The Netflix Culture Code That Changed Entertainment Forever | Reed Hastings Interview

To me, the most interesting thing about studying Netflix and talking to Reed is that it is, as a business, probably the single most relatable example—since we all watch Netflix—of 2 really simple ideas that everyone talks about but are very hard to do in practice.

The first is this notion of finding a simple idea and taking it extraordinarily seriously. Netflix has effectively been scaling up its core original model since its inception. Reed talks in our conversation about how even the DVDs were nothing but a stepping stone toward the streaming future that they envisioned at the very outset of the company's founding in 1997, and how simply letting that idea play out over decades without getting distracted can be so powerful.

The second is this notion of talent density. This is a term that now gets thrown around at every major company, and really, it was Reed and Netflix that pioneered this concept of what can happen if you set and keep a talent bar exceptionally high. We get into why that's difficult, what Netflix did to make that talent-density bar work, and how it sustained itself over decades.

This conversation really is an ode to those 2 simple concepts. And, of course, in this case, it's fun to learn about because it's something that we all watch every day.

I want to go back to your first business and the origin story of this notion of talent density that you've become very famous for. We'll talk about talent density for sure. It's one of these ideas that's now ubiquitous in most technology companies. I think you were the originator of the concept, but I want to hear how you came to learn that lesson in the first place.

Presuming that your very first team wasn't just incredibly talent-dense and perfect, what was the early origin story of that concept?

1. Talent Density

I founded Pure Software in 1990. We grew like a typical great software company, doubling. I wasn't careful about it, and I would say talent density declined. Later, when I analyzed that company, we went public in 1995, got acquired in 1997, and when I looked back at what happened, one of the major things was declining talent density.

With declining talent density, you need a bunch of rules to protect against mistakes, and that only further drives out the high-caliber people. So it was through that experience that I realized, okay, I've tried to run software like a manufacturing plant—reducing error and putting in processes—and that doesn't get high productivity or high talent.

We should manage software much more artistically, with inspiration rather than management. Typically, we humans value being nice, and we value loyalty. Yet in the workplace, that's a tension, because being nice is in contrast with being honest.

I generally like people who are nice, and yet I want people in the workplace to be honest with each other so that we're more productive. So we have to find a way to give each other permission to not be conventionally nice, and instead to be focused on the team's success, which means being very direct.

Similarly, with loyalty, we come to see loyalty as something in your family. You would never fire your brother if you were tight on money; you would share. That's what we admire, and yet in a company, what we do is lay people off.

This whole idea that a company is a family is unintentional, but it just derives from all the structures of society. All companies used to be family companies, and then corporations have grown more recently. All countries used to be family countries and kingdoms. Basically, family was the deep organizing unit, so it's natural that it spills into how we think about an organization.

The contrast is a professional sports team, and that's an admired model. It's really focused on achievement, and everyone understands that you change players as you need to try to win the championship.

It's changing the language that you use and don't use. Don't say things like, “We're a family,” or, “I treat you like my family.” That's a little bit true, but not true enough. Instead, we're a professional sports team, and we all have to fight every year to keep our position, because if we can upgrade, we must, to achieve the winning of the championship, which is producing a great company.

How do you protect against the natural way that companies seem to bleed down toward lower talent density over time? There seem to be very few organizations that get it high and then keep it at that same level, especially with scale. What are the ways that you learned to keep talent density as high as possible as the company grew so big?

As companies grow, you may be able to pay people more, so that will help. If you think of the sports teams in the biggest markets, they can afford the highest compensation, and the Yankees or the Los Angeles Dodgers often have the best players.

2. Professional Sports Team Model

It's not one-to-one, how much you spend and quality, but there is a strong correlation. I think the second thing you can do is continue to really evangelize the benefits of talent density over total quantity, so that more and more of your leaders get adept at managing for density.

I would love to talk about each stage of the funnel to creating talent density in a business, starting with how you found people in the first place, what the most reliable ways were of finding people, and then how you evaluated them. I want to talk about further down the funnel, but starting just with the top of funnel, what were the most effective ways of finding people who had the potential to be extremely talented inside one of your businesses?

I've come to look at it more like keeping a pretty broad funnel and hiring a lot of people. Over the first year, you really get to know them, and you can figure out whether you want to keep them or not.

Other people have a view of keeping it very hard to get in, but then you can stay no matter what. I think that's been more of Google's orientation, as an example, and it comes from their graduate-school background. It's really hard to get into Stanford Graduate School, and then it's hard to get pushed out, too. It's natural that they mapped themselves onto that model, and there are some benefits to that, but that's a different model.

Mine is more about having relatively open doors. We'll interview broadly and try to select what we think is the best person.

It stands to reason that maybe your 1-year attrition rate was higher than Google's or somebody else's.

Oh, quite a bit.

Do you remember what it was?

20% in the first year?

Given that that's pretty high, what would you tell people on the way in, or tell the organization about that rate itself, to make sure it didn't spook people that lots of people would leave afterward?

Well, it did spook people. I would say it's only fair to let them know what they're getting into.

Yeah.

We would say, “We're not going to guarantee you a lot, but we'll guarantee that we'll always surround you with great people and have you work on hard problems.” That was our core. You may not be happy, the hours may be long, and the food may be okay, but the essence of what we can do at work is hard problems with great people.

If your primary orientation is around job security and you're willing to put up with working with uneven levels of talent, then there are other companies that are a better fit.

There are some benefits of that, which is that you have stability in your life. If you're more of a performance junkie, and the thing that makes you vibe the most is working around incredibly talented people and running fast and loose with great teammates, then you're willing to put up with the job insecurity. Nobody likes it, but you're willing to put up with it to get the performance density.

You said “fast and loose.” Can you say more about “loose”? If you overmanage, for example, a tight process or specific hours that you have to be in the office, or a wide variety of things, you filter out performance and creativity. If you overmanage those things, what happens?

The looser that you can run, the more creative the organization will be. We talk about it as managing on the edge of chaos.

3. Managing on the Edge of Chaos

You don't actually want to fall into chaos. In chaos, the product barely gets released, it's full of bugs, people are upset, payroll isn't made, and lots of bad things happen. But it's getting us close to that edge of chaos, where there are last-minute saves and as much dynamism as you can possibly tolerate, as opposed to, say, a semiconductor factory, which is trying to reduce variation and reduce error to get rid of variance.

If you're going to be a creative organization, you want to be high variance, high creativity, and, again, managing on the edge of chaos.

I'm curious, with the 20% attrition rate, what you learned about letting people go well and the right way. How did you get really good at that specific part of the life cycle?

I think there are 2 parts to it. One is to release the moral thing. Most managers are people managers. They like people, and they don't want to hurt people, so it's very difficult for them.

One of the best things is to do large severance packages, like 4 to 9 months of salary. It feels expensive at first, but one, it makes the person who's let go feel a little bit better because they've got a bunch of money in their pocket. Two, it helps the manager do their job, because then they don't feel as bad letting the person go.

4. The Keeper's Test

It sets up a much better mutual feeling. The third thing with terminations is setting a context where it's not a moral issue.

You didn't fail. It's just like a professional sports player: We think we can get someone better here. Okay, so it's a pity for the person, but it's seen as natural as opposed to a failure.

Typically, I would say something like, “Hey, Patrick, I see you're working really hard. You're trying. I'm so sorry to tell you that, honestly, if you quit, I wouldn't try to change your mind to stay.” The reason I wouldn't change your mind to stay is that I think I could get someone in your role who could do what you're doing, plus even more. And here's why.

The way the company is set up is that if I wouldn't work to keep you, I'm supposed to let you go. In that way, we're executing on an agreed-upon framework—the whole Keeper Test framework.

How did the Keeper Test literally work? How was it rolled out across the company?

It was always there. In the original slide deck, it said, “Adequate performance gets a generous severance package.” So it's really just starting up front. The test that we encourage people to use is: If someone were quitting, would you try to get them to stay—to keep them? That turns out to be a good test relative to all the relief we sometimes feel when someone not great moves on.

Was there an episode in Netflix's history that you can remember where you were on the edge of chaos and it really either did, or very nearly did, cost you very dearly during Netflix's 25 years?

There were a couple of small things that we did wrong, and one big one: the Quickster separation of DVD and streaming.

Maybe taking the Qwikster example, what is it like to see high talent density operate against something like that?

Quickster, for your listeners, was a sad episode in 2011, when I became convinced we really had to go all in on streaming, drop DVD, and put DVD in its own company that would drift along and free us from that. Unfortunately, most of the customers were mostly using DVDs, so they were still mailing me the discs. They didn't like it. There were lots of cancellations, and the stock dropped by 75%, so it was a tough time.

5. Qwikster

Ultimately, it's the right thing to have separated DVD and streaming, but we did it too fast. The big analysis of it afterward was that lots of the executives thought it was very problematic. But they kind of said to themselves, “Geez, Reed's made 18 decisions right before this, so I'm probably wrong and Reed's probably right.” They suppressed their own significant doubts.

What we realized was that if they all knew of each other's doubts, they would have been much more likely to weigh in and probably just have us do it more slowly. We instituted a much more collective information process on decisions going forward, where everybody weighed in from 10 to negative 10 on decisions, and it's all in a big shared document so everyone sees what everyone else thinks.

That way, if we had had that decision process in place, I think I may well have thought, “These are all fantastic people, and they're all horrified at this idea.” I may be right, but let's at least go a little bit more gently to figure that out, and we wouldn't have had as deep a hole.

If you think about all the value creation that you've been a part of, or the leader responsible for, was most of that the result of a fairly nonconsensus idea? That seems like a consensus process—or at least, if not decision by consensus, being aware of what the consensus is. I'm curious about that tension there. It seems like very often nonconsensus is where the value comes from. Is that generally true in your personal history of the decisions that you made that created most of the value?

I think you want to be super careful here, because this is the source of much value. You want to be totally independent in your thinking and not consensus-oriented at all, but you want to know what other people are thinking. Otherwise, you're flying blind.

I think there's a high value on gathering information and opinions, but then not averaging them. We would never do that. We were very clear that the concept was the informed captain. We wanted to make it like the captain of a ship: The captain of the ship makes the decisions, but it's good for them to collect a lot of information.

We were very strong on no committees. Individuals make decisions, but we want them to be informed about that decision. Then it's up to them to make it.

I'm so interested in the bucket of “seems like a bad idea but turns out to be a good idea,” because there's just less competition if it sort of seems bad. What has been your process of coming up with good ideas in the first place?

I fall in love with ideas easily. I'll see some combination or insight. The original one was that DVD, which was just coming out when Netflix started, was very lightweight. This was coming out of AOL mailing CDs to everyone to install AOL on CD-ROM, so I was pretty familiar with mailing because I'd gotten tons of these through the mail.

DVD for movies was just replacing VHS, or just starting to. I kind of clicked on that. The classic computer-networking thought experiment you do is, “What's the bandwidth of a FedExed tape through the mail?” You calculate it, and it turns out to be terabits per second at low cost to send a backup tape by FedEx.

6. Contrarian Thinking & Idea Generation

You start thinking about networks a little bit differently. All those combinations made me think of DVD by mail as an extremely efficient digital distribution network that someday the internet would be faster than, cheaper than, and lower-latency than. I never thought, “I love the mail business.” I thought, “I love the network business to deliver media.”

The contrarian part of it was that when we were fundraising in 1997, everyone was excited by internet delivery. I was like, “But it's not even close.” It didn't matter—they were excited about it. It was very much a contrarian thesis that we could build a business with DVD and then transition it to streaming. It's precisely because of that contrarian thesis that we didn't have much competition in it. Because it worked, we created great value.

When did streaming first enter your mind as, “Clearly, this is the place that we're ultimately going to have to go”?

Oh, that was from the beginning.

From the very beginning?

That's why we named the company Netflix: internet movies.

Yeah. It was really just about managing the transition, even from day one. Designing the efficient system for DVDs was just a notch on the timeline getting to streaming.

Correct. It was one digital distribution network, and then eventually we would replace it with another. We knew that would be a challenge, but we knew the best way to be successful at it was to get big on DVD. For the first decade, that's all we worked on.

One of the other really cool things about your background is that for a long time you were on the boards of, I think, Facebook and Microsoft. I don't know if you're still on those 2 boards or not.

No, I'm not.

But today I think you're on the Anthropic board and the Bloomberg board. You've had Netflix itself at the center of technology, and you've had this very cool, 360-degree view of probably the most interesting era of technology development ever.

I'm curious, from those seats, what the technology landscape looks like to you today. What are the key considerations—the things that you have your attention on—that seem the most important to you from those vantage points?

First of all, because of exponential phenomena, it's always the coolest time ever to be in computer science. In the 1980s, I thought, “Oh my God, so much better than the 1960s.” I just think that'll always be true.

As a CEO at Netflix, I learned so much from being on the boards of Microsoft and Facebook. They had quite different businesses, but they made very interesting trade-offs in the way they thought about things. Both of them were very long-term-oriented in that they were willing to lose money in certain new areas for a decade.

What I loved about looking at Facebook's business was that it was ad-supported, and everything they did that was on the core, like Instagram, worked incredibly well. When they tried to do crypto or other things that were not big ad-supported businesses, it didn't work well. That's an example of how companies get good at something, and then if you can add to the core mechanism, that's great.

7. Lessons from Boards

We've always wanted to add content to the Netflix subscription to make it more and more useful, more and more enjoyable, but keep it like one big model, as opposed to also doing theatrical movies or something else as a way to expand revenue. To answer your question, I would say we're trying to find simple, large models that, if they work, you can continue to expand and expand on the core monetization engine that you've already got.

If you look at Microsoft's case, it's building high-scale software. I'm on the board of Bloomberg, which is owned by Mike Bloomberg. It's trading stations for Wall Street and media around that. He's been incredible at this long-term orientation, having this intimate relationship with the customers, becoming a trusted utility for the industry. That's been very powerful, and there are big moats for that business—really customer loyalty that he's been serving in multiple dimensions for a long time.

Anthropic I've only been on the board of for a year, and it's a wild story because it's growing so fast.

What have you learned from Mark? You mentioned what you learned from Facebook, but what did you learn from him specifically?

He's super committed. When you look at the metaverse, he's convinced that there's going to be something beyond the phones. Maybe that'll be a glasses format, and he doesn't want to be dependent on it; he wants to really be the inventor of that layer, which is extraordinarily ambitious. I probably would have just been the ad giant if I was doing that business and tried to go after TikTok. But he wants to do bigger and broader things for society. It's great because he does amazing amounts of innovation funded with what would otherwise be the profits of the company.

You've been on these great boards. You had a board yourself, of course. What advice would you give to people to either be a great board member or run a great board process themselves?

Typically, board members want to add value. The problem is, by the conflict rules, they don't really know the business. If you run an airline, you can't be on another airline's board, but you're doing that board one day a quarter, for the most part. And on one day a quarter, it is super hard to add value.

And so what you see is a lot of directors who struggle to add value, and then management has to be super polite to them. Management can't tell them, “You don't know what you're talking about,” because they run the thing. And so you see this dysfunctional thing where board members ask hard questions and management ducks and weaves, and it's not very functional.

I would say the first part is for board members to realize, “Okay, I'm not here to add value.” They can hire consultants who know the industry and are not conflicted, and they pay for the advice. So I shouldn't spend my time trying to give advice. So then what am I doing? I'm here as a board member as an insurance layer. If the company falls apart, I will step in and be part of replacing the CEO. And that's basically the entire job, which is replacing the CEO.

Well, okay. And to do that and to have the confidence to do that, you have to learn the business. So you can't be asleep. You've got to really ask a lot of questions and learn what drives the profit streams, how the business works, and what the issues are with it. But again, you're not trying to solve those problems. You're trying to get a grasp of the business so that you can determine who might be the best person to run the firm.

And if you get that right, as, say, Microsoft shareholders or the board did with likely Satya Nadella, then the business takes off, and all the advice in the world doesn't matter compared to that. If you're on a board, don't measure yourself by, “Did you give a suggestion?” Measure yourself by, “Did you get more and more prepared for the small chance that you will have to take big action?” And so it's a lot like a firefighter who drills and drills and drills and hopes that there's never a fire.

Yeah.

Okay.

When selecting people that would be that insurance layer for your own business, what did you select for? Because a lot of these boards are full of very fancy people like you that are great names to have on a website as a board of directors, and that seems to be a selection criterion versus, like, this person is actually going to be good at this insurance-layer thing. How did you select board members?

Yeah, people who I believe will be wise in a crisis. We talk through the board model. We call it extreme duty of care. Duty of care is one of the responsibilities of a director, and we amp it up so that they really have to know what's going on.

We ask directors to come to management meetings so they can watch what's going on, watch the sausage being made. Again, not so they can add value, but so they're highly informed. And so we look for people who are wise in a crisis.

A board interview process would involve those kinds of things: “Tell me about different business crises that have happened,” and whether, in case that happens, they would be wise.

How much of your time, when you were running the business full-time, was systems structuring and thinking around the business versus the marginal strategic initiative or something?

I never booked hours on my calendar to think about the culture. You end up just trying to make things better and then watching what's going well and what's not, making observations.

8. The Open Compensation Experiment

Here's an example. From maybe 2004 on, we had open compensation. Basically, the top 100 or 500 people in the company could see all the compensation throughout the company. The rationale was that they could keep similar people in a similar vein, and there would be more trust around gender and around other dimensions that could be discriminatory because the data was all out for everyone to see.

That was all true, but it also created a lot of petty rivalries. “I make a huge amount of money. This other person makes a huge amount plus $10,000 more.” And so it got pretty distracting.

Ultimately, we put the question to the VPs about 10 years later, in 2016, and they decided to take it away from themselves and from everybody else and do the traditional approach: you know your direct reports and their teams, but not the whole company. I would say that was an experiment in human nature, which got resolved pretty decisively to be less mavericky, but it ended up working a little better.

Again, we would take an experimental view on things, and that's a good example because then you can see that we're not geniuses. We're just willing to question things and try them. So we did open compensation for a number of years and then decided that its net costs were negative.

Another strategic question that always fascinated me about Netflix was how you determined how much to spend on originals and original content.

As much as we possibly could.

Say more about just the core calculus or thinking there.

Well, there's how much to spend on any one show—that's a different question—but in terms of the total budget, we would always try to shovel money into that in the hopes of creating the great next K-pop Demon Hunters.

9. The Strategy Behind Original Content

In terms of any one show, the question is, what's the likelihood, based on what we've seen, that this is going to be big? And it's also a competitive market. The very first original series that we had that helped make our reputation was House of Cards, and we had to bid that away from HBO.

As Media Rights Capital was making it, they had bids both from HBO and us, and we were not—we were a DVD company. So we had to overpay relative to HBO, and then they went with us. We had to overpay by a bunch because it's a lot of risk.

Yeah.

And then they came through and made a fantastic show. Then we were off to the races in original content. It's a simple way to think about it, almost like one would think about a venture capital portfolio or something: you want to make lots of bets, and you don't know exactly which one's going to be K-pop Demon Hunters, but having a KPop Demon Hunters is the thing that matters—that you have some dominant, massive franchise.

Very much so. But it's similar to venture capital if every A round were $100 million and there was just an A round. So it tends to be pretty much a single round to—

Fund the construction.

You do get sequels and other things you have option rights to.

Yeah.

But that would be the big difference from venture. If you think about the portfolio of content, what else would surprise people about the conversations happening inside the business, especially in the early days of developing that portfolio? What are the key considerations that matter to you as you expanded it so that it's a combination?

I mean, now it's so many things, but in the early days, you're obviously making choices. It's House of Cards; it's not something else, and there are trade-offs. What would surprise people about the conversations that led to the portfolio that you ultimately chose?

Everything for us was around reinforcing the brand and trying to figure out what the brand should be. The cable networks, by necessity, were narrow brands because they got one cable slot.

The difference was that FX and Hallmark were both interesting, doing different types of content, but the handle on the brand gave you the type of content, which was inherently pretty niche because it had one network slot. We were doing something that had all the network slots.

And so we spent a lot of time thinking about how much of the programming we wanted to be Hallmark—soft, easy, romantic stories, feel-good—versus FX, sort of cutting-edge and violent and dark, versus Comedy Central.

Okay. So, our main issue relative to the industry was that we had this incredible breadth of content to choose from. And on any new film or series, unless it's completely derivative, there are just so many variables compared to other things. So, it ends up that you can do asset allocation, which is how much in comedy, how much in drama.

Okay.

But in terms of the stock picking, it ended up being intuition and people's judgment. And then we promoted those people with great judgment who got this right again and again. We called it great taste, but they had more than taste. They had taste and judgment about whether the people would deliver, whether it would come together, and all kinds of other things.

So it became just people picking. Then it was trying to figure out how much money to put in each area, and the people in those areas would figure out how to best spend it.

The other side of the equation, of course, is the beauty of the business model: fixed cost for a piece of content, and then a growing subscriber base across which to spread those costs. But that requires that you grow the subscriber base. How did those two interrelate? What did you learn about what sorts of fixed spend on content would create great, reliable, high subscriber growth?

What I loved about Microsoft and Facebook's business is that they, at that point, basically had 1 big product, or maybe 2 highly related ones, and then it was about growing those products to $50 billion in revenue on a product. So when I started Netflix, I thought, thankfully, we can do this as 1 really big product, because entertainment is an extremely large market. Basically, every human on the planet watches television to varying degrees, but it's a deeply human thing to watch stories.

And so then the question is, okay, what percentage of that could we capture? Even today, Netflix is only about 10% of U.S. television. We've got a long way to go, and internationally it's less than that. Generally, in terms of how we think about subscriber growth, we knew that if we could produce better television, make it lower-cost and more enjoyable by being on demand, there would be a huge market for it.

So it was kind of constrained on essentially product quality. What kind of shows do we have? Now streaming is kind of flawless and not differentiated between competitors. But for a decade, we did it much better than our peers.

Is that other 90% defined as just traditional television, or does that include, like, YouTube watched on television?

No, YouTube is about 12%.

So it includes everything—sports, video gaming, the uses of the television screen?

I mean, we compete for time on mobile phones too, but we're very small there. It's not a big use case. On television, we're a big use case, but still, again, it's under 10%.

If you think about that percentage as an important thing for Netflix as a business, what are the competitive frontiers or fields on which you feel like you're competing against something like YouTube? It's easier to imagine competing against cable or network shows, but versus something like YouTube, which is a pure UGC platform, do you think about it that way? Are we competing against them, and therefore we want to do certain things to win?

Well, they're growing and we're growing, and traditional linear is shrinking. So you're right that mostly we both compete with linear TV.

But we do worry about YouTube because it's sort of a substitution threat. Does it get better and better with AI creators, and does it just become more and more of people's time? That's the user-generated world, and it's not really user-generated; it's on spec. That is, there are some very professional people who make content for YouTube, but they don't get paid for it in advance. Then they put it up and see what kind of ad revenues they get.

In our case, we pre-fund the programs, which gives them a bigger budget. They don't have to do it on spec, and that's really the biggest difference in the business model. But ultimately, it's about whether we produce content like The Perfect Neighbors, a documentary that just came out, won all these awards, and has been the number-one documentary this last month—clever, fresh-perspective content like that—or K-pop Demon Hunters, which was our hit this summer. So it's the ability to create those hits.

What is that magic like? What is shared among the people like Ted and others who have been able to reliably and consistently create, or be a part of creating, those big hits over time?

If only it were reliable and consistent. [Laughter.]

I think K-pop Demon Hunters was probably our 30th animated film.

Fascinating.

Okay, so it's not at all reliable and consistent. Again, it is a lot more like art: seeing the contrarian edge and figuring out what the story is. Imagine the pitch for K-pop Demon Hunters. It doesn't fit a set of formulas.

In that way, it's a lot like venture, in that a few of the companies will generate outsized returns.

What do you think will be the most interesting impacts of AI on the Netflix business specifically? This could mean the cost to create the content, it could mean the service, or it could mean anything else. Where does your mind go as you think about the raw capabilities of the technology?

Visual effects is 1 area where a lot of that workflow can be automated. But in terms of recognizing something like K-pop Demon Hunters at the script stage or pitch stage—which is the biggest value creator, deciding which things we back—that will be a far-distant skill.

10. AI's Impact on Storytelling

Eventually, AI might eat up everything and be better than humans at everything. But in terms of the sequencing, think of when AI is not particularly incentivized, and the companies are not incentivized, to do long-form character development. At some point, they may do that and focus on it, and then the AIs will be winning the Booker Prize and doing the best fiction in the world.

Remember, we're only interested in the top 0.001% of the stories that get written. Simply writing a story—I mean, there are a million film students; we could just go to them. So the issue is trying to find 1 that's really unusual and extraordinary, and recognizing that 1 early. I think AI will have had a lot of other effects before it hits us on that.

Can you imagine kinds of innovation in the form factors or formats of shows? It seems like we've got a couple: there's the show, the documentary, and the full-length feature movie. Can you imagine lots of different kinds of form factors starting to proliferate?

Well, let's step back a second and think about contrarian thinking generally. You love contrarian thinking, right? But you probably need to remember that contrarian thinking, most of the time, is wrong.

Right.

Once in a while it's right, and that's when you get the big reward. But you have to say that most of the time contrarian thinking is wrong, and conventional thinking is right.

For example, on formats, people have been trying to think about multi-ending, design-your-own-story, and short-form. There was Quibi; there are all kinds of things. And the enduring aspect of a film at 1.5 to 3 hours as a story has stayed strong, like the enduring form of a novel, a short story, or a TV series.

These things are tapping into something human. Video gaming is a different modality, and that's quite a bit different, but most of the hybrids between TV series that you interact with have been very small markets. It doesn't mean we won't eventually come up with a new art form that's quite different, but I don't think it's as easy as choose-your-own-adventure.

The particular thing is that we're in lean-back mode with television, and we're mostly wanting it to tell us a story. If you think of young kids—2-year-olds—half of the time they're like, “Daddy, read me a story,” and half of the time it's, “Daddy, play with me.” These are 2 different modalities. One is passive—I mean, again, I think it's very biological, and we're selected for it—and the other is very active. One of those becomes television, and the other becomes video games.

I'm also fascinated by the technology backbone and story behind Netflix—the sort of invisible part of the business that everyone just takes for granted. They can hit a button and have this beautiful thing pop up, but I know quite a lot of building happened behind the scenes. Can you tell that part of the Netflix story, what it took, infrastructure-wise and technology-wise, to make what we all enjoy possible?

Well, there's always been a sort of medium barrier to entry. I would say, first, with DVDs, we had incredible sorting and shipping machines, postal integration, and all kinds of other things. I used to spend all this time thinking about types of polycarbonate plastics that break and don't break, and we were working with pressing plants.

The biggest issue we had was making sure that the DVD would get to you without cracking or being damaged in shipping, and that it was on time.

The postal carriers didn't steal it. So there was a huge amount of machinery to ship 1 million red envelopes a day consistently, kind of FedEx-style. Then, certainly, streaming the mechanics of getting the bits to people was challenging.

We first launched in 2007, and for probably 15 years, the internet was underpowered. You had to do a lot of clever engineering things, but for the most part, there are now 100 companies that stream, and consumers can't particularly tell a difference between them. I would say that's now just become part of the base systems and has been commoditized.

What's unique is still being able to do the AI recommendations—all the deep learning on what, out of the 1,000 things on Netflix you would enjoy, you would enjoy most at what time. That's still a big area of tech innovation. Gaming is another area. We're trying to push into different types of games and figure out gaming in addition to TV series and films.

Why do gaming at all? If you're so good at the core thing and there's room for scale still—you're only at 10%—why bother with gaming?

Yeah. We used to just be movies, and then we expanded into TV series. We're really glad we did that. Then we expanded into unscripted content—Love Is Blind, for example. We've always been expanding into new categories, and gaming is just another category of entertainment.

We've got some cool things going on with the TV, where your phone is the remote control. That has higher latency, but it's easy for party-mode-type games, and it's really fun for these sorts of social interactions.

How do you know when to keep betting on something, and how long-term to be behind something? Gaming is a great example. I'm sure there are examples of things you tried that didn't ultimately work and that you stopped doing.

Sure. Let's do one of those. If you look at The New York Times, in January 2006 there was a launch of Netflix Friends.

So this was friend-to-friend sharing about films and what you were watching.

Facebook was still just at Harvard. We worked for 2 or 3 years on that. Could we get people sharing what DVDs they were picking? Could you give each other different permissions? We tried different permission schemes.

Then Facebook started doing that whole integration, where they did photos and you could share via Facebook. So then we said, “Okay, that's the problem. You don't want to set up your own network. Let's all share via Facebook.” That didn't work any better. Then we tried 1 or 2 other variants, but it was probably 8 solid years.

That's part of what got me on the Facebook board, trying to figure out more of how social was going to be. Ultimately, that probably got solved by TikTok.

How do you think about TikTok? What are your impressions of it?

It's like old cable used to be. You'd change channels and just be there, numb, changing channels, looking for something to watch. But really, it was the numbness of the new, or the endorphin hit of the new thing constantly. It's hitting that part of enjoyment.

It's very creative as a business, and all of that, and very effective. But I would say it's not a thing I want to spend a lot of time on.

When you were CEO, I'm curious how you thought about generating and keeping business power, which leads to free cash flow, and then the allocation of free cash flow. Those seem to be especially important things once you've got product-market fit, you're growing, and you're huge.

11. Power

How much would you sit down and think about where our power comes from? Is it scale? Is it some other cornered resource? Is it some set of different things? Did you really guide the decisions to get more power? How much was that specifically on your mind?

Power is a way of saying above-market margins. The theory is that we can all earn a market rate of maybe 6%, but to earn above that is because it's hard for competitors to do what you do. Then you can get an above-market margin.

We definitely spent time thinking about that— which things should we license our content to, exclusively or non-exclusively, and our deals on televisions and those kinds of things. Television makers would often want to tax us. A typical television maker thinks, “Well, Netflix, you're making a lot of money, so if I'm putting the app on the TV, I want 30%, like Apple gets.”

There would be battles over that. Power is essentially: could they sell a TV without Netflix, or how many members would we lose if Sony televisions, for example, didn't have the Netflix app? That's an example of how that worked out.

Amazon and Bezos are very famous for constantly reallocating capital back into the business to keep generating more customer benefit, which Netflix has obviously done as well. How did you think, or would you think, about the point in the company's life cycle to do more harvesting—to pay dividends, to buy back shares, to do this sort of thing? I'm curious how you thought through the capital allocator's toolkit of the things that you could do with the capital you were generating.

Well, in most businesses, that's highly material—building a lot more warehouses or something. But honestly, for Netflix, there was very little in the way of capital allocation. There was the total budget and the budget per show, but the biggest shows we have, like Stranger Things, were less than 1% of viewing in a year. We have extreme non-concentration, with lots of different budgets and spending spread out.

There was very little capex of any long-term nature. Margins were pretty close to free cash flow, and we always did buybacks with it rather than build it up.

The related tension was how profitable, how soon. It wasn't strictly a cash question; it was essentially a P&L margin question. What we decided was, let's have low margins relative to cable, which ran at around 35% to 40% margins, so that we can invest a higher percentage of revenue into content and have better content for our revenue level than we would otherwise.

That became the fundamental lens through which you ran the business, and they still run it today.

How did you know when it was time to leave being full-time CEO?

Because Greg and Ted were ready. I'd been developing them for at least a decade, and I felt like, coming out of COVID, they were ready. Unless I was going to be around for another decade and train a different set of people to take over, this was the time.

So it was really driven by them. Since they took over, they've tripled the stock, and they've done incredibly well.

How does something like the set of ideas we've talked about so far translate to a totally different domain, like what you're doing with Powder Mountain? It seems like such a wildly different project in almost every way that I can imagine. How much directly translates, and how much needs to be left behind given the different nature of the project?

12. Powder Mountain

Powder Mountain is a ski mountain and real estate development that fell on hard times in Utah. The original people running it ran out of money, so they never finished a lot of the project. We happen to have a house there and love the place. The natural beauty is insane. It's 10,000 acres.

After retiring from Netflix, I decided to take control of it, invest in it, and do a turnaround. So then it's rebuilding the staff and rebuilding the vision. I would say 90-plus percent of the talent-density, No Rules Rules model has worked extremely well: the ability to move fast, hire incredible people, and have them do things.

Everyone being very creative—the talent-density model has been worth the pain, meaning the turnover, and has created an amazing set of leaders throughout the company.

How did you approach it from the beginning in terms of the original vision and plan? It's a distressed asset that you go in and buy. How do you determine the initial vision, and what were the first couple of steps to execute against it?

There, it was a series of transactions to gain control. It took 6 months to buy out a majority of the company shareholders and have control. Everyone wants the billionaire to pay a lot, and I had to be clear with them that this thing could collapse and that if I didn't come in, it might. That was stage 1.

Then stage 2 was figuring out: okay, this is a great mountain, but if half of it were private, like Yellowstone Club, and half stayed public as it was, then it could be a real win-win. They could share operating costs and be more efficient, and we could then have a very uncrowded resort on the public side.

That gets to something that's gone on in the ski industry, which is high crowds. It gets to compete with that. On the private side, it's building a 650-home community of ski lovers where they get their own enormous ski resort, basically the size of Heavenly or likely Vail, just for the 600 homes. It's pretty spectacular in terms of what drives the ski business.

What, aside from the real estate stuff, are the most important variables or considerations that you've figured out in your studying of its history?

Skiing is about 1/8, or 1/10, as big as golf in terms of the number of people playing. I'd love to close some of that gap. It's cold, but it's very family-oriented.

You get outdoors. It’s social with your friends on the lift. It’s got some of those same properties. Interestingly, there are 25,000 golf courses in the U.S., and about 20%, or 4,000, are private golf courses.

In private golf, you get better tee times, a nice clubhouse atmosphere, and a social environment where you get to know people. That’s really what private skiing is. There are about 500 ski areas instead of 25,000, but only 3 are private: Yellowstone Club, Wasatch Peaks Ranch, and Powder Mountain. It’s a very underserved market relative to golf.

What’s most fun about it to you—the whole project?

It’s very right-brained. Everything at Netflix was very strategic, logical, and involved a lot of big competitors. In skiing, the competitors are very cooperative. I think it’s because you have 20 or 30 miles between you, so it’s a lot more collegial.

It’s aesthetic. The big wins we’ve had have been building up the art at Powder Mountain. There’s a lot of outdoor land art that’s incredibly beautiful to ski through. If you’ve had the good fortune to go to Storm King, north of Manhattan—think of Storm King on a ski mountain.

Skiing through it.

Yes, skiing through it.

Tell me about that part of it. How did you conceive of that, and how did you execute it? How does one acquire Storm King-like art for a ski mountain?

I think that, for your audience, the conceptual part is the key. We wanted to have a ski resort and differentiate it. What are we going to do in the summer? You could do zip lines and mountain biking, but it’s all been done over and over. Frankly, it’s high adrenaline, and it’s not that great a match for real estate sales. Most importantly, it’s conventional. It’s been done.

What’s interesting and scalable and fantastic, but hasn’t been done? That’s the art part. I’d been to Storm King, and Storm King is 600 acres. It’s not like a mountain, but it is outdoor sculpture and incredibly stunning. So, again, it was that synthesis of trying to do that on a mountain. Then it was bringing in the curators and getting the work going. Now we’ve got dozens of pieces already in and a lot more coming, and that side is really coming together as the heart of our summer and fall experience.

How did you decide to focus so much on education as one of the buckets of your time? We talked about Powder Mountain, but education, charter schools, and philanthropy as well are a huge chunk of your time. What was it about that sector that drew you? I’m curious for you to riff on the problems that you see in the space.

13. AI in Education

I spend probably a third of my time on Powder Mountain because it’s a joy. On the education side, I was a high school math teacher as my first job out of college, so I’ve always cared about K–12, and I’ve done a lot of philanthropy in that sector over the last 25 years.

The new big thing is AI, so it’s easy to put those together. How are we going to apply AI? It’s super well articulated by your prior guest around Alpha School: kids should be taught individually, as opposed to having a teacher stand in front of a class and lecture to them. That industrial model of the teacher as the sage on the stage needs to be replaced with individualized tutoring.

Prior to AI, individualized tutoring would cost $100,000 a year per kid, so it was out of reach for everyone. Now, with software, we can have individualized instruction, and the teachers become more like social workers. They’re helping with discussion, social-emotional learning, and a lot of the more human and emotional factors. The content transfer—what were the roots of the Civil War, how to do fractions—that’s all becoming software, hopefully as quickly as possible, because then it’s very global and kids will learn more.

What do you think we can do to speed that up the most? You mentioned it could take decades because of the regulated nature of schools. Things move slowly. What could we do that would speed that up?

It’s focusing on apps that really help kids learn more. It’s helping parents see that they all wonder, “Hey, with AI coming, and my kid’s 6 or 16, what’s going to happen to them in the workplace?” They need more and better skills than ever.

Every 16-year-old is learning about AI anyway, so it’s having them be more focused on that and less on traditional classrooms. When you think about classrooms, we use them in K–12, we use them in college, and then, in the workplace, we never use them again. You did all this classroom learning, and it has no bearing on your working life.

Again, it’s really driving up the percentage of kids’ time that’s not in a classroom. As Joe says, it’s helping kids really love school, because then they’ll continue to love learning. The classroom boredom and frustration is at the heart of it.

I’m curious, as you think about the future just broadly across all your interests—you’ve got a cool purview on the world—what most worries you, and what most excites you about the future?

14. Future Risks & Upside of AI

I’m part of the Anthropic camp, where it’s good to talk about the negatives—not because we think they’re going to happen, but because we’ll lower the chance of them happening if we’re honest and talk about them. I don’t think the AI boomer-and-doomer thing is that useful. I think we all want to acknowledge that there are some pretty significant risks, but they’re not dispositive, and we humans may be able to capture tremendous benefits by harnessing AI for a higher quality of life on a global basis.

I’m on Team Human for making that happen. I would say that’s the biggest swing factor of the next 50 years: how well we do that.

What do you think the biggest risks are?

The near-term risks are unemployment causing societal chaos and strife. If you were to get a lot of unemployment, you might get radical politicians promising to get rid of AI or promising to do other things, and that destabilizes society.

Then there’s the long-term power competition between us and, say, China. Does war become, “How many robots do you produce?” It’d be unfortunate if we both ended up having to spend a bunch of money on that because of distrust. A new Cold War would soak up a lot of GDP growth.

The benefit side would be that we cure disease and get nuclear fusion, with huge amounts of low-cost energy. Humans don’t have to work as much, maybe not at all. They get to do things like learn chess and learn how to play all kinds of games, or learn biology for fun, like you learn chess today. There’s tremendous upside to automating a lot of this and taking it to the next level.

My traditional closing question for every interview is the same: What is the kindest thing that anyone’s ever done for you?

Thirty years ago, I worked at a startup. I was a frontline engineer, and I was 28, so I was doing all-nighters all the time. I used to have coffee cups spread around my desk, and over a couple of days they would get kind of ugly and messy. The janitor, every now and then, would clean them all, and I’d come in and there would be clean mugs. I didn’t think about it that much.

15. The Kindest Thing

One morning, I woke up early. In those days, you had to go into the office because the computers were there; you couldn’t take them home. I went into the office at 4:35 in the morning, walked in, went into the bathroom, and there was my CEO washing coffee cups.

I looked at him and said, “Barry, are those my cups?” He said, “Yeah.” I said, “Have you been washing my cups all year?” He said, “Yeah.” I said, “Why?” And he said, “You do so much for us, and this is the one thing I could do for you.”

I was very moved by his humility, caring, and kindness. I thought, “God, I’ll follow this guy to the ends of the earth.” Simple gestures.

Holy cow. Great story. Amazing place to close. Thank you so much for your time.

Real pleasure, Patrick.