There was a time when the list of places you could buy serious digital advertising fit on an index card. Search, social, display exchanges, maybe video. That card is now a phone book. Your grocery delivery app sells ads. Your bank pilots ads. Rideshare apps sell ads. The gas station screen sells ads. Travel apps, dating apps, weather apps, point of sale terminals, smart TVs down to the remote button. Every product manager at every company with an audience has now sat in the same meeting and heard the same sentence: we have first party data and user attention, so we are leaving money on the table.
I understand the meeting because I have been in versions of it on both sides. The math is seductive. Advertising revenue is high margin, it monetizes an asset you already own, and retail media handed everyone the playbook plus the excuse. Amazon built a colossal ads business out of shelf space and search bars, and every board in the world noticed. The result is a structural shift I call convergence: media companies became commerce companies, commerce companies became media companies, and now every app with daily active users is converging on the same business model from different directions.
My thesis for this paper: the proliferation is real and mostly rational for the sellers, but it is creating a buyer side crisis that the industry has not priced in. A hundred subscale ad networks with incompatible measurement is not a hundred opportunities. It is a tax on marketing organizations, and the shakeout, which I expect to run from 2027 through 2030, will sort these networks into a small tier that earns permanent budget lines and a long tail that quietly becomes white labeled inventory inside someone else's demand platform.
Key Findings
- Ad monetization has become the default second business model for any app that achieves scale. The strategic question inside platforms is no longer whether, but how much before users revolt.
- Every one of these networks is selling the same two ingredients, closed loop data and captive context, and most are overpricing both.
- The buyer side cost of fragmentation, integration, measurement chaos, and minimum spends, is the industry's least discussed problem, and I give it a name: the Fragmentation Tax.
- A network's durable value tracks one variable above all: proximity to a transaction it can actually observe. I formalize this as Checkout Distance.
- Consolidation is inevitable. The long tail of app networks will not die, it will federate, selling through aggregators and losing pricing power on the way.
- Advertisers who ride convergence well treat new networks as data partnerships first and reach second.
The Gravity Pulling Everyone In
Why is this happening to every app at once? Three forces, stacked. First, the privacy reshuffle raised the value of first party relationships, and apps own logged in users with declared behavior. What the open web lost, the apps inherited. Second, growth economics: most consumer apps hit market saturation, subscription fatigue capped direct monetization, and ads are the only lever left that scales revenue without scaling users. Third, tooling collapsed the barrier to entry. A decade ago building an ad business meant building an ad server, a sales force, and an auction. Today the entire stack is rentable: sponsored listing infrastructure, off the shelf auctions, clean rooms, self serve portals. Monetization went from a moat to a plugin.
None of this is irrational. What is irrational is the collective belief that advertiser demand will expand to absorb every new supply pool at premium prices. Budgets are finite and marketing teams are more finite. Something has to ration, and what rations is buyer attention, which is the great irony of the whole cycle: the industry that sells attention is now starved of it on its own buy side.
Checkout Distance: The Only Ranking That Matters
When a new network pitches me, I run one test before any conversation about audiences or formats. How many steps sit between the ad exposure and a transaction this platform can see with its own eyes? I call this Checkout Distance, and it sorts the entire convergence landscape.
Distance zero: the ad lives inside the purchase flow and the platform observes the sale. Retail media search placements, marketplace sponsored listings, delivery app promotions. These earn performance budgets because attribution is native, not modeled. Distance one: the platform observes strong intent but not the final sale, travel search, real estate browsing, product research surfaces. Valuable, negotiable, frequently oversold. Distance two and beyond: the app has attention and demographics but no commercial signal in the loop, and it is effectively selling display with better wallpaper. There is nothing wrong with distance two inventory except its price, which is usually set by analogy to distance zero neighbors it does not resemble.
Checkout Distance also predicts how these networks behave under pressure. When budgets tighten, distance zero networks hold rates, distance two networks discount, and distance one networks invent new attribution methodologies. Watch for it.
The Fragmentation Tax
Now the buyer side ledger, which nobody selling convergence wants to itemize. Every additional network on a media plan costs an integration, a contract, a measurement reconciliation, a creative spec, a billing relationship, and a slice of a human being's week. Multiply by forty networks and you have the Fragmentation Tax: the overhead that eats the incremental value the long tail theoretically offers. Back of the napkin, suppose a mid sized advertiser adds twenty small networks averaging a few thousand dollars a month each. The media math might clear. Add the loaded cost of managing twenty incompatible reporting schemas and quarterly business reviews, and the program is underwater before the first conversion is miscounted.
The tax explains the next phase of the market better than any technology argument. Buyers will not manage a hundred relationships, so intermediaries will re aggregate the fragments, demand platforms, retail media aggregators, and increasingly the autonomous buying systems that treat networks as interchangeable endpoints. Fragmented supply plus overwhelmed buyers is exactly the environment where machine buying wins, because software does not experience the Fragmentation Tax the way a media team does. The long tail's future is federation: living inside someone else's pipes, at someone else's prices.
There is also a user side of the ledger. Apps underestimate how quickly ad load erodes the trust that made their audience monetizable, especially in utility and finance contexts where the user's job to be done has nothing to do with discovery. Platforms squeezed organic surfaces for years and trained users to route around them, a dynamic I documented in the death of organic reach. The apps now repeating that squeeze on their own core experience should study how that story ends.
What Convergence Does to the Big Platforms
The incumbents are not watching passively. Search and social giants face, for the first time, real competition for performance budgets from networks that own transactions they can only model. Their counter moves are predictable and already visible in outline: buy or build commerce closer to their own checkout, court the retail networks as measurement partners to stay in the attribution loop, and position their buying tools as the aggregation layer for everyone else's inventory. The endgame they want is clear. If they cannot own every network, they will own the demand pipe through which every network is bought.
For advertisers this creates a delicate moment. The aggregation offer will be genuinely convenient and quietly recentralizing. Accept it wholesale and the transparency you gained from working with transaction rich networks directly gets modeled away again behind a familiar dashboard. Convenience is how the industry lost measurement independence the first time. It would be professional malpractice to lose it the same way twice.
What I Would Do About It
Rank every network on your plan, and every one pitching you, by Checkout Distance, and force the price conversation to happen inside that frame. Pay performance rates only where the platform observes the transaction natively. Everything else is context and reach, priced accordingly, no matter what the deck says.
Cap your relationship count deliberately. Pick a single digit number of networks you will manage directly and deeply, where the joint data work compounds, and buy the long tail only through aggregation with strict incrementality testing. A network too small to justify a data partnership is too small to justify a direct relationship.
Charge the Fragmentation Tax back internally. Put a loaded cost per managed network into your planning model so that adding the twenty first network has to beat not just its media benchmark but its overhead. Teams that do this cut their tail spend within two quarters, and performance rarely notices.
And if you are on the other side of the table, an app considering the ads meeting, respect Checkout Distance before you hire a sales team. Build measurement before monetization, cap ad load below what your revenue model wants, and read Paper No.53 on outcome markets before you price anything, because the buyers of 2028 will demand settlement terms the sellers of 2026 are not ready to offer. The window where subscale networks command premium pricing is already closing. Enter humble or do not enter.