top of page

Who Owns the Pipes?

Nov 10
12 min read

As advertising's data infrastructure becomes increasingly concentrated, the industry needs to ask an uncomfortable question: Can the companies connecting everyone else's data and media interests ever be truly neutral?


By Rio Longacre. This Signal & Noise exclusive is brought to you by AUDIENCES.


The Data Debate and the Infrastructure Question


The advertising industry has spent the better part of the past decade arguing about who owns the data. We have debated the merits of first-party versus third-party data ad nauseam, opined about the impending death (and subsequent resurrection) of the third-party cookie, and agonized about consumer privacy, identity resolution, data clean rooms, and whether brands should entrust their most valuable customer information to technology companies whose business models depend on monetizing it.

But lost in all that debate is another question we haven't spent nearly enough time asking: Who owns the infrastructure through which all that data flows? Who owns the pipes?


This question became considerably more relevant this past week when Publicis Groupe completed its acquisition of LiveRamp, bringing one of the advertising industry's most important identity and data collaboration platforms under the ownership of one of the world's largest advertising holding companies.


The deal is significant for reasons that extend well beyond its price tag. LiveRamp has spent years positioning itself as an independent connective layer for the advertising ecosystem, helping brands, publishers, agencies, and technology platforms make their customer data interoperable. It never hesitated to use its positioning as the industry's "Switzerland" to justify its dominant market position and high rates. Indeed, its value proposition has always depended, at least in part, on the idea that it could work across competing platforms without favoring any particular participant.

The fact a platform can technically connect to everyone doesn't necessarily mean its owner is economically indifferent to where those connections lead.

Publicis, meanwhile, is not a neutral infrastructure provider. Nor should anyone expect it to be. It is a publicly traded advertising and marketing services company (HoldCo in industry parlance) whose responsibility is to its clients and shareholders. It competes for media budgets, data services, technology engagements, and, increasingly, ownership of the infrastructure underpinning enterprise marketing.


None of this means Publicis will misuse LiveRamp's technology or customer data. They are way too smart to do that. Plus, there are legitimate arguments for combining the two businesses, including the potential to accelerate investment, improve interoperability, and bring sophisticated identity capabilities to more advertisers.


But ownership matters. Incentives matter. And the fact that a platform can technically connect to everyone doesn't necessarily mean its owner is economically indifferent to where those connections lead. For years, advertisers have worried about the walled gardens controlling their media investments. Perhaps it's time we paid closer attention to the companies controlling the pipes between them.


As advertising's most valuable asset shifts from customer data to the infrastructure connecting it, the question is no longer just who owns the data, but who controls the pipes through which it flows.
As advertising's most valuable asset shifts from customer data to the infrastructure connecting it, the question is no longer just who owns the data, but who controls the pipes through which it flows.

When Neutrality Requires a Regulator


What makes the Publicis-LiveRamp deal especially interesting is that these concerns aren't entirely hypothetical. In October, the Federal Trade Commission intervened, requiring commitments designed to preserve neutral, nondiscriminatory access to LiveRamp's services and prevent Publicis from accessing confidential information belonging to LiveRamp's customers. Read that twice.


The FTC's concern was straightforward: ownership of a critical data connectivity platform could give Publicis both the ability and the incentive to disadvantage competing media agencies or steer customers toward its own services. No kidding. That's why they did the transaction in the first place.

To allow the deal to move forward, Publicis ultimately agreed to safeguards that include data firewalls, fair access, interoperability, and independent oversight. Publicis CEO Arthur Sadoun has also explicitly committed to preserving LiveRamp's neutrality. Those are meaningful protections, and they should be acknowledged.

Today, the plumbing is becoming one of the most valuable parts of the house. Publicis isn't alone in recognizing this.

Still, I find the underlying situation concerning. We've reached a point where the advertising industry's data infrastructure has become so strategically important that regulators are imposing conditions to ensure it remains accessible to competitors of the company that owns it.


Think about that for a moment. Back in the day, identity resolution and data connectivity were considered fairly unglamorous, and dare I say boring, pieces of advertising plumbing. Necessary, certainly, but hardly the kind of technology that would determine the competitive structure of the entire industry. Today, by contrast, the plumbing is becoming one of the most valuable parts of the house. Publicis isn't alone in recognizing this.


The FTC's intervention in the acquisition underscores a growing concern: when critical advertising infrastructure becomes concentrated under a single owner, neutrality can no longer be taken for granted.
The FTC's intervention in the acquisition underscores a growing concern: when critical advertising infrastructure becomes concentrated under a single owner, neutrality can no longer be taken for granted.

Infrastructure Is Becoming Strategy


For much of the history of digital advertising, the industry's economic center of gravity was media.

Agencies competed over their ability to plan and buy it, publishers competed to attract audiences, and technology companies competed over their ability to make the buying and selling of impressions more efficient. Data providers competed to make those impressions more valuable. The infrastructure connecting all these participants was important, but it was largely understood as a means to an end.


Today, this distinction is becoming harder to maintain. Identity, data collaboration, cloud infrastructure, clean rooms, activation, and measurement are no longer peripheral capabilities. They increasingly determine which companies can participate in a transaction, what information is available to inform a decision, how performance gets measured, and ultimately, who captures the economic value. The more dependent advertisers become on these systems, the more strategic their ownership becomes.

Consider what happens when a single organization can combine customer identity, third-party data enrichment, media planning, activation, and measurement within a largely integrated commercial offering. It's a new kind of walled garden, if you will.

Identity, data collaboration, cloud infrastructure, clean rooms, activation, and measurement are no longer peripheral capabilities.

This model would bring obvious advantages. Integration can reduce friction, eliminate redundant technology, simplify workflows, and potentially deliver better outcomes. Anyone who has spent time trying to connect an enterprise's fragmented marketing technology stack can appreciate the appeal.

But integration also creates dependencies.


And those dependencies have economic consequences. The company that controls the infrastructure can influence how much it costs to access an audience, which partners can participate in a transaction, and where the resulting revenue flows. Even when the underlying data belongs to the advertiser, the economics of putting that data to work may be determined by someone else. In other words, you can own the asset without controlling the tollbooth.


A brand might own its customer data while depending on someone else's identity graph to make that data addressable. While the brand may own the relationship with the customer, it relies on a third party to determine which media platforms can recognize that customer. It might retain its first-party data in its own cloud environment while depending on an external service to interpret, enrich, and distribute the signals that make the data useful.


In each case, the brand technically owns its information. But how much control does it really have? Data ownership and operational independence are not the same thing. And increasingly, I suspect the distinction between them will become one of the most important questions in enterprise advertising strategy.


As advertising evolves, competitive advantage increasingly comes not just from owning data or media, but from controlling the infrastructure that connects identity, activation, measurement, and AI
As advertising evolves, competitive advantage increasingly comes not just from owning data or media, but from controlling the infrastructure that connects identity, activation, measurement, and AI

The Walled Garden Is an Architectural Model


We typically associate walled gardens with Google, Meta, Amazon, and other large platforms that control enormous amounts of consumer data, advertising inventory, and measurement capabilities.

The trade-off has always been fairly straightforward. Advertisers get access to extraordinary scale, sophisticated optimization, and powerful consumer signals, but they surrender some control over how their campaigns operate and how results are measured. Walled gardens are easy to use and often great at delivering specific outcomes that brands can measure.


Over time, we've seen similar dynamics emerge elsewhere in the marketing ecosystem. Retail media networks increasingly combine transaction data, audience creation, inventory, and measurement. Looking at SaaS, enterprise technology providers are bringing more marketing capabilities directly into their cloud and data environments, while integrating with data sets across the enterprise to maximize impact. Agency holding companies like Publicis, Omnicom, and WPP are investing heavily in proprietary data, identity, technology, and AI.

There's an important difference between being able to connect to another platform and being able to change providers without disrupting your business.

No, these aren't identical business models. Nor do I think it's particularly useful to pretend they are all the same or even indirect competitors. But they share an important characteristic: each is attempting to control more of the infrastructure through which marketing decisions are made. The risk, therefore, isn't necessarily that these companies will refuse to interoperate. Most have compelling commercial reasons to maintain broad connectivity.


The more subtle risk is that interoperability becomes something the infrastructure owner permits or allows, rather than something the customer controls. There's an important difference between being able to connect to another platform and being able to change providers without disrupting your business.

A technology company can offer hundreds of integrations while still making it extraordinarily difficult to leave. It can provide access to competing services while making its own services easier, cheaper, or more effective to use. And it can give customers ownership of their data while retaining control over the identity, intelligence, or decisioning capabilities that make that data valuable.


None of these outcomes is inevitable, but I believe they're worth examining before we congratulate ourselves for having built an open ecosystem simply because the APIs and MCP servers work.


Walled gardens aren't just about controlling data and media. They're increasingly defined by architectures that offer seamless integration while determining how advertisers access, use, and move their customer intelligence.
Walled gardens aren't just about controlling data and media. They're increasingly defined by architectures that offer seamless integration while determining how advertisers access, use, and move their customer intelligence.

AI Raises the Stakes Considerably


Here's where I think this becomes especially interesting. We're entering an era in which AI agents will increasingly participate in marketing decisions that previously required human intervention. This means they'll identify opportunities, select audiences, evaluate inventory, allocate budgets, adjust bids, coordinate creative, and measure performance with increasing independence. Eventually, many of these decisions will happen continuously, with humans establishing objectives and constraints rather than managing every individual action.


In theory, this sounds terrific—at least until you start asking who controls the environment in which those agents operate. Let's not forget, an AI agent is only as independent as the information, tools, permissions, and incentives available to it.


Imagine an enterprise deploying an AI agent to optimize its media investment across multiple channels. The agent has access to customer data, campaign performance, identity services, and a range of media-buying platforms. Now imagine that the infrastructure providing the agent's identity resolution, data enrichment, and connectivity also belongs to an organization that operates its own media-buying business.

The more decisions we delegate to AI, the more important the underlying architecture becomes.

Does this automatically create a conflict of interest? Not necessarily, but it does introduce important questions that become more consequential as autonomous decision-making expands. Can the agent evaluate competing media opportunities on equal terms? Are certain platforms easier to access, and is access the same for all parties? Are preferential rates being offered, and does the infrastructure expose comparable data across destinations? Can the advertiser independently verify the agent's decisions and the economic outcomes they produce?


And the bias doesn't necessarily have to be intentional. An agent operating within an integrated technology environment may naturally favor the services that are easiest to access, have the richest available data, or offer the fastest execution. Over time, architectural convenience can become algorithmic preference, even without anyone explicitly programming the system to favor a particular provide.


Perhaps most importantly, whose objectives is the agent actually optimizing? The advertiser's? The agency's? The platform's? Some combination of all three? For years, we've debated whether programmatic advertising platforms have incentives to direct spending toward inventory that benefits their own businesses. Agentic advertising could reproduce some of those same conflicts with considerably greater sophistication, at much higher speeds, and on an enormous scale.


Looking ahead, the algorithms will certainly become more intelligent, but that doesn't magically eliminate the commercial incentives of the companies that own and operate them. In fact, the more decisions we delegate to AI, the more important the underlying architecture becomes.


A human media buyer can question a recommendation, compare alternatives, or challenge a vendor. An autonomous system may execute thousands of decisions before anyone examines the results. At that point, transparency, auditability, and control over the decision-making environment aren't optional safeguards. They're prerequisites for trusting the system in the first place.


As AI agents take on more advertising decisions, the stakes rise considerably. The real question isn't just how intelligent these systems become, but whose interests they're ultimately playing for.
As AI agents take on more advertising decisions, the stakes rise considerably. The real question isn't just how intelligent these systems become, but whose interests they're ultimately playing for.

Independence Should Be an Architectural Choice


I don't believe the answer is to abandon integrated technology platforms or insist that every component of the marketing stack come from a different vendor. This is not only unfeasible, but would be ridiculous. The concepts of modularity, composability, and integration are valid and generally create enormous value, but the industry has already wasted enormous amounts of time and money trying to connect systems that were never designed to work together.


But given the opportunity to rethink and rearchitect the tech stack in light of agentic transformation, I do think advertisers need to become more deliberate about which capabilities they own, which they rent, and which they allow third parties to control. This is where the architectural conversation becomes more interesting than the familiar debate about where first-party data should live and how it should be activated, especially across paid channels.

Neutrality has to be demonstrated through architecture, governance, and commercial behavior—not simply asserted in a sales presentation.

Increasingly, brands have the ability to retain their customer information within their own cloud environments while using independent services to connect that information to the broader advertising ecosystem. Rather than continually transferring entire customer datasets into proprietary platforms owned by quasi-competitors, they can generate and distribute only the specific, appropriately governed signals required to support advertising decisions.


No, this doesn't eliminate the need for identity resolution, data collaboration, or sophisticated activation technology. Nor does it mean every advertiser should attempt to build those capabilities internally. Rather, it means those capabilities can be organized around the advertiser's own data environment instead of requiring the advertiser to organize its data around a particular technology provider.

That distinction matters. The objective shouldn't be to eliminate dependencies altogether. It should be to make those dependencies visible, manageable, and replaceable without having to rebuild the entire marketing operation.


In practical terms, if an advertiser wants to change agencies, its customer intelligence shouldn't have to move with the agency. If it wants to introduce a new media platform, it shouldn't need to reconstruct its entire identity and data architecture. If it wants to deploy its own AI agents, those agents should be able to access relevant customer context without becoming permanently dependent on one company's proprietary intelligence layer.


This is also why independent connectivity providers have an important role to play. They can help brands make their customer intelligence usable across the ecosystem without requiring them to surrender control of the underlying data or commit to a single commercial stack.


Of course, independence is not automatically a guarantee of neutrality. An independent vendor can have opaque pricing, restrictive contracts, biased integrations, or its own commercial incentives. At the end of the day, neutrality has to be demonstrated through architecture, governance, and commercial behavior—not simply asserted in a sales presentation.


Advertisers don't need to declare war on integrated technology platforms, but they should fight for their independence. The key is building an architecture that preserves control over customer intelligence, enables flexibility, and prevents vendor lock-in.
Advertisers don't need to declare war on integrated technology platforms, but they should fight for their independence. The key is building an architecture that preserves control over customer intelligence, enables flexibility, and prevents vendor lock-in.

Five Questions Every Advertiser Should Be Asking


If I were evaluating a major investment in identity, data collaboration, or advertising infrastructure today, I would spend less time counting integrations and more time understanding the dependencies and potential conflicts of interest the technology creates.


There are five questions I'd ultimately want answered:


  1. Who controls the underlying customer intelligence? Not simply who owns the raw data, but who controls the identity relationships, enriched attributes, decisioning logic, and derived insights that make it valuable.

  2. What happens if we change agencies or technology providers? Can we preserve our customer intelligence, activation workflows, and measurement capabilities, or are they inseparable from the current vendor?

  3. Are all destinations treated equally? Are there differences in pricing, access, functionality, or performance that favor certain commercial relationships? And if so, are those differences transparent?

  4. Can we independently measure outcomes? A system that controls audience creation, activation, and measurement may be convenient, but advertisers should understand how results are calculated and whether they can be validated outside that system.

  5. Who controls the decisions made by AI? As agents become more autonomous, advertisers need visibility into the objectives, constraints, data sources, and economic incentives influencing their recommendations and actions.


These aren't questions designed to disqualify any particular vendor—even LiveRamp. They're simply questions every enterprise should ask before making infrastructure decisions that could shape its marketing operations for the next decade. The answers will influence architecture, procurement, and governance long before the first campaign goes live.


Owning the Data Is No Longer Enough


I think the Publicis-LiveRamp acquisition will ultimately be remembered as part of a broader transition in advertising rather than an epoch-defining event. Let's not forget that for years, the industry's most valuable assets were audiences, inventory, and the ability to reach consumers at scale. Then customer data became the strategic prize. Now we're watching the infrastructure that connects customer intelligence to media, measurement, and increasingly, AI become a major competitive battleground.


Publicis is run by smart people who clearly recognize this opportunity. Its acquisition of LiveRamp brings together substantial identity, data collaboration, and marketing capabilities, and its commitment on paper to preserving LiveRamp's neutrality will be important to the broader ecosystem. The FTC's intervention also sends an important signal: the neutrality of advertising infrastructure can be consequential enough to warrant explicit regulatory safeguards.


But I don't think this is ultimately a story about Publicis or LiveRamp per se. Rather, we're getting a glimpse of the next phase of competition in advertising, one in which controlling the connective infrastructure may become just as valuable as controlling the data or media flowing through it. If this is, in fact, where we are heading—and I would argue we are—advertisers should welcome innovation, cheer consolidation where it creates genuine efficiencies, and embrace technology that makes their increasingly complicated operations easier to manage.


What they shouldn't do, however, is confuse convenience with independence. A brand can own every customer record, maintain its own cloud environment, and invest millions in sophisticated AI capabilities while remaining fundamentally dependent on someone else's infrastructure to make any of it useful.

As more marketing decisions become automated, those dependencies may become harder to recognize and more expensive to unwind.


We've spent twenty years telling advertisers to take ownership of their customer data. Perhaps it's time we asked them to pay closer attention to who owns the pipes.


------------------------------------------------


Rio is an executive with 20+ years at the intersection of strategy consulting, AdTech, data, and media. He's a trusted advisor on customer experience, digital strategy, and marketing transformation. He's a partner at Credera, Omnicom's consulting arm. He's also a podcast host, writer, and public speaker focused on the future of advertising and AI-driven infrastructure.

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page