Saving the Open Web: Why We Need to Stop Selling Programmatic Hotdogs

A blueprint built on multiple conversations at Ad Week NYC. In detail with Scott Messer and Fern Potter . Data from a Jounce Media report.

The open web is dying. But there are enough people with the talent and passion to save it—if we have the collective will to act.

This isn't hyperbole or industry melodrama. It's mathematical reality backed by brutal market data.

The Problem: A Double Whammy

The open web faces twin existential threats that are compounding into a death spiral.

AI and the Traffic Apocalypse

First, artificial intelligence is eviscerating publisher traffic. Publishers have seen organic traffic decline by 25% over the past year, with over 600 million lost visits as users increasingly turn to AI chatbots and Google's AI Overviews for answers. Zero-click searches have surged from 56% to 69% between May 2024 and May 2025, meaning the majority of searches never result in a click to a publisher's site.

Matthew Prince, CEO of Cloudflare, has been tracking this collapse with frightening precision. Ten years ago, for every two pages Google crawled, it sent one visitor back to the original site. That ratio has now plummeted to 18:1. With AI Overviews, Prince estimates that up to 90% of searches now end without a single click to an external website. The compensation model that built the web—traffic driving ad revenue—is simply breaking down.

The revenue implications are catastrophic. As one industry expert bluntly put it: "Audience and pageviews will decline and evolve into secular declines. CPMs will increase modestly, but not enough to offset ad revenue losses". Lower traffic plus modest CPM increases equals an unsustainable business model.

But here's where it gets truly perverse: AI companies are extracting enormous value from publisher content while paying nothing for it. They train their models on the web's content, then answer questions directly, cutting publishers out of the value chain entirely. As Prince argues, "If we don't fix this, the internet will die." AI systems derive their value from content that publishers created and paid to produce, yet the incentive to create that content is evaporating.

Walled Gardens Eating the World

Second, advertiser spend is flowing inexorably toward walled gardens. According to Jounce Media's "State of the Open Internet" report, the picture is stark: in a market where supply has been treated as a commodity, strategic and financial value accrues to the companies that control demand. Demand is the scarce resource, and walled gardens—deputized by media buyers to manage programmatic budgets—exert enormous control.

The numbers tell the story. Of the $1 trillion in global paid media investments in 2025, digital accounts for 70.7%, or about $640 billion. Within digital, paid search captures $204.6 billion and walled gardens $351.6 billion. The open internet—the open programmatic marketplace plus direct reservations—gets just $87.2 billion.

But it's the trajectory that's truly alarming. From 2017 to 2025, walled gardens grew from approximately 42% to 81% of non-search digital advertising budgets. As Jounce notes, "the small and shrinking number of walled gardens have captured a large and growing share of all digital investments." The open internet's share isn't just declining—it's stagnating in absolute dollars while everything around it expands dramatically.

Here's what that dominance looks like at the highest level: seven of the top ten companies in the world by market capitalization make much of their money from advertising. We can likely add another two or three AI companies to that mix soon. These are the most valuable companies on the planet, and advertising is core to their business models.

Yet there's no champion of the open web in that top ten. Not one. The economic value created by the open internet—the journalism, the niche communities, the diverse voices, the independent publishers—flows primarily to the platforms that aggregate and control access to it, not to the creators and publishers who produce it. This is the fundamental imbalance we must address.

Further, most walled gardens operate DSPs and ad networks that have become the primary source of revenue for open internet media companies. A large share of what appears as "open internet" spend in the data actually originates with Google (Google Ads + DV360), Amazon (Amazon DSP), and ad networks operated by other walled garden media companies. Even when publishers win, they're winning on platforms controlled by their competitors.

This is the squeeze: publishers need higher yields to survive the traffic apocalypse. To get higher yields, the open web needs to become a better product. But that's precisely where we've failed—and it's not the publishers' fault.

The Tragedy of the Commons

Our conversation kept returning to Garrett Hardin's famous 1968 concept: the tragedy of the commons. In the original formulation, individual herders acting in their own rational self-interest would overgraze a shared pasture until it was destroyed, leaving everyone worse off.

Today, the open web is that pasture. But here's where most people get the story wrong: publishers aren't the ones who destroyed the commons. They're the victims of a system with fundamentally broken incentives.

Publishers should never have been put in a position where their premium inventory was commodified and priced the same as low-quality knockoffs. This was a foundational error—one driven by the promise of "user targeting" that said the audience mattered more than the context. Publishers were told their inventory was interchangeable, that what mattered was reaching the right user, not the environment in which that user was reached.

Established economic principles tell you exactly what happens when you allow quality products to be sold alongside counterfeits at similar prices: the quality products lose. Gresham's Law in monetary policy states that "bad money drives out good." The same principle applies to inventory. When fraud, MFA sites, and arbitraged garbage compete in the same marketplace as premium publisher content, the bad drives out the good.

Publishers were told to accept this. They were told that programmatic was the future, that scale mattered more than quality, that targeting the user was paramount. And as their organic traffic evaporated and walled gardens captured an ever-larger share of budgets, they fought for survival with increasingly limited options.

The pollution in the open web commons comes from multiple sources. Reputable publishers, desperate to recover a fair share of ad budgets just to keep operating, found themselves competing in a polluted marketplace. But much of the actual pollution—the MFA sites, arbitraged traffic, auto-refreshing ad slots, domain-spoofed impressions—comes from bad actors and intermediaries exploiting the market dynamics.

And here's the truly perverse part: the adtech intermediaries who take a cut from each transaction would actually lose money by removing low-quality or fraudulent inventory. Think about that. The companies facilitating these transactions have no economic reason to police quality. In fact, they benefit from more transactions, regardless of quality. More impressions mean more fees, even if those impressions are worthless or fraudulent.

This is the structural rot at the heart of the open exchange. It's not that people are evil; it's that the incentives are backwards. As Matthew Prince of Cloudflare puts it, traffic has never been a good proxy for value, yet we built an entire industry around maximizing it. Publishers were caught in a system designed to fail them.

But there's a second, more structural problem: publishers lacked a unified offering like the walled gardens. It wasn't just that the open web was "polluted"—it was a wholly less efficient place to scale outcomes, even when it delivered strong results.

The walled gardens certainly haven't been perfect. They still face quality and fraud issues, and they don't take enough responsibility for these problems. But they offered something crucial that the open web failed to provide: media that could be shown to deliver outcomes, combined with a greater level of accountability for quality and fraud than the chaotic open exchange ever achieved.

Facebook and Google offer advertisers a single login, unified measurement, consistent quality standards. The open web offered fragmentation, complexity, and opacity. The walled gardens' failings—and they have many—actually show how a better open web can fight back if it learns the right lessons. They proved that advertisers will pay premium prices for quality, consistency, and accountability. The open web just needs to deliver those same attributes with the reach and diversity that only the open internet can provide.

Advertisers rationally flee to walled gardens not only because of quality concerns, but because walled gardens made it easy. When you can't trust the quality of inventory, when your campaigns might appear next to AI-generated gibberish, and when you need to navigate dozens of relationships to achieve the same scale you get from one walled garden buy, the safest choice is to retreat behind the walls.

Now AI accelerates this dynamic. It's not a separate problem we can solve in isolation—it's amplifying every existing dysfunction. AI companies extract value from publisher content while paying nothing. They reduce traffic to publishers, eroding the core business model. And they do this while the open web is already hemorrhaging share to walled gardens because of quality and complexity issues.

We can't fix just the AI part, even if we could. We have to address the perverse incentives, the programmatic hotdog problem, the tragedy of the commons—all of it. These forces brought us here. AI just reveals how unsustainable the foundation always was.

But here's the critical insight: fixing the quality and incentive problems will actually help us address AI's impact. We're not fighting the tide—AI will reduce some page impressions, and that's inevitable. But what we can do is:

Ensure AI is held accountable for how it uses and rewards the data it takes from publishers. Whether through licensing, pay-per-crawl mechanisms, or other market-based solutions, value must flow back to creators.

Get fair value for the audiences that remain. When traffic declines 25%, but quality deals deliver 70% revenue increases through higher CPMs and better performance, publishers can still grow. Revenue concentrates in quality inventory while junk inventory (which AI can easily replicate) becomes worthless—as it should be.

Provide genuine value to advertisers. When AI cuts off search traffic, advertisers still need to build brands and reach audiences. Quality deals on the open web offer what AI summaries cannot: trusted environments, contextual relevance, and audiences in the right mindset to engage with brands.

This isn't just about display advertising, either. It's a precursor to what could happen in CTV. If we don't learn these lessons now, Connected TV will follow the same trajectory—commodified inventory, perverse incentives, walled garden dominance. The stakes extend far beyond banner ads.

The Elegance of Hardin's Solution

But here's the thing about Hardin's tragedy of the commons: it comes with its own solution. Transparency.

When herders can see what others are doing to the pasture, when there's visibility and accountability, rational actors can coordinate to preserve the commons. The problem isn't the open exchange itself—it's the opacity that allows bad actors to externalize the costs of their pollution onto everyone else.

And here's the deeper problem: the entire system has perverse incentives. AdTech companies would actually lose money by removing low-quality or fraudulent inventory. Think about that. The intermediaries taking their cut from each transaction have no economic reason to police quality—in fact, they benefit from more transactions, regardless of quality. More impressions mean more fees, even if those impressions are worthless or fraudulent.

Right now, bad actors can get kicked out of the ecosystem and immediately respawn under a new domain or seller ID. There's no persistent identification, so it's trivially easy to come back. Yet the solution exists: transaction IDs and seller IDs tied to registered, trusted payment accounts. If someone commits fraud, you blacklist them at the payment level, not just the domain level. They can't just spin up a new site and continue.

Why isn't this standard practice? Because any exchange implementing it would lose money. This is the perverse incentive laid bare. Platforms make money from garbage inventory. Removing garbage is like removing a limb. They should do it, but they don't, because their business models are built on volume, not quality.

This is the structural rot at the heart of the open exchange. It's not that people are evil; it's that the incentives are backwards. As Matthew Prince of Cloudflare puts it, traffic has never been a good proxy for value, yet we built an entire industry around maximizing it. Publishers were caught in a system designed to fail them.

I don't think we can clean up the open auction. The incentives are too misaligned, the scale too vast, the players too numerous. But we can create transparent, clean private deals that rebuild advertiser trust. This is where the tragedy of the commons meets economic reality.

The Programmatic Hotdog Problem

Here's the analogy that crystallized everything: think of open auction programmatic as a hotdog.

From the advertiser's perspective, the hotdog is outcome-driven—it's all about taste, satisfaction, and that last-bite conversion. But here's the problem: I love a good hotdog. I really do. But I wouldn't eat one when the ingredients are sourced from the lowest bidder—mystery meat from questionable suppliers, fillers from who-knows-where, processed through the longest and most opaque supply chain imaginable.

This is the open web today. Open auction programmatic has become the dollar menu of digital advertising: cheap, fast, and deeply unsatisfying. The ANA estimated annual programmatic ad wastage in the US at $22 billion, or 25% of total spend. That's not a rounding error. That's a structural crisis.

The inventory is sourced through a labyrinth of resellers and arbitrageurs. The audiences are cobbled together from the dregs of third-party data. The placements appear on sites that exist solely to capture programmatic spend—digital ghost towns with no real human engagement. And when advertisers inevitably discover they've been paying premium prices for subprime inventory, they retreat further into the walled gardens.

The programmatic hotdog problem isn't about the technology. It's about trust. It's about knowing what you're buying and where it came from.

The Solution: Source Your Ingredients from Deals

Instead of continuing down this path, we need to fundamentally reimagine how we source and package open web inventory. We need to get our ingredients from deals—the Whole Foods of the programmatic ecosystem, if you will.

More than 66% of all open exchange advertising, representing $100 billion in annual spend, is now transacted through private marketplace (PMP) deals. This shift toward curation represents the industry's recognition that quality matters. But here's the critical distinction: not all deals are created equal.

Some are Whole Foods—premium, organic, sustainably sourced, with full transparency into provenance. Others are what I call "unholy foods" (and yes, I'm rather proud of that pun)—repackaged garbage dressed up in private marketplace clothing. The industry's newest obsession with curation has spawned countless "curators" who are simply arbitrageurs with better marketing, taking a random sample (or worse) of the open web and packaging it to make a margin.

What Makes a Good Deal Actually Good

A quality deal has specific, verifiable characteristics:

It uses data to ensure high quality inventory - providing a level of inventory quality and targeting you genuinely cannot buy from the open auction. The publishers are exclusively quality publishers. The curators are either the publishers themselves or specialist companies that aggregate supply while taking genuine responsibility for outcomes and quality.

It's still priced on CPM, but delivers fundamentally different value: higher viewability, higher attention, much higher brand safety, dramatically lower fraud and MFA exposure, and ultimately far better outcomes for advertisers.

Someone is taking responsibility - and this is the key difference. In an ideal implementation, curators can restrict inventory to only suppliers they can verify at a payment ID level. (This is exactly why and how to use transaction IDs properly - but more on that later.)

The economics work for everyone: Publishers who included their ad inventory in curated PMPs saw an average 70% revenue increase on mobile in the United States and a 13% uplift for connected TV. This isn't marginal improvement - it's transformational.

Here's how the economics can actually work:

Current state: Traffic is declining 25% year over year. Even with modest CPM increases, revenues are in secular decline. Publishers are fighting over scraps while walled gardens capture 81% of growth. Deal-based future: Yes, we may lose page volume - that's inevitable with AI. But the yield on what remains can rise dramatically. Revenue concentrates where it belongs: in the best quality publishers. Higher-performing media attracts more market share from walled gardens. This translates to higher CPMs that more than offset volume declines.

The goal isn't just to survive. It's to grow the open internet from $87 billion today to $100-150 billion by 2030, with more money going to publishers and less being extracted by unnecessary intermediaries. The adtech providers who enable this will be rewarded with sustainable businesses built on value creation rather than value extraction.

Making Deals Compete on Quality

We need these deals to compete transparently on quality. That means:

Quality environments. These marketplaces should only pay out money to vendors they are proud to work with. High quality publishers, high attention. Appropriate verification with bad actors getting cut. Verifiable supply chain transparency: No hidden resellers, no domain spoofing, no made-for-advertising sites masquerading as premium publishers. Ideally one or two hops maximum - no more of the 15+ intermediary daisy chains. This should eventually move to direct supply with either the DSP or SSP as the conduit. Any supply chain war that results needs to compete on quality of product not price. Transaction ID implementation: Seller IDs tied to registered, trusted accounts and payment IDs. This prevents bad actors from simply respawning under new domains when they get caught. If you commit fraud, you're blocked at the payment level, not just the domain level. Effective Data. As Gareth Glaser points out. Much of the data being sent to the exchanges is poor, a volume rather than a precision game today. This needs to change.

But quality isn't just about blocking fraud and MFA sites. Matthew Prince offers a more profound vision: the future will reward unique, local, and high-quality content that fills what he calls the "holes in human knowledge." Traffic has never been a good proxy for value. The most valuable content will be scarce, original, and authentic—the opposite of the commodity content that currently pollutes the open exchange.

Prince uses a Swiss cheese metaphor: AI models represent all known knowledge, and the most valuable content fills the gaps. Hyper-local journalism, quirky niche expertise, original research—these are the premium ingredients that advertisers should pay for, not the mystery meat of the open auction.

This shift from volume to value requires a fundamental reimagining of what we're selling. We're not selling impressions or clicks. We're selling access to unique audiences consuming unique content in quality environments. That's a premium product worthy of premium pricing.

And here's the business model that makes it sustainable: good curators should be rewarded with market share, not margin extraction. With header bidding, we can go straight from publisher to curator - a low-hop environment where total tech tax never exceeds 20% and ideally sits under 10%. We don't need multiple actors taking 3-5% cuts each. That's the old model, and it's broken.

The Trade Desk Isn't Doing Everything Right, But It's Pushing in the Right Direction

The Trade Desk's SP500+ launch—a curated collection of 500 premium publishers—sparked intense debate. Critics dismissed it as an inclusion list dressed up as innovation. But regardless of execution, the strategic direction is correct: focusing buyer attention on quality inventory from known publishers.

We need more options here. Many more. We can't rely on a single DSP to define what "premium" means. We need the entire ecosystem—SSPs, independent curators, publisher consortia, agency trading desks—all building transparent, accountable deal structures.

But here's what concerns me: we keep thinking within the DSP/SSP paradigm. Buy-side optimization versus sell-side optimization. Demand path versus supply path. These binary frameworks limit our thinking.

Beyond the DSP/SSP Paradigm

What if we thought beyond these categories entirely? What if we built marketplaces purpose-designed for quality deals—where transparency isn't a feature, it's the architecture? Where publishers compete on audience quality and contextual relevance, not on who can stuff the most ads on a page? Where curators are held accountable for the performance of their packages?

This isn't about eliminating technology or returning to some pre-programmatic golden age. It's about using technology to create the transparency that prevents the tragedy of the commons. It's about building systems where quality naturally rises to the top because quality is what's measured, rewarded, and compensated.

Always, though, with a view toward private quality. Not scale for scale's sake. Not volume metrics that mean nothing. Quality that advertisers can trust and pay premium prices for.

A Call to Arms: Redefine the Categories

Here's the uncomfortable truth that Scott, Fern, and I kept returning to: we need to stop pretending that open web programmatic and open web deals are the same channel. They're not. They're fundamentally different products serving different purposes.

Open auction programmatic is volume trading—it's the spot market for impressions where scale trumps quality. It has its place for certain use cases, but it's not premium, and we need to stop marketing it as such.

Open web deals, when properly constructed and transparently executed, are premium products that command premium prices. They offer the targeting sophistication of walled gardens with the reach and diversity of the open web—but only when built on foundations of quality and trust.

The industry needs to unite around a simple principle: we will only recommend that advertisers buy from quality deal environments. Not open auction. Not mystery meat programmatic. Quality, curated, transparent deals.

Think of it this way: when you want a quick, cheap meal, you buy a hotdog from a cart. When you want a proper dining experience, you go to a restaurant where you know the provenance of the ingredients, the skill of the chef, and the standards of the establishment.

Don't buy a programmatic hotdog and expect a filet mignon. Get a deals-based gourmet burger—one where you know exactly where the beef came from, how it was raised, who ground it, and how it was prepared. Where transparency isn't a buzzword, it's the business model. Where "premium" isn't self-declared, it's verifiable.

The Outcomes Problem: Why Measurement Matters

There's a growing trend toward paying for outcomes rather than impressions, and rightly so. Advertisers should pay for results, not activity. But here's the uncomfortable truth: programmatic has fundamentally failed at outcome-based measurement, and that failure is killing the open web.

As Gareth Glaser, co-founder and CEO of Gamera and author of "Gareth Hates AdTech," has compellingly demonstrated, there are two primary reasons Meta consistently outperforms programmatic on post-click outcomes:

First, Meta trades on CPC while programmatic trades on CPM. Meta is optimizing toward clicks and conversions from the start. Programmatic is optimizing toward impressions and hoping conversions follow.

Second, and more critically, Meta has conversion tracking on every campaign via its pixel, while programmatic DSPs have conversion tracking but programmatic publishers don't. The people who build, place, and shape the ads—the publishers—have no insight into what's performing. Meanwhile, the people who do have performance insight—the DSPs—are structurally biased to value impressions and post-impression outcomes equally to post-click outcomes.

This creates a fundamental misalignment. Publishers can't optimize for what they can't see, and what they can see (viewable impressions, CPMs) doesn't correlate with what advertisers actually care about (conversions, ROAS, business outcomes).

Badly Structured Outcomes Create Badly Structured Incentives

Glaser makes a crucial point about MFA sites: they're not a bug in the system, they're a feature. When the primary desirable payable event in programmatic became "viewable impressions," the market rationally responded by creating sites whose primary purpose is generating as many viewable impressions as possible, as quickly as possible.

The optimization goal wasn't "viewable impressions on high-quality content" or "viewable impressions without ad clutter." It was just "viewable impressions." So the market delivered exactly that—pages stuffed with ads, optimized for viewability metrics, with content as an afterthought.

As Glaser writes: "It is my firmly held belief that badly structured outcome goals are to blame for programmatic being subject to so much bullshit. When outcomes are poorly structured, and are not tied to explicit measurements of performance in some way, the industry naturally becomes a hornet's nest of misaligned incentives and tomfoolery."

This is why Meta outperforms programmatic. Not because Facebook's users are more valuable or their targeting is magic, but because Meta's measurement infrastructure actually tracks the outcomes advertisers care about, and their optimization systems can see and respond to that data in real-time.

Quality as the Filter for Great Outcomes

The solution isn't just better measurement—it's measurement systems built on a foundation of quality inventory.

High-quality publishers with engaged audiences, premium content, and clean user experiences naturally drive better outcomes. But programmatic's measurement systems don't differentiate between quality and garbage until it's too late. By the time the campaign data shows poor performance, the budget has been spent on junk inventory.

What we need are systems that use quality as a filter upfront and include measurement systems that demonstrate the great outcomes this space can drive.

This means:

Quality-first deal structures where only verified premium inventory participates, with clean supply chains and transparent audience definitions.

Publisher-level conversion tracking so that publishers can see what's working and optimize their inventory and audiences accordingly. Break the information asymmetry that advantages walled gardens.

Unified measurement frameworks that track true business outcomes—not proxy metrics like viewability—across the entire customer journey, with log-level data available to all parties.

Performance-based deal pricing where publishers and curators are rewarded for driving actual results, not just serving impressions. Put your money where your measurement is.

Continuous feedback loops that allow buyers and sellers to optimize together toward shared outcome goals, with transparency into what's driving performance.

The open web is massively good at driving outcomes when done correctly. But "done correctly" requires both quality inventory and quality measurement—and programmatic has systematically failed to deliver either.

This is why the shift to premium deals isn't just about cleaning up the supply chain. It's about creating the infrastructure for outcome-based success that programmatic has always promised but rarely delivered.

The Path Forward: What To Do Monday Morning

Here's the uncomfortable truth: we have about a year. Maybe less.

If the open web doesn't shift dramatically toward quality deals in the next 12 months, we could easily sink below $40 billion and become entirely dominated by walled gardens—even for the inventory that's nominally "open web." This is already happening with new in-house media teams who stick exclusively to the big seven platforms and never consider open web inventory at all.

The solution isn't to fix the open auction. It's to stop treating it as the same product as premium deal-based inventory, and to make that shift now. Here's exactly what each player needs to do starting Monday morning.

For Advertisers and Brands

Call your agency and say this: "I want you buying 70% or more of our open web inventory from verified deals. Use open auction only for incremental reach, and when you do, apply the highest level of verification. I want a list of the deals you're buying, what's in them, and how I can verify their quality."

Put it in writing. Make it a requirement in your media plan. Don't accept vague promises about "premium inventory."

For Agency Media Buyers

Go into your DSP and organize your open web inventory into three distinct cohorts:

High-quality deals - Curated, verified, accountable. This gets 70%+ of budget. Quality publisher direct deals - Known publishers with clear supply paths. This gets the rest of your proactive budget. Open auction with verification - Heavy exclude lists, fraud detection, MFA blocking. Use only for incremental reach when 1 and 2 are exhausted.

Stop using low-quality deals. If a "curator" can't tell you exactly what publishers are in their deal, what verification they provide, and how they take responsibility for quality - walk away. They're arbitrageurs, not curators.

Demand accountability. When something goes wrong - fraud, brand safety issues, performance failures - get your money back or remove that curator from all future campaigns. Make consequences real.

For Publishers

Ensure your inventory is high quality and being included in agency deal lists. If you're not, find out why and fix it.

Talk directly to advertisers. Rebuild relationships. Show them your audience data, your quality metrics, your verification. Don't hide behind SSPs and hope for the best.

Separate your quality inventory from anything that touches the open auction. Create clear deals with transparent terms. Be willing to stand behind your inventory with performance guarantees.

Choose your partners carefully. Work with SSPs and curators who genuinely prioritize quality over volume. If their business model depends on garbage inventory, they're not your partner—they're your competitor.

For SSPs, Curators, and DSPs

Reduce the number of hops. Aim for one or two maximum. Every additional intermediary is friction, cost, and opacity.

Implement transaction IDs properly. Tie seller IDs to registered, trusted accounts and payment IDs. Make bad actors impossible to respawn.

Make quality the path of least resistance. Your platforms should make it easier to buy quality deals than to buy open auction garbage.

Take real responsibility. If you're curating deals, you're vouching for that inventory. If it turns out to be fraud or MFA, make it right. Your reputation is your business model.

For Industry Bodies and Trade Associations

Test new providers. Ensure there's genuine competition and innovation in providing quality media. The space has become stale, dominated by incumbents with misaligned incentives.

Create standards and certification for what constitutes a quality deal. Make it easy for buyers to distinguish real curation from arbitrage.

Call out bad actors publicly. The industry has been too polite for too long. Name and shame companies that pollute the commons.

We Need a Champion

The open web needs what it's never had: a champion at the highest level.

Seven of the top ten companies in the world by market capitalization make much of their money from advertising. We'll likely add two or three AI companies to that list soon. Yet there's no champion of the open web in that elite group. Not one.

That champion could emerge from publishers who finally organize and unite. It could come from good actors among SSPs and curators who genuinely prioritize publisher success. It could be industry experts like Fern Potter, Scott Messer, Gareth Glaser, and others creating a marketplace that performs and that we can be proud of.

I'll be honest: I'm embarrassed by what the open auction has become and my own small part in its creation. We got fundamental things wrong. We believed that not judging price on inventory quality was somehow fair or efficient. We thought selling quality next to counterfeits and "letting the auction figure it out" would work. We underestimated the greed of actors who would hoover up margin without adding value. We were naive about how perverse incentives would dominate everything.

This time has to be different. The champion we need isn't just a company or platform—it's a coalition of everyone who believes the open web is worth saving and is willing to sacrifice short-term extraction for long-term sustainability.

The New Paradigm

This path forward requires courage and coordination:

Agencies and brands: Say no to the open auction by default. Demand quality, transparency, and accountability. Pay fair prices for premium inventory rather than chasing the cheapest CPM.

Publishers: Separate your quality inventory from the polluted exchange. Build verifiable, auditable deals. Work with partners who elevate rather than exploit you. You've been fighting an unfair battle—it's time to change the rules of engagement.

SSPs and curators: Make deal creation and management easier than open auction participation. Enforce real quality standards with teeth. Your curation is only as good as your willingness to reject bad inventory.

DSPs and platforms: Build infrastructure that makes quality deals the path of least resistance. Stop optimizing solely for scale and start weighting for quality.

Industry bodies and trade associations: Establish standards for what constitutes a "quality deal." Create certification programs. Call out bad actors who pollute the commons.

The tragedy of the commons isn't inevitable. It's the result of individual actors optimizing for short-term gains at the expense of long-term sustainability. We can reverse this trend, but only if we collectively decide that the open web is worth saving—and that quality matters more than volume.

Fast Forward: Two Possible Futures

The Dystopia (If We Fail)

It's 2030. The open web has shrunk to $35 billion—less than 5% of digital advertising. Even that is largely captured by walled garden DSPs buying "open web" inventory on their terms. Independent publishers have collapsed or consolidated into a handful of mega-properties. Local news is dead. Niche communities have disappeared.

All that remains is performance marketing—bottom-funnel "buy now" buttons on Google and Meta. Brand building happens exclusively inside walled gardens. Advertisers have zero negotiating power. CPMs are whatever the duopoly decides. Privacy is entirely controlled by companies that monetize surveillance.

And AI? It's extracting value from what little content remains, training on publisher work without compensation, answering questions without sending traffic. The open web has become a ghost town—technically existing but economically dead.

This isn't science fiction. This is the trajectory we're on right now.

The Vision (If We Succeed)

It's 2030. The open web has grown to $140 billion—not through volume but through value. Publishers who focus on quality, unique, local content are thriving. They earn premium CPMs because advertisers finally understand what they're buying and trust what they're getting.

The daisy chain of intermediaries has collapsed. Publishers work with one or two trusted partners who take reasonable fees—under 10% total—in exchange for genuine curation and accountability. Transaction IDs have made fraud nearly impossible. Bad actors can't respawn endlessly because payment-level verification stops them cold.

Advertisers build their brands on the open web because it offers something walled gardens never could: diversity of voice, contextual relevance, and reach without surveillance. The relationship between advertiser and publisher has been restored—enabled by technology but grounded in trust and mutual value creation.

AI companies? They pay for the content they use. Whether through licensing deals, pay-per-crawl mechanisms like Cloudflare's infrastructure enables, or other market-based solutions, value flows back to creators. Unique, high-quality journalism is recognized as the premium product it always should have been.

We don't have endless DSPs and SSPs. We have simpler, more intelligent connections between advertisers and publishers built on the right incentives. Quality matters more than volume. Scarcity drives value. Provenance is verifiable. Transparency isn't a buzzword—it's how the system works.

A champion of the open web sits among the top companies by market cap—not because it extracted value from publishers, but because its success is tied to their success. When publishers thrive, the champion thrives. Incentives are finally aligned.

Which Future Do We Choose?

The programmatic era had good intentions. It promised efficiency, scale, and democratization. But it ultimately failed to do the right thing, and it failed on market share. The walled gardens won because they offered what programmatic couldn't: consistent quality, simple access, and trustworthy outcomes. They proved advertisers will pay for accountability—even imperfect accountability.

We must learn the lessons of the past to ensure this next generation is one we can be proud to be a part of. We got it wrong by believing volume mattered more than value, that targeting users mattered more than respecting context, that "letting the auction figure it out" would somehow produce quality. We were wrong.

But we're not dead yet. The open web built the internet as we know it. It enabled independent journalism, niche communities, diverse voices, and the democratization of information.

There are enough people with talent and passion who want to save it. The question is whether we have the collective will to make it happen—and whether we have the courage to act now, not next year.

Because next year might be too late.

We don't want a world where all marketing is a "buy here now" button on Google and Meta. We don't want journalism to collapse because AI extracted all its value. We don't want the open web to become a footnote in the history of the internet.

The future doesn't have to be a choice between walled gardens and a polluted open exchange. We can build something better—a transparent, quality-driven open web that rewards creators, serves audiences, and delivers results for advertisers. A web where the best ideas win, where diverse voices thrive, where quality isn't just a buzzword but the foundation of the entire ecosystem.

That's the web worth fighting for. That's the web we can still build.

But we need to start Monday morning. Not next quarter. Not when the data looks worse. Now.

The clock is ticking. What we do in the next twelve months will determine whether the open web survives the decade.

Choose wisely.