The Advertising Industry Is Doing a Gilded Age Reenactment and Nobody Wants to Talk About It

A long essay about markets that stopped being markets, the one corner of advertising that escaped the squeeze, and Jay Gould.

I.

On the morning of September 24, 1869, which was a Friday, a 33-year-old named Jay Gould sat in an office above the Gold Room on New Street in lower Manhattan, watching the price of gold climb toward $160 an ounce, and tried to calculate whether he was about to become the richest man in America or the most publicly ruined one.

He had been planning this particular morning for about six months. The plan, to the extent a plan that relied on bribing the President's brother-in-law can be called a plan, was straightforward. Gould and his business partner, Jim Fisk, a former circus performer who had talked his way into a railroad board and then into controlling interest of the Erie, had been quietly buying up every ounce of gold they could get their hands on. At the time, gold and U.S. paper currency were not the same thing. They traded at fluctuating rates, like two currencies belonging to different countries that happened to share a postal system. If the price of gold went up, short sellers, meaning people who had bet the price would fall, had to buy gold at the new, higher price to cover their positions. If Gould and Fisk could push the price high enough, the shorts would panic, gold would spike, and the two of them would walk away with something like a quarter of the net worth of the United States.

The only thing that could stop them was the U.S. Treasury, which at any moment could dump its own gold reserves onto the market and crush the price. To prevent this, Gould had arranged for Abel Corbin, an obscure financial operator who happened to be married to President Grant's sister, to spend time at Grant's weekend house in Saratoga, explaining to the President at great length why it would be very bad for the economy if the Treasury sold gold right now. Grant, who was not an especially attentive President on economic matters, had been largely convinced.

On the morning of September 24, gold was climbing. $147. $150. $155. The short sellers were beginning to panic in the way that short sellers do, which is to say visibly and loudly and in some cases literally on their knees on the floor of the Gold Room begging buyers for mercy. Gould was watching this happen and preparing, I assume, to order champagne.

Then, at about 11:45 in the morning, Ulysses S. Grant figured out what was going on.

His wife Julia had, on his instruction, sent a warning letter to his sister Jennie ordering her to close out her husband Abel Corbin's speculative positions immediately, and Abel Corbin had seen that letter, and Grant had learned of his seeing it, and Grant now understood that his own brother-in-law was not a concerned patriot asking patriotic questions but a compromised participant in the corner itself. Grant, now correctly furious, immediately ordered the Treasury to dump $4 million of gold onto the market. The price collapsed in about fifteen minutes. Half of Wall Street was ruined by sundown. Thousands of investors who had bought gold at $160 watched it fall to $130 before they could unwind their positions. The newspapers, trying to capture the specific quality of the catastrophe, started calling the day Black Friday.

That is the origin of the phrase Black Friday. I had to look it up to be sure. I didn't quite believe it either. The word we now attach to doorbuster sales at Best Buy started, in 1869, as the description of a Gilded Age market crash caused by two guys trying to corner gold with the help of the President's brother-in-law.

I have been thinking about Jay Gould a lot lately, which, if you know me, is almost entirely out of character. I am not a 19th-century finance guy. I am not a historian. I run a company that sells out-of-home advertising (billboards, transit ads, airport displays, the screens above the bar in a hotel lobby), and my professional life mostly involves spreadsheets, coding, and conversations with people who own land near highways or want to rent ad space on land near highways. Jay Gould should not be occupying this much real estate in my head.

He is anyway, because the advertising industry in 2026 looks, structurally, a great deal like the financial industry in 1869. I realize that is a lot to claim in the fifth paragraph of an essay. I'll spend the next nine thousand words trying to justify it. You are welcome to stop reading at any point if the argument doesn't hold up. I would understand, and I would also, honestly, be a little impressed that you got this far.

What I want to do here is something I don't see many people in my industry doing out loud, which is to describe what is actually happening to advertising in the 2020s as honestly as I can, and to suggest that we have been somewhere quite a lot like this before, and that the way out of it last time is probably similar to the way out this time. If that sounds grand, it is grand. It is also, I think, true. Let me show you what I mean.

II.

The New York Times sold $662 million worth of advertising in 2014.

I want you to sit with that number for a minute, because the number by itself does almost nothing. It's just a quantity of money, and the thing I'm about to do requires you to have a feeling about it. So: 2014. The iPhone 6 had just come out. True Detective had aired its first season (the good one). LeBron was in Miami. Gas cost $3.36 a gallon, which I know because I looked it up, and which is a sentence that sounds invented but isn't. The New York Times had just won four Pulitzers in two years, run the Snowden stories, and published what was, by almost any measure a person could come up with, the most important English-language journalism of the decade.

And they sold $662 million in advertising.

Ten years later, in 2024, they sold roughly $508 million of advertising. That's about a quarter less, in nominal terms. In real terms, after adjusting for the dollar losing roughly a third of its purchasing power over the intervening decade, the Times's advertising business shrank by something close to half. In the same ten years, the Times tripled its digital subscriber base, built Wordle into a product line, acquired The Athletic, launched Cooking and Games as subscription services, and grew its total revenue from $1.6 billion to almost $2.9 billion. Every part of the business grew except one. The advertising business, the business that had for a century defined what it meant to be a newspaper, ended the decade materially smaller than it had started it.

The reason this is interesting, and the reason I started this essay with it rather than with the actual argument I want to make, is that it is not a story about the New York Times. It is a story about what happened to the thing the New York Times was selling into. If you imagine that the Times has been, for the past ten years, a very skilled fisherman standing next to a lake, this is not a story about whether he knew how to fish. This is a story about the lake.

Here is what the lake did:

In aggregate, outside of five specific companies, the global advertising market has not grown since 2018.

I had to read that sentence four or five times before I fully absorbed it. The number comes from WARC, which is essentially the Congressional Budget Office of the advertising industry. If they publish something, nobody serious argues with it for about eighteen months. Their finding, as of 2024: if you subtract Alibaba, Alphabet, Amazon, ByteDance, and Meta from the global ad market, whatever is left has been flat since 2018. Not growing slowly. Not growing with inflation. Flat. Eight consecutive years.

I'd like to do a kind of roll call here, because I think the sentence only starts working when you realize who "whatever is left" actually includes. It includes NBC and CBS and Fox. It includes every newspaper on earth. It includes every magazine publisher, every terrestrial radio station, every podcast network, every cable channel that hasn't yet shut down, and a couple that have. It includes Spotify. It includes Disney, as in the whole of Disney's advertising business, which includes ESPN, which includes Monday Night Football. It includes Netflix's ad tier (new) and Paramount+ (struggling) and Peacock (who can say). It includes the reconstituted remains of Twitter, which lost almost $6 billion in cumulative ad revenue after Musk bought it and replaced half of the advertisers with bots and the other half with whatever the opposite of an advertiser is. And it includes the one that stopped me cold. The New York Times.

All of them. Together. Did not grow. For eight years.

I checked this number. I want you to know I checked it, because I did not believe it the first time I read it, or the fifth. I pulled the underlying data. Global print advertising was worth $75.9 billion in 2016. By 2022 it was worth $37.3 billion. That's not a decline. That's an evaporation. Amazon's advertising business, which essentially did not exist before 2013, was worth $37.7 billion in 2022. So to a rough approximation, Amazon's advertising revenue equaled the entire global print advertising industry's. An industry that took roughly two centuries to construct. Absorbed, in dollar terms, by a company that used to sell books.

U.S. linear TV advertising peaked in 2018 and has been in structural decline since. Current projections have it dropping from about $60 billion to under $48 billion by 2027, which is a phrase analysts now write without emotion, the way obituaries describe long illnesses. Global audio advertising (podcasts, streaming music, terrestrial radio combined) has been flat at roughly $27 billion for five years running and is projected to decline through 2029. X's top advertiser in 2022 was AT&T, spending about $33 million a year on the platform. By 2025, AT&T's annual spend on X was $100,000. Disney's spend fell from $27.7 million to $500,000. These are not small businesses rounding to zero. These are Fortune 100 advertisers who used to send eight-figure checks to one of the top twenty media owners in the world, and who now send checks that a mid-sized dental practice could match.

The entire remainder of the advertising industry has spent the last eight years fighting for share of a pie that stopped getting bigger the summer TikTok showed up.

You remember that summer, or you remember a version of it. In August of 2018, ByteDance merged Musical.ly into a new thing called TikTok and released it globally. Most people in advertising did not notice. It looked, at the time, like a novelty app for teenagers who wanted to lip-sync. I have a vivid memory of a senior marketing executive at a consumer brand I was chatting with in late 2018 telling me, with real confidence, that TikTok was a fad and that "the kids will be back on Instagram by Christmas." He meant it. I can still hear the way he said the kids.

TikTok had, in 2018, approximately 55 million global users. It now has over 1.7 billion. It is the fifth member of the oligopoly this essay has been describing. It is, in 2026, one of the five companies that absorbed every incremental dollar of advertising growth over the last eight years. The reason I'm putting this fact next to the WARC stagnation data is that it is not a coincidence. The last time the global ad market for everyone not in the club was growing is, almost exactly, the same summer the fifth member of the club showed up at the door. Nobody let it in, exactly. It just arrived, and then within five years it was eating everyone else's lunch, and the advertising industry spent most of the intervening period trying to figure out whether to be afraid of it, whether to advertise on it, or whether to pretend it was still going to go away.

That summer, in memory, feels both very recent and impossibly long ago, which is approximately how long eight years is. Most of the things that felt fragile in 2018 are now structural. Most of the things that felt structural in 2018 are now gone. The advertising market you and I work in is one of the things that felt structural and is now, for everyone not in the club, effectively stalled.

And yet.

I've been giving you a list of categories that have flatlined or shrunk since 2018, and the list is long enough that it can feel, reading it, like the entire advertising industry has been slowly compressed into the shape of five California companies. But that isn't quite true. There is one advertising category, and I want to be specific about this because it is the entire reason I started writing this essay, that has not flatlined. That has, quietly, compounded. That has crossed dollar-value milestones in the same eight-year window in which everything else I've named has been fighting to hold its ground.

It is the kind of advertising that hangs on the side of a building in Des Moines.

U.S. out-of-home advertising, meaning billboards on Interstate 80, signage at O'Hare, the wrapped side of a bus in Chicago, the screen above the bar in a hotel lobby, crossed $9 billion in domestic spend for the first time in 2026. The growth is not spectacular. It is slow, it is physical, it takes cement and permits and the cooperation of actual landowners to happen. But it has compounded for most of the years that everything else I've described has not.

The question I kept arriving at, the one I could not stop asking, was: why. Why would a 150-year-old medium, made of vinyl, steel, and aluminum and occasionally a single enormous LED, be growing in the same decade that print halved and linear TV entered a managed decline and the New York Times watched its advertising business quietly shrink by roughly a quarter?

The answer, I think, is structural, and it is the same answer that explains why the Gilded Age lasted as long as it did and why it ended the specific way it ended. A billboard on Interstate 80, standing in a field outside Des Moines, is not logged into anything. It does not have a Google account. It does not report its impressions to a walled garden. It cannot be bought inside Ads Manager. It is visible to every driver who passes. The only entity that knows how many drivers passed is the company that owns the field, and the only way to put an ad on it is to ask them.

That is the shape of a market the robber barons of 2026 cannot own.

The rest of this essay is about why, and about what that means, and about how we got here. It is going to take a while. I'd recommend a second cup of coffee.

III.

There is a phrase for what Alphabet, Amazon, Meta, ByteDance, and Alibaba have been doing for the past eight years, and the phrase has been in use for approximately 155 years, which is to say it showed up roughly the moment industrial capitalism at scale became a thing human beings had to talk about. The Atlantic Monthly used it for Cornelius Vanderbilt in August 1870. It stuck. It is now, probably, the single most recognizable epithet in the history of American commerce.

They called them robber barons.

I want to spend a minute on the phrase itself, because I think most people who use it today don't know where it comes from, and the etymology does a surprising amount of the argumentative work in this essay. Robber baron is an English translation of the German Raubritter, which literally means "robbing knight," and which was the medieval Rhineland term for a specific kind of feudal lord. A lord who would set up a toll booth on an unauthorized stretch of river and demand payment from merchants whose goods happened to float past his castle. The key word is unauthorized. The toll wasn't for services rendered. It wasn't for protection or improvement of the river. It was for the right to keep moving. The lord owned a stretch of the waterway, and if you wanted to ship your flax to Cologne, you paid him to not stop you.

The phrase applied itself to 19th-century American industrialists because industrialists were doing, in an industrial context, exactly what Raubritter had done on the Rhine. They owned the rails, the ports, the refineries, the coking plants, and the telegraph wires (the infrastructure between producers and markets), and they charged whatever the traffic would bear to let commerce flow through. This was not, in the strictest sense, theft. The railroads really did transport the wheat. Standard Oil's kerosene really was cleaner and cheaper than the alternatives. The telegraph lines carried messages. What made the barons robber barons was not that they produced nothing. It was that they had arranged things such that commerce could not happen without passing through them, and they collected a toll every time it did.

The data, from that period, is staggering. In 1890, the wealthiest 1% of American families owned 51% of the country's real and personal property. Standard Oil, at its peak, controlled approximately 90% of U.S. oil refining. J.P. Morgan, at the height of his influence, led a network of banks and trust companies that a congressional committee documented as holding 341 directorships across 112 corporations, a web the public had by then given its own name: the money trust. Morgan personally organized a bailout of the U.S. Treasury in 1895, selling gold directly into federal coffers to prevent the government from running out of it, and then a few years later orchestrated a private-sector rescue of the entire American banking system during the Panic of 1907, an intervention so consequential that Congress, partly in response to having watched a single private citizen perform the function of a central bank, created the Federal Reserve. A private citizen was liquid enough, and connected enough, to twice keep the American financial system functional. There was a period in the 1890s when if Morgan had taken a long weekend, the Treasury might have defaulted.

(Side note: the fact that J.P. Morgan personally bailed out the federal government twice should be in more essays. It comes up almost nowhere. It keeps not coming up. If you're ever in a meeting where someone is making a vague point about government capacity, try this line, please. "J.P. Morgan personally bailed out the federal government twice" will end any meeting you deploy it in. It is a conversation-killer of the highest order, and it is also true.)

Here is the move that matters for my argument. The robber barons were not, on balance, stupid men running inferior companies. This is the thing that took me longest to absorb when I started reading about them, because the cultural caricature of the Gilded Age industrialist is a mustache-twirling villain sitting on a pile of gold coins, and the reality is a lot more uncomfortable. Rockefeller's oil was genuinely better than his competitors'. It burned cleaner. It came in consistent grades. It cost less per gallon to produce because Standard Oil had vertically integrated the entire supply chain. They owned the barrels, the pipelines, the tank cars, the wholesale depots. Carnegie's steel was structurally superior to its competitors'. The Bessemer converters at the Homestead works could produce steel rails that lasted decades. The rails his competitors sold had to be replaced every few years. The railroads actually moved the wheat. They moved a lot of wheat. The American industrial economy that emerged from the Gilded Age became, within thirty years, the largest economy in the world, and most of that infrastructure was built by the people we now call robber barons.

What made them robber barons was not the quality of their product. It was the position of their tollbooth.

Rockefeller is the cleanest example, and the one whose story is most useful for understanding what's happened to advertising. He did not win the oil business by making better oil than anyone else, although he did. He won it by making a deal with the railroads. The South Improvement Company, formed in 1872, was technically a railroad pool: an agreement between the major Eastern trunk lines to stabilize rates and limit competition. What the public didn't know, and what Ida Tarbell would eventually expose, was that Standard Oil had negotiated a second, secret provision. Rebates. For every barrel of oil Standard shipped, the railroads would give Standard a discount. That was bad enough. The next part was worse. For every barrel of oil Standard's competitors shipped, the railroads would also give Standard a cut, a payment called a "drawback," paid by the railroad to Standard Oil every time a competitor used the same railroad to move a competing product.

Let me state that again, because I don't think it quite lands the first time. Standard Oil was being paid a percentage of its competitors' revenue every time its competitors used the railroads. It was extracting a toll on commerce it wasn't part of. The competitors often didn't know this was happening. They just knew that their margins kept compressing, their shipping costs kept going up, and for some reason Standard Oil kept being able to undercut them on price. By the time the arrangement was exposed in the early 1880s, half of the independent oil producers in Pennsylvania had already been rolled up or driven out of business, and Standard Oil controlled 90% of the market.

If that arrangement sounds familiar, if it sounds to you like a description of something other than 19th-century oil, that is because it is.

IV.

The advertising industry is consolidating at three layers simultaneously, and I think that's the thing most people miss when they talk about this. We usually talk about one layer at a time, meaning Google's power, or the holding company merger, or the media owner roll-ups, and the one-at-a-time framing obscures the real picture. The real picture is that every layer of the stack is concentrating at the same time, in the same window, for the same underlying reasons. That has happened exactly one other time in American history, and we gave that time a name.

Layer one: the platforms. Google, Meta, Amazon. These are the Standard Oils of the 2020s, which is a comparison that will annoy people who work at those companies and that I stand behind anyway. They own the channels where ads are served. They own the targeting infrastructure. They own the measurement. They set the prices, they grade the results, and the advertisers who use them have, as the operative term of art goes, nowhere else to go. A 2024 analysis put it this way: advertising inside a walled garden is operating inside a system that is optimized first for the platform's growth, second for user experience, and third, if there's any budget left, for whether the ad actually actually worked. That is not a conspiracy. That is written into the business model. A platform that put advertiser outcomes first would be leaving money on the table, and the market, being what it is, would punish it for doing so.

(Mild digression: I don't actually think Google is evil. I've met people who work at Google. They are largely nice, largely smart, and largely doing what any rational actor in their position would do. Same for Meta and Amazon. If I ran a business with a 30% operating margin and shareholders who expected it to grow every quarter, I would also protect the measurement layer. I would also make it hard to export my data. I would also build walled gardens. The problem isn't that these companies are run by bad people. The problem is that the incentives are what they are, and the incentives are not aligned with the people paying the bills. This is a structural observation, not a moral one. I want to be clear about that, because the structural observation is much more dangerous than the moral one. You can replace the leadership of Google. You cannot replace the incentives.)

Layer two: the holding companies. This is the railroad consolidation of the current Gilded Age, and it happened, with almost novelistic timing, in November of 2025. Omnicom acquired Interpublic Group for $13.5 billion. The deal had been announced in December 2024 and spent the year working its way through regulators on three continents. When it finally closed, the combined company had $26 billion in annual revenue, eliminated approximately 14,000 roles, and collapsed the Big Six global advertising holding companies into the Big Five. Vanderbilt, who in 1870 rolled up two competing lines into the New York Central and Hudson River Railroad (the year after Jay Gould's gold corner, which is to say when the dust of that particular catastrophe had barely settled), would have recognized the playbook exactly. You buy the rival, you fire the middle management, you vacate the office space, and you emerge with a single network that sets the prices for every advertiser in your Rolodex.

It is not just the scale. The holding companies have also spent the last decade buying their own data layers, so that when a CMO hires an agency, the agency delivers them to its own proprietary data platform, which measures its own campaigns, which reports against its own benchmarks. Publicis bought Epsilon for $4.4 billion in 2019. IPG bought Acxiom for $2.3 billion the year before. Dentsu bought Merkle back in 2016. What these acquisitions built, collectively, are mini walled gardens inside the larger walled gardens. Closed measurement systems inside closed media systems, each of which has a financial reason to want you to spend more rather than less.

(Aside for anyone who worked at IPG: I'm sorry. I don't know what else to say. 14,000 people is not an abstraction, and I have a couple of friends in that number. History also doesn't know what to say to you, which is partly why history keeps happening.)

Layer three: the media owners. In the same quarter the Omnicom-IPG deal closed, Clear Channel Outdoor, one of the largest outdoor media company in the Americas, was taken private by Mubadala Capital and TWG Global in a $6.2 billion all-cash transaction. Outfront Media, the second-largest outdoor company in the United States, announced a $20 million strategic investment in an independent marketplace to modernize its sales operations. These moves happened within a 90-day window, were mostly not covered by the mainstream press, and represented the largest structural reorganization of the physical advertising layer in about three decades.

I know about all of this in some detail because I work in it. This is the closest I'm going to come in this essay to mentioning my own company by name, and I'm going to leave it at that, because the essay is not about my company. If you want to know what I do for a living, my LinkedIn is right there. You can figure it out. What I want to note is that the physical layer of advertising, meaning the screens you pass on your way to work, the billboards you see from the highway, the signs at the airport, is restructuring in parallel with the other two layers, and for the same reasons. Media owners want to be part of something bigger, or part of something private, or part of something that can negotiate with the platforms from a position of scale rather than weakness. The individual rational move, for each of them, is consolidation. The aggregate outcome, across three layers simultaneously, is the Gilded Age.

Three layers of one industry, consolidating at the same time, for the same reasons, in the same window, is not a coincidence. It is a pattern. The pattern has a name. We already gave it a name a thousand words ago.

V.

I want to take a detour through Pennsylvania. I know this seems like a digression. It is a digression. It's also, I think, the most useful way to understand what's happening to advertising right now, and I'll explain why in about nine hundred words.

The Pennsylvania oil fields, in the decade after the Civil War, were the largest industrial boom of the 19th century. Oil had been discovered in commercial quantities at Titusville in 1859. Within four years, the hills above Oil Creek were producing more petroleum than the rest of the world combined. Towns appeared overnight. Pithole City went from a wheat field to 15,000 residents in nine months, which is genuinely a wild sentence to have to write about a place that doesn't exist anymore. Wildcatters became millionaires in a summer. The price of a barrel of crude oil was $10 in 1860, collapsed to ten cents by the end of 1861 as too many wells came online at once, recovered to $4 in 1862 when the producers organized to set a floor, then spiked back up above $12 during the Civil War years, then collapsed again to $2.40 by 1867. Nobody had any idea what anything was supposed to cost, because the entire industry was less than a decade old and every month some new well or refinery changed the math.

This was, essentially, advertising's open web era, except with more beards and significantly more gunfights.

Thousands of small producers, none of them with pricing power, all of them drilling as fast as they could, all of them competing to the death. No one owned the market because no one could own the market. If you had a well, you had oil. If you had oil, you could ship it. If you could ship it, you could sell it. The railroads that carried the oil out (the Erie, the New York Central, the Pennsylvania, the Atlantic and Great Western) charged the going rate. Everyone had access to the rails. The rails were the rails.

John D. Rockefeller arrived in this chaos in 1865 as a 26-year-old Cleveland bookkeeper who had decided to get into the refining business. He did not have the largest refinery. He did not have the best wells. He didn't even have oil. He was strictly a refiner, buying crude from the Pennsylvania producers and turning it into kerosene for urban lamps. What he had, eventually, was a deal with the railroads. The deal I described earlier. The rebate structure. The drawback system. The secret rate schedule that meant Rockefeller's shipping costs were a fraction of his competitors', and that every barrel his competitors shipped generated a small payment to Rockefeller from the railroad itself.

The Pennsylvania producers did not know this was happening. They just knew that their margins were compressing and Standard Oil's weren't. They knew that every time they tried to raise prices, Standard undercut them. They knew that every refiner in Cleveland, in Pittsburgh, in Philadelphia, was either being bought by Standard or driven out of business, and they did not understand why a company in Ohio had more financial leverage than companies operating fifty miles from the wells. By 1872, seven years after Rockefeller started, the "Cleveland Massacre" saw 22 of Cleveland's 26 refiners sold to Standard Oil within a three-month window. By 1878, the Pennsylvania producers had lost. The Oil Exchange in Titusville, which had once been the beating heart of the American petroleum industry, closed its trading floor. The towns that had grown up in the Allegheny hills, meaning Pithole, Babylon, Red Hot, a dozen others, emptied out. Most of them no longer exist. If you visit Pithole today you can walk through grass and occasionally see a rectangle of foundation stones where the opera house used to be.

I tell that story because it is almost perfectly the story of the open web in advertising, and because I think the parallel is so exact that once you see it, you can't stop seeing it.

The small Pennsylvania producers were not bad at oil. They drilled wells, they struck oil, they moved barrels. They did every single thing a petroleum producer was supposed to do. They lost because they could not see the arrangement between Rockefeller and the railroads, the invisible structure under the visible market, and because by the time they could see it, it was too late. The rebates had been in place for years. The drawbacks had been pooling into Standard's coffers every time a Pennsylvania barrel moved east. The competitive playing field had never been level. It had just looked level, which is worse, because it meant the producers spent years believing the game was winnable when it hadn't been, structurally, since 1871.

Every publisher I have ever spoken to about their digital advertising business, in any detail, has sooner or later arrived at a version of this realization. The New York Times is not bad at selling ads. The Washington Post is not bad at selling ads. NBC is not bad at selling ads. The publishers who have been squeezed over the past eight years are, almost without exception, operationally competent. They know their audiences. They have good sales teams. They produce premium inventory. They lost because they could not see the arrangement between the platforms and the identity layer, the invisible structure under the visible market, and because by the time they could see it, it was too late. The walled gardens had been built. The auction infrastructure belonged to the platforms. The measurement was marked by the graders. The competitive playing field had never been level. It had just looked level, which is worse, because it meant the publishers spent years believing the game was winnable when it hadn't been, structurally, since about 2014.

The Pennsylvania wildcatters were not bad at oil. They were bad at not owning the railroad.

We are not bad at advertising. We are bad at not owning the rails.

There is one form of advertising, however, in which the rails cannot be owned in the same way. I will come back to it in a few sections. I mention it now only to plant a flag: the thing that saved the category I work in is not better measurement, not better data, not better audience targeting. The thing that saved it is that its infrastructure is physical. You cannot own the rails when the rails are not rails. The rails are a billboard, in a field, on a hill outside Des Moines, made of aluminum, steel and vinyl and the right to occupy a specific piece of American ground.

Hold that thought. I'll return to it.

VI.

Okay. I've been making a confident argument for about five thousand words now, and I want to stop and ask if it's actually right.

The obvious problem with the Gilded Age analogy is that every time someone argues we're in a new Gilded Age, they're usually wrong. People have been saying we're in a new Gilded Age since approximately 1982. The argument resurfaces every decade. Sometimes it's about bank consolidation, sometimes it's about Reaganomics, sometimes it's about tech generally, sometimes it's about private equity. At some point it starts to feel like the boy who cried wolf, except the wolf is J.P. Morgan. If I'm being honest, the burden of proof for "this time is actually different" is on me, not on the reader.

So let me take the three biggest objections seriously.

Objection one: is this really concentration, or is it just the natural dominance of better products? Google Search really is better than the alternatives. Meta's ad targeting works at a technical level that no one else has matched. Amazon's retail media is effective because it sits next to purchase intent in a way that no other placement does. Maybe the 56% combined market share reflects real value creation, not market failure. Maybe the stagnation of everyone else is the market correctly routing dollars to the most effective place for them.

I think this is partly true, and I want to grant it. I think Google Search really is a better product than what existed before. I think the quality of targeting on Meta is a genuine technical achievement. If you had told me in 2005 that one company would be able to place the right ad in front of the right person as reliably as Meta now does, I would not have believed it. The products work.

But "the product is good" is exactly what Rockefeller's defenders said in 1890, and they were right. Standard Oil's kerosene really was better than the alternatives. That did not change what the market had become. The Sherman Act did not get passed because Standard Oil's product was bad. It got passed because a market with one dominant provider, however good, stops being a market in any meaningful sense. The market becomes the provider, and the provider becomes accountable only to itself. Whether or not the outcomes it produces are good is then a matter of its own benevolence, which is not a reliable structure for a democracy to rest on.

Objection two: isn't advertising a different kind of industry? Oil is a commodity. Steel is a commodity. Advertising is more like a creative service, and maybe commodity-market logic doesn't apply to it.

I used to think this was a strong objection. I've talked myself out of it. The physical substrate of advertising, meaning the impressions, the screens, the seconds of attention, the targeting signals, are absolutely commodities. What sits on top of them (the creative idea, the brand strategy, the emotional hook) is a craft, sometimes a great one. But the rails underneath are pipes, and the pipes are owned. The "craft" argument works fine at the level of a single agency creative director making a single campaign. It does not work at the level of an industry being reshaped by companies that sell pipes.

Objection three: what if this is all just cyclical? What if the industry concentrates for five years, antitrust pressure arrives, the walled gardens crack open, and we return to something closer to the fragmented 2005 market?

I thought about this one hard. I don't think it's cyclical. Here is the reason. The structural force that keeps the Gilded Age configuration stable is not legal or political. It's technical. More data means better targeting. Better targeting means more revenue. More revenue means more data. Once you are the company with the most data, you are also the company best positioned to acquire more data, and the next company has to build a data business from scratch while you compound your lead. This is the feedback loop that no 19th-century monopoly actually had. Rockefeller's competitive advantage was operational. He ran his refineries better, and operational advantages can be matched. A feedback loop is not a matchable advantage. It is a widening moat. Over time, it doesn't narrow. It widens.

So the three biggest objections, as far as I can see them, don't quite hold. I could be wrong. It has happened before. But I've looked at the argument from every side I know how to look at it from, and the Gilded Age parallel keeps holding up. At this point, convincing me out of it would require showing me a piece of industry history I haven't seen yet. I'd genuinely welcome the attempt. If you have that piece of history, write me. I'll read it.

What I want to do now is talk about how the Gilded Age actually ended, because that is the section where the essay either earns its argument or fails to.

VII.

The Gilded Age ended. How?

This is the question I think matters more than any other question in this essay, and I think most people get the answer wrong. The caricature version of the answer is "antitrust broke up the monopolies." That's not quite right. The real story is more interesting and, for our purposes, more useful.

The Gilded Age was dismantled by three specific pieces of federal infrastructure, enacted over roughly a 26-year period, none of which actually dismantled the industries they regulated.

The Interstate Commerce Act of 1887 was the first. It was also, structurally, the most important, and probably the least remembered. Its job was to regulate the railroads, the rails that had been the backbone of Rockefeller's advantage, Vanderbilt's dynasty, and Jay Gould's larger portfolio. The law did not nationalize the railroads. It did not break them up. It did not limit their size. What it did was require them to charge published rates, treat shippers equally, and operate under federal supervision. The railroads stayed enormous. They just became legible. A small Pennsylvania refinery in 1888 could, for the first time, see what Standard Oil was paying to ship a barrel from Cleveland to New York, because the rate was now required to be public. The tollbooth didn't come down. It just had a posted price list.

The Sherman Antitrust Act of 1890 came three years later, and it is the one most people remember, mostly because it was used, eventually, to break up Standard Oil in 1911. But notice what happened in that breakup. Standard Oil became 34 separate companies. Those 34 companies included entities that are now called ExxonMobil, Chevron, Amoco, Mobil, BP, Conoco, and about fifteen others. The combined value of the 34 spin-offs within a few years exceeded the value of the original Standard Oil. The "breakup" didn't eliminate scale. It eliminated coordination. The pieces had to compete with each other, which they did, imperfectly and with lots of cheating. But they had to, which was the point.

The Federal Reserve Act of 1913 was the third. It was prompted, in part, by J.P. Morgan's personal bailouts of the federal government. The observation was that perhaps the U.S. should not be reliant on a single billionaire in lower Manhattan to keep its banking system solvent. The Fed didn't break up the banks. It created a central institution that could perform the bailout function without needing a phone call to Morgan. The banks stayed large. They just stopped being the only game.

The pattern, across all three reforms: the Gilded Age corrections did not dismantle the industrial revolution they responded to. They made its infrastructure public, its pricing legible, its behavior accountable. The rails kept running. The oil kept flowing. The banks kept banking. What changed was what they were allowed to hide.

This is, I realize, an unglamorous version of the ending. The movie version is Teddy Roosevelt riding a horse in a tall hat and announcing he has slain the monopolies. The real ending is a lot more bureaucratic. A lot of federal rate schedules. A lot of published price lists. A lot of depositions and consent decrees and compliance filings that nobody remembers now because the compliance layer, once it exists, becomes invisible. The invisible layer is what held the post-Gilded Age economy together for seventy years. It is also, I think, almost exactly what the advertising industry is about to require, and what will eventually come, and what, when it arrives, will look nothing like the dramatic antitrust story most people are expecting.

The thing that solved the Gilded Age was not breaking up the powerful. It was making the powerful operate in public.

What this means for advertising: the reform, when it comes, will not look like Google being dismembered. It will look like regulatory pressure toward measurement that advertisers own, pricing that is published, and identity layers that are portable. The platforms will stay enormous. They should stay enormous. They're useful. What will change is that some of the dollars moving through them will become legible in ways they currently aren't. The advertiser will stop pretending that the platform's self-graded test is an audit.

This is, I think, already happening. The European Union's Digital Markets Act went into effect in 2024 and forces interoperability requirements on the largest platforms. Several state-level privacy laws in the U.S. have pushed the identity layer toward portability. Cross-channel measurement, meaning the ability to see how a campaign performs across Google, Meta, and everything else in a single dashboard, has gone from "marketer wish list" to "table-stakes procurement requirement" in the last three years. The Interstate Commerce Act analog isn't a single piece of federal legislation. It's a ten-year accumulation of smaller pressures, each of which forces a little more transparency into a system that was designed to prevent it.

That's the remedy. And it's not a surprise what the remedy is, because we invented it 140 years ago and wrote down how to do it.

VIII.

I want to talk about out-of-home advertising for a minute. The category I work in, the one I've been circling through this entire essay, the one I promised to return to and now am.

Here is the thing I think is true about out-of-home, and that I didn't fully understand until I started writing this essay: it is structurally the Interstate Commerce Act of the advertising industry, except nobody had to pass a law to get it there. It just was.

Let me explain what I mean. Out-of-home has three structural properties that no other major advertising medium has in 2026, and I want to name all three because any one of them alone would be interesting, but all three together are the thing that explains why the category has kept growing while everything else has stagnated.

One: it cannot be bought inside a walled garden. There is no Google Ads dashboard for billboards on Interstate 80. There is no Meta interface for the signage at O'Hare. The inventory sits outside the platforms, by physical fact. It has to be bought through a separate marketplace, which means the walled gardens do not control it, do not meter it, and do not grade the test. A CMO who buys out-of-home buys it from the media owner or through a marketplace that sits between them, and in either case, the data about what they bought, what they paid, and what it did belongs to them. This is unglamorous. It is also the thing.

Two: the audience is not logged in. A billboard reaches everyone who drives past it. There is no account to de-platform, no algorithm to demote the view, no privacy law that changes the measurement, because the measurement is, fundamentally, how many cars passed and how many eyes were plausibly on the screen. Out-of-home is the one major advertising category that was never built on an identity layer. The identity-layer erosion that is currently reshaping the rest of digital advertising (the cookie deprecation, the Apple privacy changes, the state-level privacy laws, the GDPR overhang, the impending AI-driven identity chaos) does not touch out-of-home at all. The category is, structurally, post-cookie. It has always been post-cookie. It was post-cookie before the cookie existed.

Three: the supply side is the opposite of concentrated. Unlike the platform layer (three companies) or the agency layer (five companies), the OOH inventory market has no dominant owner. Lamar, Outfront, Clear Channel, and JCDecaux combine for about 66% of the U.S. market. The remaining 34% belongs to independent operators. Family-owned billboard companies in Tulsa, regional transit networks in the Pacific Northwest, place-based media companies that specialize in gyms or grocery stores or urgent-care waiting rooms. There is no Standard Oil of billboards. There isn't even a Rockefeller trying to build one. The structure of the category makes a Standard-Oil-style rollup nearly impossible, because the inventory is tied to specific physical locations owned by specific counterparties with specific local relationships.

(Brief tangent on this third point: the reason there can't be a Standard Oil of out-of-home is that the inventory is physically constrained in a way digital inventory is not. If Meta wanted to double its ad inventory tomorrow, it could. Just show more ads per scroll. If Google wanted to triple its inventory, it could. Index more sites, show more placements. But you cannot double the number of billboards in Manhattan. The permits are finite. The buildings are finite. The highway frontage is finite. The FAA restricts what can be built near airports. The DOT regulates what can sit on Interstate right-of-way. Physical reality is the moat, and physical reality is not disruptable by a software release.)

Out-of-home will not replace digital. I want to be completely clear about this, because the maximalist version of the argument (billboards will eat Google) is stupid, and I don't believe it, and I'm not making it. Digital will continue to be the dominant form of advertising for as long as any of us are alive in this industry. What out-of-home does, structurally, regardless of whether anyone planned for it to, is exist as proof that a transparent advertising market is still possible.

In a world where 83% of digital ad revenue is projected to be inside walled gardens by 2027, the existence of a major channel that cannot be walled is a corrective pressure on the whole system. Not because advertisers will move their whole budget to out-of-home. They won't, and they shouldn't. But because having an alternative with published prices, advertiser-owned data, and independent measurement changes the negotiating position of every CMO in every room with every platform. The existence of the alternative is itself a form of regulation. It is the Interstate Commerce Act, but it's a market condition rather than a statute. It's what happens when one piece of the infrastructure refuses to be enclosed.

I don't think the people who work in out-of-home fully appreciate this yet. I talk to media owners all the time. Family billboard companies in Oklahoma, transit operators in Boston, place-based media executives in Los Angeles. Most of them see themselves as operators of a slightly old-fashioned business that happens to be profitable. They do not yet see themselves as the last functioning open market in American advertising. I think they should. I think that's what they are, and I think that's why, in a decade where everything else in advertising has either been absorbed by five companies or structurally flatlined, the category has quietly, unheroically, kept growing.

The thing out-of-home is, in 2026, is the thing oil in Texas was in 1905 after Spindletop. Not the end of Standard Oil's story. Just the evidence that Standard Oil's story wasn't the only one that could have been written.

IX.

I want to tell you something personal, because the essay is long enough already that I feel like I've earned the right to get a little personal, and because I think it matters for what comes next.

I joined the company I now run in the middle of 2018, which, if you've been paying attention to the arithmetic of this essay, is the exact year the ad market outside the top five platforms stopped growing. This is not a coincidence I planned. It's a coincidence I noticed in retrospect, writing this essay, while pulling WARC data for the passages you read twenty minutes ago. When I came in, I was an early employee at a small company trying to build a technology layer underneath an industry (out-of-home) that was badly underserved by software. We had a working product and not very many customers. I spent my first year basically optimizing Google Ads and cold-emailing people.

In July of 2018 (a couple of months before TikTok launched globally, for those of us keeping a mental timeline, though none of us knew at the time that that date was going to matter), I was on a plane somewhere and had a conversation with a marketing executive of a consumer brand you've heard of. He was explaining to me why he was spending 80% of his budget on Facebook and Google, and he said something that I have been quietly turning over in my head for almost eight years. He said something along the lines of: "I know the measurement is probably bullshit. But it's the only measurement we've got, so we use it, and we optimize against it. What choice do we have?"

What choice do we have.

That was the sentence that put me on the path to the argument in this essay, although I didn't know it at the time. Because the answer to what choice do we have is, when you look at it historically, almost always more choices than it currently feels like you have. It felt, in 1878, like every oil producer in Pennsylvania was going to lose. Most of them did. Some of them didn't, and the ones who didn't mostly didn't because they found a part of the industry that Rockefeller couldn't easily control. A specific refining technique, a specific regional market, a specific piece of logistics that Standard Oil hadn't yet encircled. The producers who survived the Gilded Age were the ones who figured out which parts of the industry could not be enclosed, and who built their positions there.

I spent the next six years watching the advertising industry consolidate at every layer simultaneously, and I spent that time, mostly, building software for the one category that structurally couldn't be enclosed. Not because I had a grand plan about it. I didn't. I was mostly just trying to make buying and selling billboards easy. But somewhere around 2022, after the cookie announcements and the iOS privacy changes and the first wave of agency consolidation, I started to realize that the category I was in was not a sleepy backwater. It was, structurally, the one remaining place in American advertising where the market still worked the way markets are supposed to work. Published inventory. Known prices. Data that belonged to the buyer. No platform grading its own test.

I became CEO in the spring of 2024. By that point, the Gilded Age parallel had moved from "interesting historical frame" to "something I thought about most days." I started reading more Gilded Age history. I read Ron Chernow on Rockefeller and Morgan. I read Matthew Josephson. I read Ida Tarbell's actual 1904 exposé of Standard Oil, which is, genuinely, one of the most underrated pieces of business journalism ever written. Everything kept confirming the pattern. Everything kept confirming that what I was watching happen in my industry had happened before, and that the shape of the resolution was already available if anyone wanted to look it up.

Which is part of why I'm writing this essay. I am aware that I am biased. I run a company in the exact category I'm defending. If you finish this essay thinking "the author was not neutral," you will be correct. What I'd ask you to do is not to discount the argument because of who's making it, but to test it on its own terms. The WARC data is not mine. The Omnicom-IPG deal is not mine. The Clear Channel take-private is not mine. The 90% market share figure for Standard Oil is not mine. The Gilded Age happened to people a century before I was born. I'm just describing the shape.

The reason I started this essay by telling you about Jay Gould is that Jay Gould keeps showing up in my head, uninvited, most days. I spend a lot of my professional life watching the same pattern reassemble itself. That's the honest version. Make of it what you will.

X.

Okay. Let's deal with the thing. I can feel you thinking it. I would be thinking it if I were reading this. The author runs an out-of-home advertising software company and has just spent nine thousand words arguing that out-of-home advertising is structurally important. Of course he thinks that. He'd be fired if he didn't. Why are we listening to this?

You're right. I am biased. I have a stake in the outcome. I would not be writing this essay if I did not run this company. If any of that makes you want to discount the argument, I understand. I would do the same in your position.

The thing I'd ask you to do, though, is not to discount the argument because of who's making it, but to test it on its own terms. Strip out the CEO. Strip out the company. The facts of the essay don't need me in them. The three-layer consolidation is happening whether I write about it or not. The Omnicom-IPG deal closed in November 2025 whether or not anyone in my business noticed. The Clear Channel take-private went through. The walled gardens are heading toward 83% of digital ad revenue by 2027 independent of anything I think about it. The Gilded Age parallel holds or doesn't hold based on the historical and industry facts, not based on who's drawing it.

What I'd offer in return for you discounting my conclusion is this: I have been sitting with these facts for eight years. If I'd wanted to make an easier argument, I'd have made one. The easier argument is "here are four case studies of successful campaigns, please book a demo." I have written that argument before. It is a perfectly fine argument. It is also not this essay. This essay is the one I actually think is true, and I'm writing it at ten thousand words because the shorter version doesn't capture the shape of what's happening.

The biased version of the argument and the true version of the argument can coexist. I think they do. You are welcome to hold me accountable if they don't.

And now that I've made the concession, I want to use the remaining time to say something specific about what I think happens next, because I think an argument that doesn't make a prediction isn't really an argument. It's a diagnosis.

XI.

Three forecasts. Each falsifiable. I'm putting them in writing so you can come back in five years and tell me where I was wrong.

Forecast one: Regulatory pressure on walled-garden measurement, not walled-garden existence. Over the next five years, the political and regulatory pressure on Google, Meta, and Amazon will focus increasingly on what they're allowed to keep proprietary. Identity graphs, measurement methodology, auction transparency, data portability. Not on whether they're allowed to be big. The playbook will look more like 1887 (Interstate Commerce Act) than 1911 (Standard Oil break-up). This will be unsatisfying to people who want a dramatic antitrust moment. It will also be, in the long run, more durable, because measurement transparency is a harder moat to rebuild than corporate structure.

Forecast two: Advertiser-owned measurement becomes table stakes. The CMO of 2030 will not accept "the platform says it worked." They will demand independent measurement that they own, can export, and can compare across channels. The companies that win the next decade of advertising technology will be the ones that operate the transparent layer. The marketplaces, the measurement networks, the cross-channel attribution systems. Not the ones that operate the walled gardens. The walled gardens will still be profitable. They will also, gradually, become less central to how media planning happens, because the measurement moat will erode. This is a bet, and I am making it in public, and I will live with the consequences of being wrong about it.

Forecast three: The consolidation at all three layers continues, and then reverses at the media-owner layer. Over the next five years, the Big Five holding companies will probably become the Big Three. The platforms will not lose share. They will consolidate it. But the media-owner layer, which is currently consolidating via the Clear Channel take-private and the ongoing rollup among independents, will eventually re-fragment as new independent operators enter the market, drawn by the structural advantages of the categories the walled gardens can't reach. The action will move to the physical edges of the industry, because that's where the rails can't be owned the way software rails can.

Any of these predictions could be wrong. Probably at least one of them is. I'm putting them in writing anyway, because I think the worst thing an essay like this can do is dodge making a claim while pretending it has.

The next ten years of advertising are going to be more interesting than the last ten. I don't say that casually. I say it because the structural conditions that made the last ten years what they were are starting, finally, to crack. The walled gardens are under pressure they haven't been under before. The measurement moat is narrower than it looks. The physical advertising category is growing in a way most industry observers haven't noticed yet. The Omnicom-IPG deal has produced the largest advertising-industry labor dislocation in modern memory, and 14,000 newly unemployed advertising professionals, most of them very good at their jobs, are about to start companies.

The industry is not stuck. It only looks stuck from inside the 2018-to-2026 window. Step back, and the pattern is the early phase of the resolution, not the endless extension of the problem.

That's what I think is true. I could be wrong.

XII.

Jay Gould died in 1892, at the age of 56, of tuberculosis. His estate, adjusted for inflation, was worth approximately $22 billion in 2026 dollars. He spent most of his last decade being sued. By the federal government, by competing railroads, by widows of shareholders whose savings he had incinerated in various rounds of stock manipulation, by Abel Corbin (yes, that Abel Corbin. The relationship did not survive Black Friday, which you can imagine), and by an entire category of smaller investors who had been ruined by Gould's various campaigns.

The thing that ended his grip on American finance was not a competitor. It was not a better product. It was not a more ethical rival who outworked him. It was, slowly, over years, a legal framework that made what he had done, retroactively, a crime. The Sherman Antitrust Act was passed in 1890. Gould didn't live long enough to see it applied in full force. The bigger actions came a decade after his death. But the structure that would unmake the Gilded Age was already being built while he was still alive. He could see it coming. He did not have the legal framework to stop it. He had the money. He didn't have the laws.

I am not predicting that for Google. I am not predicting it for Meta or Amazon or Omnicom or Publicis or any of the media companies currently taking themselves private or getting acquired. I don't think anyone currently running these companies is committing a crime. Nobody in the advertising industry is trying to corner a gold market with their wife's sister's husband. The companies in question are, mostly, well-run, led by capable people, producing genuinely useful products, hiring thoughtful employees, and doing what any rational actor in their structural position would do.

But the shape of the correction, when it comes, and it's coming, will probably look more like the 1890s than like whatever we've been expecting. It will be quieter than an antitrust movie suggests. It will take longer than reformers want. It will look, in the moment, like regulatory pressure and market repositioning and a few strategically timed essays on LinkedIn. In retrospect, it will look like an industry becoming legible.

That's what I've tried to describe here. I don't know if I got it right. You've made it to the end of ten thousand words about advertising and the Gilded Age, which either means I got enough of it right to keep you reading, or that you have too much time on your hands. Either way, I appreciate the attention.

The billboard on I-80, outside Des Moines, standing in the field, the one that doesn't have a Google account, that cannot be bought inside Ads Manager, that reports its impressions to no walled garden, is going to still be there in ten years. That is not a prediction. That is a property of physical infrastructure. The question is not whether it will be there. The question is whether the rest of the advertising industry, over the next decade, will come to look a little more like it. A little more transparent, a little more legible, a little more accountable to the people paying the bills. Or whether the billboard will remain, as it is now, one of the last places in the market where the market still works.

I know which outcome I'm rooting for. I also know which one the history books, eventually, tend to deliver. The Gilded Age was not permanent. Nothing structural in capitalism is.

Jay Gould's estate was worth $22 billion in modern dollars. The thing he couldn't buy, at any price, was the thing that would eventually be written about him. I think about that most days. I thought you might want to know.

— Chris

Citations available here. If you find an error, please write me. I'll fix it.

Citations available at adquick.com/blog/gilded-age-sources. If you find an error, please write me. I'll fix it.