The principal problem: Don’t measure what you sell

The WPP whistleblower case has put principal media back in the headlines. But the real issue isn’t that agencies make margin on media — it’s that the same players now own the marketplace, the media, and increasingly the measurement. That’s a structural conflict. And it matters more now than it ever has, because the numbers it produces are the ones marketing is taking to the boardroom.

Most people in our industry have heard that WPP is in trouble. Fewer have paid close attention to the specific mechanism at the centre of it — and I think the mechanism is the more important story.

A whistleblower lawsuit filed by former GroupM executive Richard Foster alleges that WPP’s media buying arm was generating close to one billion dollars a year in undisclosed income from principal media trading. Not disclosed to clients. Not passed back as savings. Recorded internally as “non-product related income” — with a 15% year-on-year growth target attached to it.

When WPP filed court documents in response, they inadvertently surfaced data that gives the case its sharpest edge: 97.4% of their proprietary inventory was not being used by their largest clients. The product existed primarily for the agency, not for the advertiser. I’m not writing this to pile on. I’m writing it because it perfectly illustrates the structural tension that exists right across our industry — and because that tension is only going to become more consequential.

To understand why, you first need to understand what principal media actually is, and why it exists.

From agent to reseller

Principal media is when an agency stops acting as an agent for a client and starts acting as a vendor. The agency buys inventory in bulk — at a significant discount — and resells it to clients at a margin. Think of it like buying options on media, or buying wholesale to sell retail. The agency takes a position in the inventory; the client gets a lower effective rate than they’d find on rate card; and the agency captures the spread between what it paid and what it charged.

The economics make sense. Agency margins are among the second-lowest of any professional services sector. Procurement teams have spent fifteen years squeezing fees. If you’re running a media business and your advisory revenue has been compressed to almost nothing, finding another way to monetise your buying power is a rational response. The ANA’s 2024 report on principal media found advertiser cost savings of 10–15%. For many CFOs, that’s the end of the conversation.

My view is that principal media should be allowed. Procurement pressure is real, fee compression is real, and these products are a rational market response. Agencies have to make money somewhere. The question isn’t whether principal media exists — it’s where the conflicts it creates become intolerable.

The problem isn’t that agencies make a margin on the media. It’s what happens when the people trading the media also own the marketplace — or control the measurement.

Conflict one: Owning the market you’re trading in

Imagine a hedge fund that also owned the ASX. The fund decides what gets listed. It sets the order in which inventory is displayed. It determines the routing of trades. And it’s simultaneously placing bets on outcomes using its own capital. Nobody would accept that architecture in financial markets. We have something close to it in media.

The US Department of Justice antitrust ruling against Google in April 2025 essentially validated this argument in court. Judge Brinkema found that Google had “unlawfully monopolised the publisher ad server and the ad exchange markets.” The ruling identified the structural problem precisely: when a single entity controls the buy-side demand platform, the exchange where inventory is matched, and the sell-side server, it cannot be a neutral market. Internal documents cited in the ruling showed the exchange was extracting take rates that could only be sustained because of that structural control. The ruling doesn’t speak to intent — it speaks to architecture.

This is why The Trade Desk’s model is so important as a reference point. TTD owns no media inventory. It doesn’t have a publisher side. It has no interest in routing demand toward any particular media inventory because it has no financial stake in what gets bought. Its independence is the product — and in a market where structural conflicts are becoming more visible, that independence becomes more valuable with each passing year, not less. Crucially, that same freedom from inventory is what makes a DSP like TTD a uniquely credible participant in measurement too. Because they have no inventory to protect, no routing incentive to disguise, and no conflict of interest to manage, their position in the ecosystem is structurally cleaner than any integrated player can ever be. Independence from media ownership isn’t a constraint — it’s the foundation of trust across the entire value chain.

The point isn’t that every integrated player is acting in bad faith. It’s that the architecture creates incentives that are very hard to resist — and very hard for advertisers to audit. That’s the structural problem, and it has nothing to do with the intentions of the people involved.

Conflict two: Measuring what you sell

The second conflict is in some ways more insidious, because it’s the one that touches the numbers marketing takes to the boardroom.

Think about what we ask of financial markets. The seller of a stock cannot write the company’s own financial statements. If a company could report to the market that revenue was $X when it was really $Y, that would be objectively bad — and every governance structure in capital markets exists to prevent exactly this. We have independent auditors. Ratings agencies. Regulators. The whole infrastructure of market trust depends on the separation of selling and measuring.

The author Henry Innis

In media, that separation has been quietly eroding. Many major inventory owners now offer lift studies, brand measurement tools, and media mix modelling programs as part of their commercial offering. I want to be careful here, because some of what’s been done is genuinely valuable. The release of open-source modelling frameworks — where platforms make their research methodology publicly available for independent measurement vendors to scrutinise and build on — is exactly the right behaviour. MMM APIs and MMM Data Platforms, which allow third-party measurement providers to integrate directly, are a good example of media inventory owners making themselves easier to measure independently. Many, if not all, platforms have been relatively open about making data accessible to measurement partners. And in Australia, major TV networks have been doing genuinely interesting work on experimentation inventory — making bonus inventory available to support large advertisers running properly structured tests. This is what a media inventory owner acting in the ecosystem’s long-term interest looks like.

The concern isn’t the tools. The concern is when those tools become commercial partnerships. If the measurement programs evolve to operate as commercial partnerships with an inventory owner — when the entity selling you the media also becomes, even if by proxy, the entity running your measurement program — the independence of that measurement is structurally compromised, regardless of how good the underlying methodology is. The tool might be excellent. But the commercial relationship has changed what it’s optimising for. This isn’t widespread yet, but I can see a world where measurement may evolve quickly in this direction.

This is the connection I think the industry hasn’t fully joined up yet. Principal media and measurement aren’t separate problems. They’re often linked to the same problem of ensuring we are driving strong outcomes for our shared clients.

MMM isn’t the casualty here — it’s the solution

I want to make this point clearly, because I think the conventional framing gets it backwards. People often talk about principal media arrangements and measurement being in conflict. I think independent MMM is actually what makes principal media arrangements work — for both sides.

If you’re an agency entering a principal agreement, the commercial interest you’re taking creates a legitimate question from your client: are you recommending this because it’s right for my business, or because it’s right for your margin? The only honest answer to that question is measurement that neither party controls. An independent MMM that shows the principal inventory performing against business outcomes is the mechanism of trust. It’s how the agency demonstrates its principal position is delivering, and how the client validates that the arrangement is worth maintaining. Without that independent layer, the agreement is running on faith — and faith has a limited shelf life in procurement reviews.

The same logic applies to media inventory owners. If your inventory is genuinely performing, you should want independent measurement. It’s the thing that proves the case. The inventory owners who are resisting independent measurement are, almost by definition, signalling something about their confidence in their own product’s performance.

A bad product tries to control how it’s measured. A good product tries to perform brilliantly against outcomes. Independent MMM is how you tell which is which.

Why this matters more now

None of this is entirely new. The ANA published its first major report on media transparency in 2016. The industry said it would clean itself up.

What’s changed is the context. Most major brands are not growing fast right now. When you’re growing 15% a year, imprecise measurement is an irritant. When you’re growing 2%, it’s existential. You need to know exactly what is and isn’t working, because every budget decision is load-bearing.

Marketing is also, finally, earning back a seat at the boardroom table. That’s a hard-won gain, and it depends entirely on the credibility of the numbers we bring into the room. If those numbers are produced by measurement programs that have a commercial relationship with the media being measured, the credibility of marketing’s case to the board is only as strong as that relationship is independent. Which is to say: not very.

This is the long-term risk that I don’t think gets enough attention. It’s not that the numbers are necessarily wrong. It’s that when they’re eventually interrogated — by a CFO, a board, an auditor, or a new CMO asking uncomfortable questions — the conflict of interest will be the first thing on the table. And once it’s on the table, all the numbers come with it.

What should actually happen

• As advertisers: Do not buy measurement from organisations trading principal media in the same category. Keep those two relationships structurally separate.
• As procurement: Require all measurement vendors to disclose third-party commercial agreements transparently — including any partnership with media inventory owners that could influence methodology or output. Make this a standard clause.
• As agencies: Elevate independent third-party measurement as a non-negotiable condition of any principal media strategy. It’s the mechanism that preserves trust on both sides of the agreement. Build it into the contract, not as an afterthought.
• As media inventory owners: Invest in making yourself easier to measure independently — open APIs, portable data, research published for independent scrutiny. This is the model that builds long-term relationships. Stop building measurement programs that create commercial dependencies and start making your ad product easy to validate against outcomes.
• As DSPs: Your independence from media inventory is your most valuable structural asset — in buying, and increasingly in measurement. A DSP with no inventory to protect has no conflict to manage, and that positions you as a trusted partner across the whole value chain in a way that integrated players simply cannot match. Lean into it.

The long game

Structural arrangements across the industry will be scrutinised or they won’t. But the underlying dynamic — where the same players own the inventory, the marketplace, and the measurement — must change. And the longer it persists, the more fragile the trust that holds the whole ecosystem together becomes.

The path forward isn’t a heavily regulated one — it’s a reputational one. Brands and agencies have enormous collective leverage here if they choose to use it. An industry that insists on independent measurement, transparent marketplace architecture, and disclosed conflicts will outcompete one that doesn’t. The logic is astoundingly simple. The hard part is choosing to act on it.

The good news is that the path I’m describing isn’t anti-platform. It’s pro-ecosystem. The inventory owners making their data portable, building open APIs, publishing their methodology for independent scrutiny — those are the ones building durable relationships. That’s the version of this industry I’d like to see more of.

A bad product tries to control how it’s measured. A good product tries to perform brilliantly against outcomes.

That’s the test. It’s time more of us started applying it consistently.

Related Content

Analysis SaaSpocalypse Now: The software inferno will birth an intelligence era

Analysis All about the cash: Why the Laundys bought Nine Radio