What this covers
This is a conversation on Making Media in which Peter, with Matt and Dom, surveys the media business. Nearly all the substantive claims come from Peter. He starts with local news, which he says has no working for-profit model. Newspapers were funded by the bundle and a distribution monopoly that made readers pay implicitly, and the internet broke both. He treats that claim as one the industry hasn't faced. He also says the New York Times nearly failed around 2008 and took a high-interest loan, so its subscription success is no template, and that subscriptions can't be every company's pivot because people will only pay for so many things.
The middle of the conversation covers distribution and new formats. Peter traces platform dependency to 2005, when a business-side Forbes staffer's key job was getting stories onto Yahoo's homepage. He expects the codependency to persist because readers arrive through side doors. He calls YouTube roughly the fourth-largest cable provider in the U.S., hard for analysts to read because of its algorithmic personalization, and a different product from TV: creator-to-television moves have failed, and YouTube's own movie effort was dropped. Podcasting is early and its bubble small, and he compares it to television, which began as radio plays. AI, he says, is already mostly harmful, producing a tragedy of the commons of near-free mediocre content.
The last stretch turns to curators and cycles. Netflix shows how fast conventional wisdom flips, and Peter argues it could abandon a successful business partly because it wasn't a legacy company. He questions how special Casey Bloys and Richard Plepler's HBO reputation is, noting that music executives are mostly number-crunchers. He treats personality-led media as cyclical and capped in size, and audience capture as the price of serving a small audience that cares.
Kafka argues that media conventional wisdom flips every few years, that durable media businesses depend on a specific audience that truly wants the product rather than on data or scale, and that platform dependency, AI, and the subscription shift are reshaping economics in ways that are often misread.
- Local news lacks a working for-profit model because the internet unbundled the newspaper's distribution monopoly and nobody has funded local gathering sustainably
- Media companies that serve a specific audience that cares, or that reach massive scale, can survive, while the middle is squeezed
- Platform dependency on Google, YouTube, and social distributors has existed since the 2000s and is now more acute, and AI is flooding the market with cheap content
Niche, dedicated audience economics, not quality alone, determine which media ventures can survive.
- There is no for-profit business model that works in local news right now, and the industry has not yet faced up to that, unlike bigger media companies whose existence is not existential.
“I just feel like there is not a for-profit business model that works in local news right now, and we have yet to face up to that.”
- Newer niche media outlets such as Puck, Punchbowl, Semafor, and Axios are all targeting the same well-heeled Washington, New York, and Los Angeles business consumer, leaving other niches underserved.
“It's not a coincidence that Puck and Punchbowl and Semafor and Axios are all going after the same well-heeled Washington, New York, L.A. business consumer.”
- The key to a viable niche media business is whether a group of people really wants and needs what you make, which is more basic than having first-party data about them.
“I think it's more basic than first-party data, by the way. It's is there a group of people that really want and need what you're making.”
- An advertising business must either have a very targeted, specific audience or massive reach, with little viable middle ground.
“If you're going to be in an ad business, you need very targeted, very specific audience or you need massive, massive reach and there's kind of nothing in the middle.”
- Subscriptions cannot be every media company's pivot because there is a limit to the number of things people will pay for.
“There is a limit to the number of things people are going to buy. So subscriptions can't be everyone's pivot.”
- Making content only for the audience you care about is fine, but doing so intentionally puts on blinders that cause you to miss broader cultural trends, as Bill Simmons does with NBA basketball.
“But when you're consuming that stuff, when you're making that stuff, you're putting on blinders intentionally, you're just missing stuff.”
- The New York Times writes by default for its core audience of older, liberal, rich readers, whether or not editors are consciously thinking about it.
“The New York Times version of that is, oh, it's older, liberal, rich people are the New York Times core audience.”
- The New York Times nearly failed around 2008, taking a high-interest loan to stay afloat, and staff doubted the subscription model would work, so its success was not a straight line that other publishers can copy.
“They had to take a high-interest loan to stay afloat.”
- VR headsets are unlikely to take off until they reach a lightweight form factor, and 3D movies and TV failed for a similar reason: the experience was not mind-blowing enough to justify wearing glasses.
“those things have to get to that weight and form factor before they can take off. And it seems like technically, we're just a long way from that.”
- The New York Times is the only mass-market subscription news product in the world, with the Wall Street Journal second but aimed at a business audience, so the rest of the field is not competing for the same scale.
“They are the only subscription mass market news product in the world, I think, definitely in the U.S. The Wall Street Journal is second, I believe”
- Newspapers keep printing agate sports scores because a subset of paying readers wants them in print, and cutting them would cause those subscribers to drop off, which makes the decision hard even when the content exists online.
“there is an audience that is buying that paper because they want that thing in there, and you still have to supply it.”
- Avatar 2 had large box office numbers but almost no cultural impact, which suggests Avatar 3 may underperform because viewers who saw Avatar 2 may not return.
“It's f****** wild that, that many people could buy it, take it to something and have so little cultural impact.”
- Streaming businesses are currently judged by Wall Street on profitability rather than growth, a shift that is largely independent of the underlying business model.
“the issue of streaming is that Wall Street now cares about profits and not growth, and that's pretty independent of the business models”
- Media outlets program for the small audience that cares, such as cable news, and that audience capture skews their worldview, though it is not a bad thing because it reflects a successful business.
“But you can get that audience that cares about it to come back, so you program for it. And then your world view becomes skewed by audience capture”
- The podcast market is small, roughly 1 billion dollars from Spotify and a few hundred million from others, so calling it a bubble is funny because it is tiny compared to real bubbles.
“It's like $1 billion from Spotify and a few hundred million from others.”
- WWE was early in streaming and was basically running its own Netflix, but it could not make the streaming business work, because wrestling is a niche product rather than football-scale.
“they couldn't make that streaming business work. They were basically doing their own Netflix for a while because it's wrestling. It's not football.”
- There should be a media business for owners of long-haired Persian cats, since a small but widespread group that cares intensely could sustain either an ad or subscription business.
“there should be a media business that reaches people who love long-haired Persian cats.”
- Bill Simmons loves NBA basketball and his audience that cares about it listens to him, so his content reflects that group rather than a universal taste, which is how creators build businesses around passions.
“If you listen to Bill Simmons, you think everyone loves NBA basketball. They don't. He loves it and the audience that cares about NBA basketball listens to him, and that's what they make.”
- Personality-led media businesses are cyclical and renewed by the Substack and influencer world, though they were previously tried with figures like Walt Mossberg and Kara Swisher, and Puck is trying to scale one again.
“It seems pretty cyclical to me because when I was going to work with Walt Mossberg and Kara Swisher that was the model people were newly interested in.”
Digital disruption dismantles legacy bundles and models, and subscriptions let ventures bypass legacy practices.
- When Netflix lost one million subscribers in 2022 and critics urged it to act like a traditional Hollywood studio, it refused and appears to have been right, which undercuts advice to copy legacy practices such as staggered releases.
“When Netflix lost 1 million subscribers in 2022 and everyone was predicting their doom... Netflix has said, screw it, we're not going to do that, and they seem to be right.”
- The subscription model creates a virtuous cycle because the creator and the paying audience agree that the audience likes the work and the creator will make more of it, though it carries downsides.
“you and the person who's giving you money, have made an agreement that they like what you're making, and you're going to make more of it”
- A podcast can cost between zero and five dollars to make, and some will never make significant money while others will, so media organizations overestimate the cost of trying the format.
“the show you guys are making cost somewhere between $0 and $5 for you to make. There'll be audiences. Some of these things will never make money”
- Prior to the internet, local newspapers were funded by bundled business models and a monopoly on distribution, which made people pay implicitly; the internet broke that bundle, so local content is now available free or cheaply elsewhere.
“prior to the internet that was covered up because you had business models that bundled all that stuff together and you had a monopoly on distribution, the internet blows all that up.”
- Recorded music companies resisted streaming for a long time, and TV companies are still trying to preserve the cable bundle even though its decline is clear.
“recorded music selling, LPs and disks and really resisting streaming for a long time. I mean watching the TV folks go through right now, still trying to keep the bundle intact”
- Although copyediting jobs have disappeared, tools like Grammarly and spell check are desirable, so one should not cap automation even though it eliminates work and sometimes creates more errors.
“But you wouldn't want to say, 'Well, those jobs must be capped, and we can't have Grammarly, and Microsoft Word spell check.' We want those things in our life.”
- Podcasting is still early, with audiences expected to keep growing, but the podcast economics have retreated, and the podcasting bubble is tiny compared with other bubbles, around 1 billion dollars from Spotify and a few hundred million from others.
“It's like $1 billion from Spotify and a few hundred million from others.”
Media businesses remain dependent on distribution platforms and must accept codependency with them.
- Media businesses have been dependent on distribution platforms since 2005, when one business-side staffer's job was getting Forbes stories onto Yahoo's homepage, and the cycle keeps repeating with Facebook and others.
“the most important person in our newsroom was not an editorial, she was part of the business side, and she was in charge of getting yahoo.com to put our stories on the homepage.”
- Every media operator now says it must own its platform, a lesson that is reasonable but arrived late, and Facebook's abandoned partnerships show why publishers keep getting burned.
“We were going to work with Facebook and Facebook literally paid us to make live video and they pivoted out of that”
- Most readers will not come to a publisher's homepage directly; they find content through side doors and distributors, so media companies will have some codependency with platforms no matter what, as cable networks did with carriers.
“most people aren't going to come to your home page. They're going to find your content through a side door. They're still going to get it through a distributor.”
YouTube is under-covered and opaque, a distinct product whose long-tail audience has outgrown television.
- YouTube is hard for analysts to understand because its personalized, algorithmic nature makes it hard to see what people consume, and YouTube may benefit from that opacity.
“it's just harder to get your head around if you come to YouTube because it's so personalized, because it's so algorithmic.”
- Kids watch both YouTube and TV but expect different things from each, so YouTube is a different product and is unlikely to host movies or to act as a graduation path into television.
“You don't see people graduating from YouTube to television. There's an attempt to do that every couple of years, it doesn't work.”
- YouTube is dramatically under-covered by media peers, with roughly 9 billion dollars in top-line revenue last quarter and a cable-like TV business that makes it effectively the fourth largest U.S. cable TV provider.
“I think they did $9 billion last quarter, top line. They've created a cable TV business just kind of on the fly... They're now the fourth largest cable TV provider essentially in the U.S.”
- YouTube and television are different worlds, and attempts to move YouTube creators up to television have repeatedly failed, while YouTube's own attempt to host movies was abandoned.
“You don't see people graduating from YouTube to television. There's an attempt to do that every couple of years, it doesn't work.”
- YouTube quickly created what is now roughly the fourth largest cable TV provider in the U.S., and will likely become third, though its purpose is unclear to Kafka.
“They've created a cable TV business just kind of on the fly. ... They're now the fourth largest cable TV provider essentially in the U.S., they'll probably be #3 pretty soon.”
- YouTube's audience is massive and spread across a long tail, with Mr. Beast being a tiny fraction of it, and it is a major staple of young people's media diet that the industry has not figured out.
“Mr. Beast is a huge star, but he's a tiny fraction of YouTube's audience.”
- YouTube is under-covered partly because it sits inside Google and partly because its personalized, algorithmic nature makes it hard to see what people consume, while TikTok and Instagram have gotten more coverage recently.
“I think it's harder to understand what people are consuming.”
- MrBeast's planned Amazon show is a skeptical move because it is not clear his YouTube audience will follow him there.
“I don't know that your YouTube audience is going to come and find you on Amazon, but we'll see.”
Celebrity curator brands are fading under metrics, since executives are number crunchers and reputations are often unsupported.
- Success is now much easier to measure, and numbers are front and center, so executives like Ted Sarandos cannot defend shows that are not being watched by pointing to quality, unlike the older era of celebrity editors.
“success is much easier to measure in a lot of ways than it used to be... now the numbers are front and center.”
- The model of building a media business around a single individual is cyclical rather than secular, and such businesses are naturally capped at a certain size even though they can be highly profitable.
“It seems pretty cyclical to me... I think you're naturally capped at a certain size.”
- HBO's reputation for great content may be specific to the company, since comparable curators exist at other studios but are not famous, and the reputation tends to attach to whoever is publicly visible, such as Richard Plepler before Casey Bloys.
“Even when Casey Bloys was the guy doing the work, Richard Plepler was the guy who got the reputation for being artist-friendly”
- Music label executives are not musicians and do not discover talent; their jobs are mostly number-crunching, with people below them doing the talent scouting, which weakens the case for celebrated content curators.
“none of the people who are running the big labels are musicians... their job is to crunch the numbers.”
- The era of magazine celebrity editors such as Anna Wintour and David Remnick is largely over, with only a few remnants left, though current editors like David Haskell at New York Magazine still build products with national cultural reach.
“The idea of magazine celebrity editors that was a big thing. That era has gone, Anna Wintour and David Remnick.”
- WBD chief David Zaslav presented himself as a lover of content and Hollywood, but Hollywood did not buy it, and that gap is part of why HBO's curation story is complicated.
“David Zaslav has made the mistake of trying to present himself as like, 'I love content. I love Hollywood.' And everyone in Hollywood said, 'No. We don't like you.'”
- When Peter interviewed Vince McMahon about wrestling economics for Forbes, McMahon had publicly acknowledged steroid use and was unready for questions about how money flowed to talent.
“he obviously was not ready for me to ask him about how it worked and how the money was distributed to the talent.”
- Strong taste-driven curators like HBO's Casey Bloys are not common across the industry, and the music business shows that top executives are usually number crunchers while artist-friendly reputations come from people below them.
“The music business, none of the people who are running the big labels are musicians or if they are musicians, it's not relevant to what they do.”
- Interviewing Vince McMahon about wrestling economics for Forbes, Kafka found that he was unprepared for questions about money distribution to talent, and that executives who have not been coached are rare and most valuable.
“It's rare that you find someone, especially running a company who hasn't been coached or prepped in some way.”
- Bill Simmons acts as if he never paid attention to metrics, but Kafka believes he does, because successful media operators must monitor metrics even when they say otherwise.
“I asked also Bill Simmons about paying attention to metrics and he acts as if it's something he never did, I'm like, of course, you do. But if you weren't, you have to.”
- Netflix's defense that its shows are high quality conflicts with audiences that do not watch them, as critics say the taste is bad while viewers say they like what Netflix makes.
“Their taste isn't good. The audience likes what they're making is the retort to them.”
- If Netflix thought spreading out show releases helped it, it would already be doing so, because the company changes stated positions when data supports it.
“If they thought that spreading out the release was helpful for them, they'd be doing it.”
- A content curator's reputation at HBO, such as Casey Bloys, may be attributed to the executive when the real creative credit lies with the company's creative leaders, and Bloys reports to David Zaslav, who has alienated Hollywood by pretending to love content.
“Casey Bloys has to report up to David Zaslav.”
AI is mostly harmful to media, though it also boosts productivity and displaces skilled workers.
- AI is already having a mostly bad effect on media by producing low-cost, mediocre content at almost zero cost, creating a tragedy of the commons in which producers compete with free near-equivalents.
“It's made low-cost content for 0 cost and still bad, and it's a tragedy of the commons.”
- AI is double-edged: it boosts productivity and enables tools like Runway ML's effects work, yet it puts specialized manual effects workers out of work, similar to how proofreading jobs disappeared while spell-check remained valuable.
“But even that positive version of the story means there's people who do that special effects work manually... they're now going to be put out of work by the robots.”
Podcasts enable deeper, personal conversation, and their video and format conventions are still early.
- Putting podcasts on video is expected by platforms, but it is unclear whether video on Instagram and TikTok actually generates audience, and it adds real production cost.
“it's totally unclear whether -- when it comes to like Instagram and TikTok, whether that is actually generating audience.”
- Podcasts allow deeper conversation than live stage interviews because removing the audience lets the interviewer push a question again and tolerate awkward pauses, whereas on stage guests can filibuster.
“onstage it's much harder to ask someone a difficult question because they can just filibuster an answer.”
- Podcasting is still early, and new media mediums create new formats over time, just as early television programming was simply radio plays before it developed its own language.
“When television showed up, they didn't have words to describe it... so the original television programming was just radio plays. New mediums create new formats.”
- Podcast audiences appreciate the personal connection with hosts, which is why people pay for podcasts they already get for free.
“Audiences really appreciate the personal connection. It means something to them.”
Local news is a public good, and business-side forces shape what reaches audiences in journalism.
- Peter chose to focus on video around 2008 because tech and business were converging there, and audiences would care about why video appeared on their screen even if they did not care about particular companies.
“I consciously decided around 2008 or so that I wanted to focus on video as a beat in part because there was business because there was tech”
- Local news should be treated differently from general media companies because knowing what is happening in a community is a public good, not merely a business.
“being able to know what's happening in community is a public good.”
- Since 2005 at Forbes, a business-side employee was the most important person in the newsroom because getting stories onto the Yahoo, MSN and AOL homepages determined whether a story moved at all.
“the most important person in our newsroom was not an editorial, she was part of the business side”
Good journalism needs synthesized perspectives and unscripted guests, not polished, coached executives.
- The best-case candid guest tells only their version of the story, so responsible journalism needs to synthesize multiple perspectives, and a heavy accent or reluctance to appear limits the narrower slice of people a journalist can reach.
“It's still only their version of the story and one way of doing a responsible journalism is to synthesize a bunch of different perspectives”
- Most business executives have been coached or prepped, which makes them less interesting, and the tension of interview work is wanting guests who tell good stories without being overly scripted; book tours add money incentives that change the conversation.
“It's rare that you find someone, especially running a company who hasn't been coached or prepped in some way.”
Strategic reversals and business fads show that media conventional wisdom keeps changing.
- Conventional wisdom about Netflix flipped from subscriber growth at any cost to having many subscribers and being profitable, illustrating that media mantras change every few years.
“The Netflix was winning a couple of years ago was subscriber growth to the moon. And now the Netflix winning today... is we have a lot of subs, but they're also profitable.”
- Netflix is a rare example of a company abandoning a successful business for a new one, and it could do so partly because it was formed in the late 1990s and was not a legacy business entrenched for decades.
“One of the very few examples of a company saying, this successful business we have now, we're going to move away from it”
- Netflix is fully capable of screwing up its strategy, and its video game push could go to zero, with Wall Street not giving it credit for the risk.
“They're fully capable of screwing it up.”
- Netflix has repeatedly reversed public statements it had made, such as saying it would never do ads or sports before moving into them, which shows the capacity to change strategy openly.
“Netflix, again, they said they weren't going to do ads, they're doing ads. They said they weren't going to do sports, they're doing sports.”
- Business fads like Web3 spread through herd mentality: people were earnestly discussing it a couple of years ago and then never mention it again, though it was real.
“there's also just a herd mentality and people do one thing against the other and a year ago, 2 years ago, I was having earnest conversations about Web3”