Run the Numbers Sep 21, 2026 51m 34m saved
With Blake Saunders, founder and Managing Director of Core Advisors
Bending Spoons bought AOL for about $1.5 billion, at roughly four times EBITDA. Six months later the company went public and trades at about 16 times.
Private equity would not touch it. Blake Saunders says the buyers were too frightened of what AI is doing to consumer habits to look at what AOL actually is, which is a subscription business with paying customers.
"The problem is private equity is so freaked out about AI and what it's doing to consumer habits that they were basically able to use this contrarian position to look past that and say, this is a real subscription business."
Saunders has spent more than 15 years in media investment banking and started his own firm last year. He sells media companies for a living, which means he knows which revenue lines buyers pay up for, which ones they discount, and what a multiple is actually measuring.
The full interview is covered here so you can skip it. 51 minutes of audio, 17 minutes of reading.
Here are the 10 lessons that matter.
Key Takeaways
Subscriptions get the highest multiple and events get the lowest, because buyers pay for what is repeatable, programmable and scalable
A media company can run 50 events, tax the team and generate almost no contribution
Advertisers now pay media companies to make them show up inside AI answers, a service the industry has already named GEO
Consumers have never paid the full cost of content — subscriptions were 15% to 25% of a newspaper's revenue at the height of the business
Media's monetization risk is the opposite of software's: content is worth most the day it is published
Vice raised, produced and never synchronized audience with advertisers, and went through bankruptcy
Affiliate media trades at 3 to 5 times EBITDA, subscription businesses in double digits and creator businesses around 15
A big podcast deal can be a loss leader — the network wins the talent's advertisers, not their IP
Bending Spoons bought AOL at about 4 times EBITDA and trades at about 16, six months later
The B2B rollup math is buy a platform at 10, bolt on at 4 to 6, sell at 12
Time kills deals, which is why the banker's job is partly to keep everyone even-tempered at 3am
1. Media's New Audience
Saunders defines media as content distributed to affect some kind of change, and says the thing being changed is no longer only a person.
The audience now includes the models
I was originally thinking to affect people, but I think more and more what you're seeing is people are realizing that they need to also also impact the LLMs and these agents
Blake Saunders
That is not theoretical for anyone publishing. Newsletter and podcast subscribers now arrive from ChatGPT, Perplexity and Claude, which was not happening a year ago. The episode put numbers on the gap: a newsletter with 500 articles behind it picks up 25 to 35 new subscribers a day from search and one to three from the models — a long way behind, but a channel that did not exist.
Advertisers have noticed faster than publishers have. Saunders says they now come to media companies with a specific demand: a chief executive asked what the best company in a category is, the model did not name them, and they will pay whatever it takes to fix it. An industry has already grown up around it, with new agencies standing up quickly and media companies adding a product line they call GEO, the search-optimization business rebuilt for large language models. What makes it unstable is that the models keep changing, so the methods that work keep changing with them, and Saunders expects the engines to get better at spotting a company that has simply arranged to be mentioned across a lot of publishers.
He also notices it in his own working day, where more of his time goes to talking to agents than doing the work directly.
2. The Monetization Ladder
Asked how media makes money, Saunders starts with attention and points at the analyst who mapped it.
He credits Mary Meeker, formerly of Kleiner Perkins and now running Bond, for the slides that showed the gap between where attention sat and where the money sat: newspapers held less attention than their revenue implied, digital held more, and the gap closed over 10 or 15 years. What is left is five ways to earn: advertising, subscriptions, licensing IP, events and data.
The ranking between them comes down to one test.
Buyers pay for what repeats
this is something that like you want to beat the drum when you're looking at media companies is like what's repeatable, programmable, and scalable
Blake Saunders
Subscriptions win on that test, because a credit card on file supports cohort models, acquisition costs and retention curves, and public and private buyers both pay up for it. Advertising is usually the largest line, and buyers examine how long the advertisers have been there and how they feel about the product. Events sit at the bottom, because the business restarts every year and has to persuade ticket buyers and sponsors to show up again.
Fifty events is work, not a business
a media company can have 50 events, but it's taxing the team. It's really not generating that much contribution
Blake Saunders
Gustafson pushed back from the sponsor's side: an event puts a warm body in a room and is completely attributable. Saunders' answer was that attribution is exactly what events lack in practice. Companies allocate the money and then send the wrong people to activate it, and a big brand can charge $50,000 or $75,000 for a dinner that yields nothing. Programmatic advertising can be tied to a return on ad spend; an event ends up as a story about touching someone who converted six months later.
3. Where The Data Pays
Asked whether the ranking changes between business and consumer media, Saunders said the logic holds in both, with two differences.
In B2B, conferences and exhibitions can carry real revenue because brands have budget for exhibiting. In consumer media the large events exist but are paid for by advertising.
The line that is changing fastest is data, and he means it in two directions at once: data that gets brands to spend more, and data worth feeding to a model.
Companies are finding assets they did not know they had
people are bumping into like hidden gems where they're like, oh wow, this data that we've been recording through this app, calorie trackers, like all these things, it basically changes the complete perspective of the company
Blake Saunders
4. Why Advertising Wins
Gustafson came from software, where advertising is looked down on and subscription is treated as the higher form, and asked why, given that Google, Meta and Amazon fund themselves on it.
Saunders' answer is that the platforms sell in an auction and everyone else sells by hand.
One media property sells advertising in hand-to-hand combat
So a one-off media property, it's like hand-to-hand combat where you're selling ad sellers into the field effectively and having meetings and having to basically pitch and say, this is why you should advertise with us.
Blake Saunders
The platforms do not have to persuade anyone, because they hold the attention and there is always another bidder. If one advertiser walks, the next is already in the queue. The bargain they offer, which is to hand over all the data in return for the platform finding the customers, has a consequence built into it: the platform learns how much money the advertiser is making and is structurally motivated to take more of it. Customer acquisition costs rise, and on Saunders' reading they should, until the point where it stops being profitable for the advertiser.
The advertisers bid the price up themselves
it's the most beautiful business model in the world in the sense that you get your own customers to increase the prices so you're not the bad guy
Blake Saunders
The other half of the advantage is signal. Google's utility layer, Gmail included, produces data that makes the advertising products better.
5. Nobody Pays Full Price
Gustafson's next observation is one Saunders says is impossible to unsee: the social platforms do not pay for the content that keeps people there. YouTube seeds premium video, and the ecosystem around it funds itself — he describes live-streamers in their twenties carrying expensive cellular backpacks so their signal never drops. Every platform has tried a program to fund premium video; what actually makes them addictive is a parasocial relationship with creators, friends and family, which competes with time people would have spent talking to each other anyway.
Then he widens it into the claim the rest of the section rests on.
Consumers have never covered the cost of what they consume
one of the fallacies in media, and this goes back to newspapers, is like consumers have actually never paid the full cost of the product
Blake Saunders
At the height of the newspaper business, before the internet, subscriptions were 15% to 25% of total revenue; advertising, classifieds and other lines paid for the rest. That is why Netflix has added advertising and why Hulu ran subscription and advertising together from the start. A viable product, in his framing, has to earn out on advertising because the content is worth more to the consumer than the consumer will pay.
Gustafson's counter-example was a newsletter, where the marginal cost is what you pay Stripe, and where charging a subscription and running an ad feels like too much. Saunders' answer: an ad done well is not distracting and should carry some value of its own — people complain about the countdown on a streaming service and then watch the Super Bowl for the commercials.
6. Hold Or Sell
Saunders' history lesson is short. The internet took a sledgehammer to media monetization. Before it, families like the Hearsts and Condé Nast held portfolios of newspapers that became radio and television, with EBITDA margins of 20% to 40%, and a structural protection: there were four major affiliates in a market and no fifth station coming. Ownership rules kept a single owner from holding the newspaper and the radio station, which preserved a diversity of voices and, incidentally, the economics.
Google broke it by being better at finding the local mechanic a customer. Yellow Pages went to zero.
The families that survived did it by allocating capital somewhere else. Hearst built B2B investments and broadened what it puts money into, on the view that buying more television stations was not a good use of it.
The risk profile of a media business is also the opposite of a software business, and this is the part Saunders is most emphatic about.
Content is worth most the moment it lands
your content, the most monetizable event is at the beginning
Blake Saunders
Raise money, produce a library, and if the audience is not there at the moment of publication, the ability to monetize it may be a tenth of what it was six months earlier. His worked example is Vice, which had a large built-in audience and then added content faster than it synchronized audience growth with advertiser demand.
Vice built the machine and never matched it to the money
They basically just had this massive content creation machine that wasn't properly monetized with the audience and advertisers.
Blake Saunders
Gustafson's framing of it is that a media company grows two customer bases at once, like a marketplace, and they have to move together. Saunders agreed, and said that is what creator-led media gets right: one voice, a scaling audience, and almost no cost of production. A newsletter is a keyboard and a computer, and you pay the platform when you have revenue, which is why creator businesses reach scale without raising money and then add podcasts and events on top.
7. What Multiples Mean
Confirmed on the record: media businesses trade on a multiple of EBITDA, not revenue.
Where those multiples sit depends on what the buyer thinks happens next.
A multiple is arithmetic, not a verdict
a multiple is being guided by like the buyer's perception of like future growth
Blake Saunders
Affiliate businesses that depend on the open web for their audience trade at three, four or five times EBITDA. Subscription businesses with predictable revenue, and he names The New York Times, are in double digits. In fast-growing categories such as creators, he hears 15 times. A corporate buyer is working back from an internal hurdle rate, and some corporate development teams make it harder on themselves by refusing to credit a terminal value.
He is blunt about how the headline deals get read.
The multiples people quote are outliers
most of the buyers in the space want profitability. Like they're not going to just pay a revenue multiple.
Blake Saunders
And behind a famous price there is usually less competition than the number implies: in his experience of the bigger deals, there were not 10 buyers lining up but about one and a half, where the half was an interested party used to convince the real buyer it had competition.
8. Podcast Deals And IP
On libraries, Saunders will not generalize. Music rights consolidated hard while interest rates were low and he doubts valuations go much higher. Video libraries were the ones people struggled to argue for, on the view that nobody wants a 40-year-old film, but the ability to remake old material with AI may be pushing that value back up.
Podcasting is the medium he thinks is genuinely evergreen, with a measurement caveat: subscribing to a show can make an app download the whole back catalog, so engagement numbers are not a clean read. What makes the format valuable is that a good episode with a guest who did not do 20 other shows keeps earning. He uses Acquired as the example, which publishes its rates and, he believes, monetizes about 80% of the back catalog at the same level as current episodes — one recommendation from a large account can drive 20,000 more downloads of an old episode.
The big talent deals are advertising arbitrage.
A network buys the ad sales, not the show
it sounds like he signed a multi-year $80 million Distribution and ad sales agreement. That means they don't own his IP.
Blake Saunders
That was Dax Shepard's deal. Saunders assumes it carries a consistency requirement and a clawback if the show stops publishing, and says the economics may not clear on the show alone: a network that wins a large podcast wins access to its advertisers, and if it captures $1 million of a brand's budget this year it expects to double that next year across five more shows. Retention matters as much as acquisition, and some of these deals have not sold through at the level the buyer expected.
Underneath all of it is the thing he says people underestimate.
A sales organization cannot be switched on
I think what people underestimate is how hard it is to build a strong sales effort and the robustness that it takes.
Blake Saunders
Spotify has been building one for years, and the reason it is defensible is that budget allocation is a relationship business as much as a data one.
9. Bending Spoons Buys AOL
Asked who the smart investors in private media are, Saunders said the record is poor across the board and pointed at one recent example instead.
Bending Spoons, the Italian company that went public a month and a half before the recording, spent about $1.5 billion on AOL and also owns Evernote. Saunders calls it Constellation Software for the bargain bin of the internet, and he has written an S-1 breakdown of it. Half the business is AOL.
The arbitrage was four times to sixteen times in six months
They bought AOL for roughly 4 times EBITDA, and now they're trading at like 16 times EBITDA, and that's like 6 months later.
Blake Saunders
What made it available at four was fear. Private equity could not see past what AI might do to consumer habits, which let a contrarian buyer take a real subscription business with real customers at a distressed multiple. The market has rewarded it so far.
The value-creation plan for most media buyouts is less exotic. In B2B media it is a rollup: buy a platform at 10 times EBITDA, acquire smaller fragmented businesses at four to six times because they are smaller, and sell the combined thing at 12 when comparable businesses trade at 14. On the consumer side, the newer money goes into creator channels that already throw off cash, and pays for the apparatus — more content, a bigger audience, more advertising.
10. Time Kills Deals
Gustafson said he has been through deals with about seven moments where each one looks dead, and asked how much of dealmaking is surviving them. Saunders said part of his job is psychological: stop the founder counting money they do not have, and keep the work moving on the days the other side goes quiet. People who have spent 5, 10 or 15 years building something get caught in the swings and stop answering diligence questions, which is how a process actually dies.
The enemy is elapsed time, not the counterparty
I think time kills deals and like at 3 o'clock in the morning, I've had lawyers decide that they want to stop working and I had to call everybody and wake 'em up and say, you're going to get this done right now.
Blake Saunders
Nothing is destined to close. He has had deals blow up the day before signing for reasons nobody controlled, and he points at how quickly the ground moves now: a new feature from OpenAI in the middle of a process is enough to frighten every buyer in it.
Nothing counts until the money lands
It ain't real until the check clear.
Blake Saunders
Bonus Insights
Content to commerce works when the product is not the media company's problem
Gustafson raised Doug DeMuro, whose YouTube reviews led to Chernin Group backing him and to Cars and Bids. Saunders' read is that the media property is the front door and the business behind it is something that would otherwise be a commodity; the advantage is earned media and an audience extension for free. It works for DeMuro because he can talk about the cars on the auction platform. He thinks Hodinkee has been hard, and that the reason is the watch business rather than the model.
Creators are about to start shipping software
Because building tools is now cheap, Saunders expects creators to turn what their audience keeps asking for into products: vibe-code the first version, hand it to one developer, sell a subscription. He also points at courses selling for as much as $10,000, which he frames as monetizing the knowledge behind a two-minute clip — MrBeast's audience is children, but his operating knowledge from running Feastables is what people will pay for.
Bundling is being rebuilt by the AI platforms
Gustafson totted up Hulu, YouTube Premium, YouTube TV and Disney Plus at about $250 a month and asked whether rebundling is coming. Saunders said everyone has the same moment when they see their subscriptions added up, that bundles are hard because two companies have to agree whose product is stronger, and that some subscriptions are now paid for a better experience rather than for access at all. He expects the AI engines to launch their own subscription products, with publishers pushing back on their content being taken, and something like Apple News assembling publishers under a technology layer.
A family buyer does not need a valuation
Asked how much of media dealmaking is narrative, Saunders said scarcity and story matter up to a point and the financials arrive anyway — but that he has been on the other side of a chief executive or family owner saying they want the asset and will pay what they want. Gustafson's comparison was a conversation with a former MLB chief financial officer about sports teams: there are only so many, and the price is whatever someone pays. Saunders' addition is that teams have worked out over 30 years, partly as legacy assets, partly because there has always been a next buyer.
The lightning round
More valuable, a podcast or a conference: podcast. Best media acquisition ever: AOL. Worst: the telecom companies buying cable and content, on the theory that owning consumer attention across several products would produce a magical advertising moment. It did not, because none of them built a unified ad product that could take budget from Google and Meta — which is why, he says, they are all deconsolidating now.
The book he gives to operators
The Outsiders, William Thorndike's study of capital allocation. Saunders' point in recommending it is that people in M&A get lost in multiples when the job is the basic allocation decision, meaning buybacks, dividends and acquisitions, which is the part a chief financial officer actually controls.
Saunders' bottom line is that media value follows whatever is repeatable, programmable and scalable, that consumers have never paid what content costs so advertising is not optional, and that the best current example of both is a contrarian buyer taking a subscription business off a frightened market at four times EBITDA.
Products, Companies & Tools Mentioned
Core Advisors (Saunders' own boutique investment bank, started last year, advising media companies on deals)
Bending Spoons (Went public a month and a half before the recording; bought AOL for about $1.5 billion at roughly 4 times EBITDA and trades at about 16)
AOL (Half of Bending Spoons' business and, on Saunders' telling, a real subscription business private equity was too frightened of AI to buy)
Evernote (Another of Bending Spoons' purchases in what Saunders calls the bargain bin of the internet)
Google and Meta (The auction model that made advertising the most profitable business in media, with Gmail's signal data behind the targeting)
Amazon (Named with Google and Meta as an empire funded by advertising)
YouTube (Seeds premium content and increasingly funds or equips the creators who make the rest of it)
Netflix and Hulu (Both earn out on advertising because subscriptions alone do not cover what consumers expect)
The New York Times (His example of a subscription business trading in double-digit multiples)
Vice (Raised heavily, built a content machine, never synchronized audience and advertisers, and went through bankruptcy)
Hearst and Condé Nast (The newspaper families that diversified early; Hearst into B2B)
Acquired (Publishes its rates and, he believes, monetizes about 80% of its back catalog — his model of evergreen podcast content)
SiriusXM (The best audio monetization team, buying talent partly as a way to bring advertisers into the rest of the network)
Spotify (Has spent years building the sales organization that makes audio advertising defensible)
Apple News (His example of a technology layer assembling publishers into something like a bundle)
Substack and beehiiv (Newsletter platforms a creator pays once there is revenue, which is why those businesses scale without raising money)
Monarch (The app that showed Gustafson he spends about $250 a month on streaming subscriptions)
Cars & Bids (Chernin Group's marketplace built behind Doug DeMuro's audience — the content-to-commerce case that works)
Hodinkee (The content-to-commerce case he calls hard, because selling watches is hard)
Feastables (The operating experience behind MrBeast that Saunders says people would pay for a course on)
Bond (Mary Meeker's firm; her attention-versus-monetization slides are where Saunders starts)
OpenAI, Perplexity and Claude (Now sending subscribers to publishers, and the reason media companies are selling a service called GEO)
Books & Resources Mentioned
The Outsiders – William Thorndike (The book Saunders gives operators, because capital allocation is what a CFO actually controls)
Business Breakdowns (Saunders' earlier appearance on the Colossus show, where he answered the same question differently)
Saunders' Bending Spoons S-1 breakdown (His written analysis of the company behind the AOL deal)
Listen to the full episode
🔴 YouTube
Watch the full episode:
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: