied to rebate flows estimated at $176 million over four years. Two colleagues got fourteen and a half years and four years. The US trade press largely did not pick it up.

That one is not a metaphor either.

So: a company under this much pressure, shedding this many people, insisting this loudly to clients that it is the strategic layer, has simultaneously built the most structured partnership a DSP has ever signed with a holdco.

You do not have to be cynical to wonder what, exactly, got structured.

October 2025. Reuters covers WPP's new self-serve AI platform, Open Pro, under a headline WPP presumably did not write but cannot really dispute: WPP cuts out the agency to help brands create their own ads with AI.

Per Reuters, Open Pro is aimed at smaller brands that do not use full-service agencies. Reuters notes, flatly, that agency groups are under pressure because brands doing their own AI creative and planning pay agencies less. Cindy Rose framed it as expanding the total addressable market.

And it is worse than it sounds, because Open Pro does not merely let a brand make an ad without an agency. It lets them publish, connecting through to WPP's own media tooling for bid optimization and programmatic management, or straight out to the major platforms. The full journey, brief to buy, with the agency removed as a courtesy.

Google was among the pilot users. Google, which is both a WPP client and the company whose ad stack has spent fifteen years demonstrating what happens to intermediaries.

So WPP's answer to being disintermediated is to sell the disintermediation itself. At a discount. As a subscription.

Somewhere in Adland there is a slide reading if we cannibalize ourselves first, technically we win. I would very much like to see that slide. I suspect it has a hockey stick on it.

A former WPP Media employee contacted ADOTAT to say she was let go, and that the function she performed did not evaporate into a model or a cost line. It moved.

Not to AI. Or not only to AI.

It moved to The Trade Desk.

Her framing, as she put it to me, was not that she had been replaced by a machine. It was that she had been replaced by a vendor.

One person's account is one person's account. She is describing her own role and her own team, not a global policy, and she has an obvious interest in an explanation that locates the cause of her layoff outside her own performance.

I would be a poor journalist not to say so out loud. I deal with all of that properly in part three, with every caveat it deserves and an honest statement of what I could and could not corroborate, including where the answer is nothing.

But it points at the question that actually matters, and it is not the question the trade press has spent five months asking.

Everyone has been asking whether WPP and The Trade Desk are fighting.

Wrong question. It was always the wrong question. It treats this as a relationship drama, two companies who had a row and made up, when the temperature between them was never the interesting thing.

Here is the right one.

A brand hires a media agency because it believes it is buying judgment. Judgment about which impressions are worth what. Which supply paths are honest. Which audiences are real. Which measurement can be trusted. And, crucially, judgment exercised on the brand's behalf and against the interests of anybody trying to overcharge it.

That is the entire premise of the agency business.

It is why the model survived programmatic, survived in-housing, survived every eulogy written for it since 1985.

So: when a client writes a check to WPP Media, how much of that judgment is now being formed inside somebody else's platform, and would anyone at that client be told if the answer changed?

On Thursday, August 6, WPP reports its 2026 interim results. Cindy Rose and Joanne Wilson take analyst questions on first-half performance and progress on Elevate28.

The same day, The Trade Desk reports its quarter.

Both of them. Same day. Four days from now.

I am publishing this before either call, deliberately, so that the argument stands or falls on what was knowable this week rather than on whichever narrative each company chooses to wrap around its own numbers on Thursday morning.

I have questions. Judging by my inbox, I am not alone.

ADOTAT has no financial relationship with WPP plc, WPP Media, The Trade Desk, Index Exchange, Omnicom, Publicis Groupe, Dentsu, InfoSum or Google, and holds no position in any of their securities. No named individual or company reviewed this copy before publication, and none was offered the chance to. This post is not sponsored. ADOTAT carries advertising, and any advertising appearing alongside this piece was sold independently of its content; no advertiser was told what this post would say, saw it before publication, or had any say in whether it ran, what it concluded, or who it named.

ADOTAT is otherwise funded by reader subscriptions, which include a tier giving subscribers direct access to me by phone; subscribers include advertisers, agency staff and platform employees, and none of them saw or influenced this reporting, nor does a subscription confer any control over what ADOTAT covers or concludes.

I do not consult for, advise, or accept fees from any company I write about. Nick Manning, quoted here, is a retained expert witness for the plaintiff in Foster v. WPP, which he discloses in his own published work and which is noted again inline. Any correction after publication gets a dated update at the foot of this post and an entry on the ADOTAT corrections page.

The rest is behind the paywall, and so is most of ADOTAT+

What is in the other eighty percent of this piece: the billion-dollar figure that only exists in public because somebody sued. The bigger number underneath it, and what it says about where roughly half of WPP's profit comes from. Both balance sheets side by side, which is when this stops being funny. Why WPP does not simply build its own bidder, and why the honest answer is not weakness. What my inbox from inside WPP actually contains. And the question that ends the whole thing.

What is in the subscription: the documents. Court filings, transcripts, product documentation, and the material people send me that never appears in a post. Not my summary of them. The primary sources, so you can read what I read and tell me I am wrong.

And my phone number.

That is the part nobody advertises, so I will. Subscribers can call me. If you are sitting across a table from a holding company and want to know whether the answer you just got was a real answer. If you are looking at a scope of work and something in it does not smell right. If you are about to sign something and want a second read from somebody with no stake in whether you sign it.

I have spent two decades learning which questions land in this industry and which ones get deflected. That is worth more on a phone call than it is in an article.

What I will not do: work for you, represent you, review your copy, or tell anyone we spoke. I take no money from the companies I cover and I am not going to start taking it indirectly. If you are in a live negotiation with a company I am reporting on, say so at the start and I will tell you what I can and cannot discuss.

It is a thousand dollars a year. If you spend eighty million on media, that is a rounding error against the thing it is protecting. If you spend nothing on media, you are probably not who this is for.

UPGRADE NOW