But Netflix seems to be the exception, and that’s pure bait for Decoder — how does that structure work, and how do decisions actually get made with two CEOs? A lot of Netflix’s success is in the marriage of content and technology, and having those split up seems like a recipe for conflict or, at the very least, confusion.

Greg and I talked about this a lot — so much of the solution at Netflix is the company’s unique culture, which gives employees an enormous amount of freedom. That culture was first laid out in a famous 2009 slide deck, which former Meta COO Sheryl Sandberg once called the most important document to ever come out of Silicon Valley. That deck was refined into a memo in 2017; the new version Greg’s here to talk about is a shorter, even more streamlined version of that memo.

I have always loved the original culture deck. It’s written in a pretty brash and direct style — it made clear that just saying the company had values wasn’t enough and that the real indicators of a company’s culture were the behaviors and skills that it values and rewards. Most importantly, it said that Netflix wasn’t trying to be a family — its approach to hiring and firing would be more like a sports team, where low performers would be given generous severance packages and replaced with better players on a consistent basis.

This newest version of the memo is not quite so in your face as the original. In fact, it’s a lot nicer and more corporate. I asked Greg why that was, and his answer was fascinating: as the company grows, the memo needs to resonate with a very broad audience, and the tone matters. You’ll hear him talk about this in detail.

Another big change Greg and I talked about is that Netflix removed an oft-quoted section of the original deck labeled “freedom and responsibility.” You’ll hear him talk about needing to more clearly say that freedom requires even more responsibility.

This conversation really hit on so many of the big themes that come up so often on Decoder, and I also found a way to ask Greg what he thinks of the Samsung Frame TV. You know I had to.

Okay, Netflix co-CEO Greg Peters. Here we go.

Disclosure: I executive produced a Netflix show called The Future Of back in 2022.

This transcript has been lightly edited for length and clarity.

Greg Peters is here, the co-CEO of Netflix, welcome to Decoder.

Thanks for having me. It’s great to be here.

I am really excited to talk to you. You have been the co-CEO for about 18 months now. You were the chief product officer and the chief operating officer before that.

That’s correct.

There’s also a lot going on in streaming generally. We seem to be in a time of bundling and consolidation. And then, on top of it, there’s a new culture memo at Netflix, which is a big deal. Netflix’s culture memos tend to set the tone for the whole industry. Our show is about work charts and decisions and corporate culture, so I have a lot of questions about that, but I want to start at the start, which is you used to work at TiVo, and I used to cover TiVo every minute of every day. Are you ready for a full hour of TiVo questions?

We can do that if you want. Just to be clear: I did a startup that was acquired by Macrovision, which eventually became Rovi, which eventually became TiVo. So there’s a long history of those names there.

All of that is why I used to cover every minute of the twists and turns of that company. But let’s set that aside. Let’s start with Netflix. I feel like there are a lot of similarities between what happened on that side with cable and what is going on with streaming now. I want to start with a big question: How do you think of Netflix right now? Is it a streaming company? Is it a TV company? Is it entertainment because you also have games now? What’s the big-picture definition of Netflix?

We definitely think of ourselves as an entertainment company. We seek to be an entertainment company that does what’s very hard for other companies to do, which is bring a real capability and competency to the creative aspect of it. You see those competencies in many traditional media companies. Disney and movie studios and broadcasters around the world do a great job of these things. But we also try to bring a real tech and product capability and sensibility as well.

I think you see that in other companies: Google is amazing at that; Apple’s amazing at that. But it’s very rare to be able to bring these two centers together and be strong at both. That’s what we’re trying to do and be a global entertainment company as well. Partly what we’re trying to do is help entertain the over half a billion people around the world that currently come to us for that entertainment, and hopefully grow that number over time.

Reed Hastings, the former CEO of Netflix, once famously said the company’s biggest competitor was sleep. This is a statement of purpose about the attention economy. You’re only awake so many hours a day; your eyes are only looking at so many things. This is a zero-sum game. The thing that people will choose over everything is sleep if they can. Is that still the framework that you’re using?

Generally, yes. I think that the import of that is still relevant. I don’t think of it as strictly zero sum because there are some magnifying effects across different things, and I think that what we see is that your enthusiasm, your fandom for a certain title, can show up on social media, for example, in a way that’s additive. 

But in the general sense, the entertainment landscape is highly competitive right now. Humans on the planet have more opportunities to entertain themselves than ever before. In that sense, we have to make sure that the entertainment we’re providing is highly compelling, it’s highly attractive, and it can win more of those hours. We call them moments of truth — that’s our way of thinking about if we are doing a good job of entertaining our members.

One thing in the framework of “there are only so many hours a day and we’re competing for attention” that I think is challenging, and I’ve never heard anyone really pull apart well, is that Netflix famously pays for content, Hollywood famously pays for content, and your streaming competitors pay a lot of money for content. That’s a huge part of your economics. The user-generated content platforms do not, or they pay extraordinarily low rates if they do. 

I look at something like TikTok, and it’s full of pirated content. It’s a playground of copyright infringement. It’s very clever. As a former copyright attorney, I find myself very entertained by the innovation that occurs on TikTok on that scale. You look at something like Suits, which is a phenomenon that started with a whole bunch of copyright infringement on a whole bunch of social media platforms. What’s that pipeline like for you? Is that, “Okay, that’s fine, that’s just our marketing,” or are you actually competing with the social media platforms directly?

Well, we’re clearly competing at the margin, back to your attention economy model. But I would say that we’re doing two different entertainment jobs, so I don’t think of the competition as strictly substitutable or head-to-head in that regard. In some sense, those social media channels can operate as a way of magnifying fandom and be a great mechanism for people to talk about the shows they’re watching.

You mentioned Suits, for example. I think when we saw a huge amount of engagement with Suits, we saw part of that magnified and reflected in what people were talking about on social media. You mentioned copyright. I’d say, at the extreme, we, in the industry and generally governments around the world, are supportive of protecting the rights of the people that hold those copyrights. But people talking about the shows that they love, that’s a good thing.

Just on the basic economics, it’s still challenging to me. Those platforms get to take a lot of content for free or very, very low rates and address a big audience, and then their model is to sell targeted advertising. Netflix’s model, until recently, was you paid a whole bunch of money for content at very high rates, and then your model is subscription fees are shot higher. But now your model is targeted advertising. Are you getting the higher rates from the ads to support how much money is going into making the content, or are you finding the pressure of the social media platforms that are just getting it for free to be a real driver of your decisions?

I don’t think the competitive dynamic that you’re describing there is as sharp as you’re relaying. Evidence of that is that YouTube, for example, has largely been unsuccessful at delivering the kind of content that we deliver. There are a bunch of examples like this. Cobra Kai is my favorite. It started on YouTube but didn’t do particularly well there. Then we picked it up from them, and it’s become a gigantic show. I think they’re serving one need — one entertainment need state, if you will — and we’re serving a different one.

Then you talked about ads. My principle orientation toward what we are doing in ads is that it allows us to offer consumers around the world a lower price point, and that’s a powerful thing to do because now we can say to folks, “Hey, you can have access to all the amazing stories, the films, the series, the games that we are doing at a lower price,” and that’s super important for them. Then we can also work with advertisers to make up the difference between that and what we would otherwise charge on subscription via the advertising revenue that we get there.

Do you think that the introduction of ads is changing how you think about running the company? You’re at the Cannes Lions festival right now, which is the big advertising festival in France. When you ran a subscription business, I’m guessing you didn’t have to spend as much time talking to advertisers in France. It’s tough work, I feel very badly for you, but I’m guessing that’s a new piece of the puzzle. Did you have to build up that capability?

Yeah, for sure. It’s a new capability from a technology perspective, so we’re building technology that supports what we hope is at some point in time — and we’re still very much in the process of building this — the best-in-class streaming solution for consumers, our members, and advertisers. We’re also building capabilities that we haven’t had in the company before. We never had ad sales, for example. That’s a whole new muscle for us to go build, so it’s definitely about growing that.

.Maybe the question behind your question — and a question that I’ve gotten internally — is, “Are we programming differently as a result of now adding advertising to our model?” And the answer to that question is no because, at the end of the day, whether it’s subscription or advertising, both are anchored in something that consumers want to watch, that our members are dying to watch.

That engagement is at the center of both models. It’s mutually supportive in that regard. We very much think our programming job is to think about what our members need and how we can give them that, and that drives and then fuels both models in a positive way.

I want to come back to that because I think that perceived shift is really important, especially as you add advertising, because the market for advertising is dominated by other big tech companies. Like you said, YouTube is not out there selling premium programming the way that you are, but it’s still all just advertising.

I want to unpack that, but I also want to make sure I get to the Decoder questions because I think they will help me understand how that works in your brain. Famously, there are two CEOs at Netflix. There’s you and Ted Sarandos. Historically, having two CEOs has gone sideways. BlackBerry didn’t work out. It had two CEOs. How is that structured between you and Ted, and how is that working after 18 months?

Having two CEOs is an incredibly powerful model, and it allows us to essentially embody what I mentioned at the beginning, which is we seek to be a company that brings two very strong centers of excellence and capability on the creative side and on the product and tech side together. If you look around the world at our competitors in the space, they typically are run by one CEO, and it’s a CEO who’s either really anchored in one or the other of those worlds. I think that reflects the strengths and weaknesses that those companies have. Ours is a very powerful model. It’s hard to pull off. I think the reason we’ve seen it not be done so often historically is because you need a culture and a set of folks that are bought into this model.

That’s where I go back to our culture and think about it as an enabling function for why the co-CEO model works at Netflix. You think a lot about what we do in culture, the willingness to be transparent with each other, to have honest conversations, to be selfless, to put the company first. These are all things that make the co-CEO model work. Whereas, in other places without those cultural elements, it’s very hard or it’s harder to pull off.

We did do a little bit of a historical look, and I would say that the performance of co-CEO-led companies is bimodal. It either significantly underperforms peers or it significantly overperforms peers. The real discriminating factor there is, do you have that culture and do you have a set of CEOs who are bought into the model and see it as a positive rather than something that’s a compromise or obstacle they have to work around?

Maybe I over-index on the historical underperformers. I think the way people think about it reductively is that you are the tech CEO and Ted is the content CEO. Is it that simple? What’s the overlap?

There’s definitely truth to that. We both feel that we’re responsible for the company at the end of the day. The next level of leaders both report into us. The major decisions that we make as the company, we feel like we have to make together. We always say speaking to one of us is speaking to both of us, so we feel like we owe each other clarity and transparency in those conversations. But it is also true that one of the reasons this model functions for us is that we get the benefit of that collaboration, but also, when you think about speed of decision-making, we have lanes that we are primary in and we know that we can focus on.

If there’s a big content strategy question, depending on how big it is, we’ll both weigh in. But maybe for something that sits below that, Ted is going to have a much better-informed position than I will and he’s going to make a call more quickly. Quite frankly, even more so, our chief content officer, Bella [Bajaria], will be doing that. That allows a speed of decision-making that is part of why this works.

So how is Netflix structured? You said that line below reports to both of you. How does that work?

That next set of leaders reports to both of us. Now, we have a primary responsibility, so you think about content, marketing, legal, comms, and publicity — those are on Ted’s side. On my side, for example, we’ve got product tech, ads, games, finance. That’s an example of how we are taking primary responsibility for the activities that sit at that level.

On your side of the house, you’ve got the core engineering functions, the core product functions. There’s a real dynamic between the product, how it works, and what people experience, which is a bunch of content tiles and content recommendations. Where’s the balance there? 

I think a lot of companies think video streaming is just like a white-label product, and you can just buy it and put in your content and you’re off to the races. Netflix is famously not that. How do you think about that mix? Because everyone’s kind of standardized on your approach. Everything kind of looks like Netflix now. Is there still room to innovate?

There’s definitely room to innovate. In fact, we just launched a new version of our TV UI, which is where the majority of engagement takes place on Netflix. I also think it’s important to distinguish what is copyable and what’s not. You’re absolutely right that we do an update to our UI, and I can click a stopwatch and basically count the time until our competitors align on the pixel component of that version. And it ranges. The better competitors can do it in 18 to 24 months, and then some of the folks that take a little bit longer to get going, it can be years before they get there, but they all get there eventually.

An important component of what makes the user experience is all the things that then populate those pixels. How do we think about title recommendations? How do we present titles and use assets in different ways that make it more compelling for different users around the world? Because you and I might both be interested in season 3 of Bridgerton and be super excited about watching it, but we might be interested in it for different reasons, and our ability to communicate those differences effectively is super important to unlocking the value of that incredible storytelling. That’s harder to copy. I’d say no one’s done as good a job as we’ve done with it. Some folks are not even personalizing, which is remarkable to me at this point in time.

Who’s not personalizing?

Well, I’ll let you do your research.

I’ve heard some rumors. I’m wondering if you want to say it out loud.

No, it’s what you’d expect, which is the more traditional media folks are slower to that process. That’s a strength that we can bring that they don’t necessarily have the same fluency of accessing. But then, on the flip side, you’ve got incredible tech companies that live, eat, breathe personalization, but then the quality of their content offering is not the same and their ability to produce amazing shows at scale around the world is something that they just don’t have. We see both sides, and really, the magic for us is bringing those two things together in a way that magnifies them.

That’s the user experience that’s the most forward-facing part of the product stack. Netflix is famous for spending a lot of time and money on core engineering problems. Moving video on the internet was not a solved problem in the early days of Netflix. How much of your time is spent on video and coding and caching network distribution deals and peering arrangements?

Well, it’s still an important part of how we think we can improve the user experience. It’s remarkable to me that you mentioned encoding. And encoding efficiency… I would’ve thought five, seven, eight years ago that we had grounded out on the ability to squeeze more efficiency, which is delivering higher-quality video under less and less good network conditions. Yet something we participate in in the industry writ large is uncovering even more advances in that space, which is shocking to me, but it’s incredibly impressive.

We think that there are more opportunities out there to do that level of work and deliver better experiences to our members. I’ll give you other examples. As you mentioned, content delivery, moving video around the internet, part of this is what we’re working on in games as well. Ultimately, what we want to do is be able to deliver game experiences to the TV, and we’re going to do that by rendering the actual game in the cloud and streaming it essentially as a video. That’s an opportunity to leverage the infrastructure work that we do by being deeply embedded as a client on TV silicon and working with TV manufacturers to make sure we can bring those high-quality experiences to those devices.

Game streaming is one of those things that I hear about constantly on the show. For a while, you would look at Microsoft and be like, “Game streaming is going to happen.” It has not happened. Why do you think it hasn’t taken off in the way that everyone wanted it to?

Well, I’d say I’m convinced it’s going to happen as well.

Wait. Is it going to happen like self-driving cars are going to happen, or is it going to happen like video on the internet is going to happen?

That’s where I was going to go with this, which is, “What is it that’s going to happen?” We should be a little more clear about what that means. If you think about the pinnacle of what that could be, which is, let’s call it, console-like experiences at console-like latency, that is a very high bar, technically. The cost dynamics of that where we think about the infrastructural requirements around that are super hard.

Quite frankly, that part of the market is incredibly well served right now. You’re talking about the apex of an ecosystem that’s evolved to support that. If you think about the software development model, the software development costs, what the consumer model is, it’s a solved problem for so many people around the world. I would say, back to streaming, where’s the opportunity? The opportunity is actually not to come in from that level but to come up from the bottom and say, “Hey, there’s a range of game experiences that we can deliver to folks that are very capable technically today, where the cost dynamics around it are manageable.” If you think about family game night, couch co-op, there’s a whole set of quite impressive interactive experiences that I think we can do.

One of the challenging parts of games is people play them all over the place. I play them on airplanes, where I don’t have the connectivity. With video, you have an out. You can just put the bits on my phone and you’re not streaming them to me anymore. But my phone can probably just play them back. With a game, there are all kinds of things that might happen. You can’t just put the whole game on my phone necessarily, or my phone might not be as powerful as your server that’s making the game to stream it to me as a video file. How are you solving that problem?

We’re delivering games in multiple different ways. Cloud is one way to do it. And I’d say cloud is relevant because it unlocks TVs, which haven’t really been a source of gameplay except as the screen in which connected devices like consoles manifest themselves.

The TV manufacturers think they will be able to do that. This is what I mean by self-driving cars as a comparison. Everyone thinks they’re going to be able to do this.

We think we will be able to do it in the way that I described it. That envelope of what we can deliver will change every year. We’ve seen it grow, and I don’t know when it will get to that ultimate level, but it’s going to be a while away, and we’ve seen that again and again from the folks that have tried it. Back to your question, back to the phone — the phone is a way more capable general gaming device than a TV is. It has way more compute power in that regard, so we will deliver downloaded games to your phone in a way that is totally different from what we would do with a TV.

Do you want to get to that AAA set of games at the end of the day, or are you going to compete for the more casual games that people are playing right now?

If you step back and look at what we do, we’re in the business of working with creators, the most amazing storytellers on the planet, and having them build compelling universes that we can then deliver to people around the world in a variety of different ways. So that could be non-interactive modes like film and TV, and it could be interactive modes like games. Ultimately, to give fans the most compelling immersive experiences, eventually, we’ll get to something that looks like a AAA model. I would definitely not rule that out.

We’ve walked the same path in other places in the company. If you think about when we add new genres and things like that, sometimes we start smaller and then develop more capability and more aspiration to do it at the highest level. I think that’s a pretty common trajectory for us to follow, and I expect that’s where we’ll go with games as well.

Games seem like a new interesting engineering problem. You said you were surprised the industry can keep figuring video and coding out. Are there core engineering problems that would give you an advantage if you solved them today that you’re putting resources against?

Sure, there always are. Back to games, for example, the ability to virtualize multiple game instances on a single GPU. If we can do that more efficiently, that yields a better, more effective use of infrastructure. That’s an example I would put on the list. Another example is just how we think about recommendation systems.

We’re moving and significantly improving how we dynamically assemble the UI for you. That allows us to then, as you are navigating the UI, actually change what we are presenting in real time based on explicit and implicit signals we get from you around your needs for Netflix today. This is an evolution. We talked about how we personalize today and where other folks are in that process. The next step for us is to think about what you need from us on a Thursday night versus what you need from us when your family is assembled in front of the TV on a Sunday afternoon. It’s different, and the UI should show up in a different way to serve those needs differently.

How is your product team structured? I’m guessing this is a pretty functional organization. That’s a lot of goals. We have to figure out game streaming, we’re working on personalization, we want this more semantic UI. How have you structured the product team to chase all of those ideas?

We are a functionally organized group. Within that structure, we’ve got product leaders that are responsible for various different product categories, and then they have respective engineering teams that they go with, with a little bit of a horizontal layer to it. We’ve got a platform team that’s really focused on infrastructural components, services that are generally available to multiple different end-user-facing feature teams. That’s the core model.

And then the big Decoder question. You have a lot of decisions to make. You have a co-CEO to make decisions with. How do you make decisions? What’s the framework?

Well, maybe the first order of business — and this gets to the culture and how we operate as a company — would be, “Is this a decision that I should make, or is this a decision that should be made somewhere else in the company?” Part of what we want to do is enable folks throughout the entire company to make as many decisions as possible. That might be the first order, like, “Okay, is this my decision, or should it be somebody else’s decision?” And then, if it is my and Ted’s decision, partly, we’re trying to figure out what things are most material to the business. Companies can do a lot. You’re at a certain scale and there’s a lot that you can do. We’ve always been pretty choosy about the strategic extensions that we’ve made.

We’ve been very focused as a company. Our orientation, rather than some of our peers who have a “let’s just try something, shotgun stuff out there, and see what sticks” approach, we take the other extreme, which is, we start with the first principles argument around why this is critical to the business. The qualitative version of this might be, “Can you imagine a Netflix 10 years hence where we are not doing this?”

There’s a more wonky version of this, which is that you think about the materiality to the business: Do we have permission to play? Is there leverage off the core in success at scale? Does it lever back to the core? There are different ways of thinking about the same question, but ultimately you’re saying, “Okay, we’re going to do a limited number of things. Does this materially drive the business?” And if it does, then our orientation is, “We are going to do it to win.” We may not be great when we start, but ultimately we will learn and grow. That’s very much, again, how we think about what our culture is, which is an iterative process of trying to get better and better and seek excellence. That’s how I think about top-level decision-making.

This brings us to the culture memo. There is a new version of it. It just came out. The first one was released in 2009. It was a deck that had over 100 pages. Reed Hastings even wrote a book about this deck.

And Netflix streamlined it into a doc in 2017. So it went from being a 100-page PowerPoint to one sheet. Now there’s a new one, which is even more streamlined. Two questions: Why make such a big deal about these culture memos? And why does it keep getting shorter?

You typically have this triumvirate. There’s culture, strategy, and execution. There are probably 100 business books that will tell you that execution is the most important, or strategy is the most important, or culture is the most important. We think, first of all, it’s important to note that you have to be good at all three to be a really successful company. So you can’t be awful at one. But if I were to think about a hierarchy, I would put culture at the top. The rationale for that is if you have an amazing culture and maybe you’re mediocre at strategy or execution, an amazing culture allows you to get better at the rest. It’s a vehicle to improve, so that’s why we make a big deal out of it.

If you really think about what we’re trying to do with the culture memo and why it’s changed so much, we’re trying to constantly do a better job at articulating the practices we can employ as a company to grow and get better — to learn, to seek excellence, and to constantly strive for that. That’s the rationale. And then, why does it get shorter? Probably because we get better at articulating it. What’s the Mark Twain quote? If I had more time, I would’ve written you a shorter letter. It’s an investment in that. And one, understanding the things that matter. So it’s about seeking signal from noise and trying to constantly refine that articulation to have the maximum value for the time that somebody spends on it.

Just before we started talking, I went back and read the first one from 2009. I have to say, sitting here in 2024, I would describe the original deck as refreshingly harsh. It’s mean. It’s not a nice document. It’s very blunt. It opens by saying, “Even Enron says it has integrity. But we’re not going to lie to you. Here’s what we actually value.” The new one is a little softer. It’s a little more corporate. It’s nicer. Is that just meeting a younger employee base? Why make it nicer?

Well, it’s interesting because you called the original one “mean,” and I wouldn’t have used that word.

It’s very direct.

Direct would be the word that I would use. What we’re constantly trying to do is communicate effectively who we seek to be and who we think we are. What we found is that, in the original articulation, the tone of it incorrectly communicated a sense that it was a harsh and maybe cutthroat place, which actually, it’s really not. We want to be very clear that we do think excellence in having the colleagues around you is super important. We have this “dream team” concept in the latest version in which we are trying to be very clear about this being more of a sports team model than a family model. We’re going to seek the best player in every position. To the degree the business evolves or moves and we think there’s a change that needs to be made, we will make it.

We want to be very clear and honest about that. But we also want the tone to reflect what people’s working experience is like inside of Netflix. It’s interesting you brought it up because I think you’re looking at these two points, which is the original doc, and then where we are today. What happened in the middle actually got softened a little too much, I think. We moved the pendulum too far in the other direction, and partly what we were doing in this version is trying to bring more of that directness in while still accurately conveying what it’s like to work at Netflix.

That middle period coincided with a lot of controversy around the content on Netflix, particularly Dave Chappelle’s specials and whether Netflix’s employees were being reflected in the content that was on the service. There was a lot of tension inside the company at that time. I’m noticing, in the new memo, there’s no longer this famous section titled “freedom and responsibility.” Are those things connected? There is a section that discusses the diversity of the content on the service and how you’re going to meet the audience where they are. But there isn’t the sense that Netflix’s employees across the board are going to get a say on what appears on the service.

We actually explicitly say the opposite. You’re capturing two things that are pretty distinct. So one, being direct and setting clear expectations about who we are as a company so people can make good decisions about whether they want to join us. We wanted to be very clear about the fact that we are trying to program for over half a billion people around the planet. Those people’s tastes are very varied, and we have to bring a wide variety of content to be satisfying for those folks. We are definitely supporting artistic expression and a wide variety of creators to do that. We wanted to be very clear about that and take that head-on so folks could understand and decide that “hey, Netflix is a place that I can work for.”

And then [the section titled] “freedom and responsibility,” I would say that’s a different thing. What we found is that, in the many years since that original culture memo, people heard increasingly about the freedom part and less about the responsibility part. We always meant that to be a balance. Essentially, what we were trying to say is, “If everyone’s super responsible, we don’t have to have a lot of rules because we can rely on people’s judgment and people will do great things.”

What people ended up hearing was, “Oh, I can join Netflix and essentially all the decisions that I make are mine, and I don’t have to worry about that overarching responsibility to our collective corporate goals.” So we said, “Okay, this is an opportunity to try and clarify that, rearticulate it in a way that grounds into the original concept.” That’s why we’ve shifted the language.

One of the key values that you’ve had the whole time that I think connects to that is something called “People over Process,” where you’re just going to trust employees, where the expense policy is just “act in Netflix’s best interests.” That works when you’re the challenger, when you’re the startup and maybe you’re focused on one thing. It gets harder when you’re big and the company can do a lot of things. How do you keep that aligned at scale?

I don’t think of it that way. You got to size, which is an important dynamic. But going back to one of the core principles we have behind the culture memo is this idea that a lot of the way companies are run right now comes from the Industrial Revolution and industrial processes. There’s this idea that there’s a manufacturing process, and you get hyper-focused on how you can be amazing at that process. You think about defect reduction and it’s great because, if you’re doing that one thing, you can do it amazingly well. Look, if you’re TSMC and you’re building chips, that is your bread and butter. You do that all day long, and that’s an important part of corporate success.

For us, what we found is that refinement actually comes with brittleness. What happens is you get really good at one thing and then, when you want to do something new, the brilliance and the optimization that you’ve built in creates a lot of rejection of the new. Either it’s incompetence or lack of capability or even what we see in big companies, where all the systems are refined to say “we do X, we don’t do Y” or “we don’t do X and Y.” You get a lot of antibody rejection of Y.

A lot of what we are doing is trying to figure out how we can be a company that can go through multiple iterations of changing the strategy — adding a Y, adding a Z — and have a culture that accommodates that. Partly what we’re trying to do is have amazing judgment, have people in place that we have high expectations of, that are unusually responsible compared to their peers, and that gives them the ability to be flexible and evolve and not have the brittleness of an overly processed company. We like a good process. There are some components of having people work together as a collective unit that require scaffolding and expectations and how you set that up for success. But a lot of the process that gets in is not about that.

Big company process is about how you defend against the employee that expensed something wrong, or whatever, and a lot of that stuff, when you look at it, smart people with good judgment are not going to go do that stuff at the end of the day. We don’t think about that as good process. We reject that. We’d rather lean into, “Let’s hire good people. Let’s set the expectation that people are going to have great judgment, and then that gives them the flexibility to deal with the incredible nuance that exists around what we’re doing.”

You think about programming for over half a billion people — with countries in different situations and every title in a different situation. The user experience changes in these ways. Whether you’re a creator, an engineer, or a game developer, you want that flexibility to engage with that nuance and make smart decisions based on the specific situation that you’re dealing with, and that’s what we try to foster and encourage.

One more question here, then I want to put this into practice and come back to ads and culture. One of the most famous aspects of this memo, some of the most famous vocabulary words, are “loosely coupled, highly aligned” and “disagree and commit.” The second these words were introduced into American business, everyone started saying them. I used to work at AOL, and people would tell me we were loosely coupled and highly aligned. People say these things, but it’s hard to achieve them. You have new people coming in, and Netflix is a very different company. I think about those ideas, and what I see is you need a lot of institutional knowledge to stay highly aligned.

People need to know what you’ve done, what you might do, the ideas that you’ve had, all that nuance you’re talking about. When you have employees coming in, when you have a ferocious battle for talent and AI, for example, people are coming and going, how do you maintain that? It feels like that relies on some amount of history and institutional knowledge to keep going. Otherwise, you’ve just got people reading wikis all day long.

To some degree, yes, but I think there are mechanisms you can employ that don’t rely on tribal knowledge. As an example, “knowing what your north star is” is a very strong mechanism to support alignment. In addition to this culture memo, we have another doc that we don’t publish, which is our strategy bets.

This is a formalization of ambiguous decisions, meaning the company could do A or B. We’re very clear about, “Okay, it’s ambiguous, but we are going to do A and not B, and this is why.” That is another touchstone that allows employees to operate independently to that loosely coupled model but with a sense of how this ladders up into an aligned position.

This brings me to ads and the culture of the company and the market that you’re in. Ads are a tribe — you are in Cannes, you need an ad sales division. There are lots of ways to build an ad sales division, but that’s a different thing. Historically, at the big tech companies, the ad sales division does not look or feel or act like the product division, and there’s often some tension there. How are you bringing in a whole group like that one that feels like the future of your revenue without changing the culture of the company?

It’s funny that you mention it because we have tribes already. We had tribes before this. Let’s break it down. Kernel engineers who work on our content delivery network versus mobile UI engineers, that’s a different tribe. They have different ways of thinking about the world. We’ve always been balancing different groups of people that think about things differently.

That’s why we have these memos because, quite frankly, an evolution of the culture memo was going from the original deck that you talked about, which I would call highly anchored in Silicon Valley — there was a libertarian philosophy or ethos that you could read in there. And partly what we were trying to do over these multiple iterations was tease apart the actual outcomes, the excellence-enabling things that we wanted to identify, versus the behaviors and the articulations, and trying to make those more general.

That way, whether you’re two creatives in Amsterdam who might have a very different form of communication when you talk about direct feedback versus two engineers in Japan, you can have an honest conversation with each other about what we’re good at, what we’re not. We think that’s an important enabling factor that we want to have around the world, or around all those tribes. You mentioned ads as being a super distinct thing, and I don’t think of it as that distinct when it comes down to it because we’ve always been trying to blend these different base cultures or different expectations and then figure out how to layer on top of that a way of working and a set of expectations that allow us to be better together.

I don’t disagree, but it feels like the premium advertising, the brand advertising world, going to Upfronts and meeting with marketers and meeting their demands, that just places a different set of pressures on an engineering organization. You can see it at Google, at Meta. At scale, this is what they do. That’s who you’re competing with now, but you’re also competing with linear TV for their ad dollars. How do you think about, “Okay, we’ve got to build a whole bunch of ad tech now that competes with giant incumbents like Google and Meta but that also takes the brand dollars away from linear television”?

Part of this is knowing what role in the ads ecosystem we can or should play. What’s the strategy? Our position is the ability to bring what is amazing about TV advertising. So, if you think about the creative formats, how we were able to position brands next to titles that were culture-defining moments, essentially, which is, of course, where advertisers are excited to be. And then, you need to also pair that with what is great about digital advertising. So, if you think about targeting performance measurement, the ability to do that at all levels in the funnel.

Our job is not to be the most amazing gigantic-scale digital advertiser, because Google and Facebook are going to do a better job of that. Our job is not to be just a TV advertiser because other folks can do that as well. So we’re trying to bring these two things together in a way that we’re uniquely suited over a five- to 10-year period to magnify the impact of both of those worlds, to create a premium, highly creative space. You couldn’t do these kinds of formats on TikTok but also bring enough targeting, enough brand performance, in these different measurements to try and give advertisers a sense that they’re getting something for their dollar as well. So that’s how I think about blending those.

You mentioned TikTok. I see a lot of younger people, they open Netflix, they just stream whatever in the background, and then they look at TikTok. That’s a real dynamic. TikTok, Instagram, YouTube — on the phone, they can convert. You see the thing, you push the button, you go buy whatever. TikTok wants to sell me a lot of off-brand power tool batteries. I feel like I’m going to light my house on fire, but that’s where they’re at right now. Netflix, you can’t convert on the TV. It’s much harder. Is that a challenge for you, to go to advertisers and say, “Look, you can’t actually shop, but maybe you’ll get more take-up, even though the kids are holding their phones”?

Again, it gets to what role we can play. The demand conversion advertising, that plays a role in the ecosystem, so 100 percent that’s there. Or maybe that moment can’t happen on the phone regardless — if you think you’re going to go out and buy a car, that’s likely not happening there. But there is a lot that we can do that basically operates above the funnel in ways that brand advertisers believe is more compelling. Obviously, we should be able to measure that performance as well. How are we doing in terms of brand affinity awareness? How are we thinking about building demand at certain levels? Our goal would be to operate at that part of the funnel and then think about fulfillment as happening in a wide range of channels after that fact.

You listen to Meta’s quarterly earnings, or you just read its statements, and it feels like it can turn the knob on ad load and make more money and hit its numbers if it needs to. There’s a product manager at Meta who’s like, “It’s fine, Mark [Zuckerberg], I got it.” And then there are more ads in Instagram Reels today. You’ve been raising prices, you’ve been hiking the streaming subscription prices, and that seems to be going fine. Would you ever increase the ad load? Is that a lever that you have at your disposal?

It is a lever that’s at our disposal. It’s a lever that you use at your own peril. I would say my goal is that we will change the ad load — we’ll decrease it over time. If we fulfill that promise that we just talked about, we should be able to increasingly provide higher-relevance, higher-value ads. That’s great for the member, they have a better experience as a result, and it’s great for the advertiser. Quite frankly, advertisers will pay more for that kind of experience because it’s more valuable. I think we should be able to do that.

You can decrease ad load and say, “Look, these slots are more valuable because there are fewer of them. We’re going to charge higher prices.” Has that played out yet? Have you seen evidence that’s true?

We’re very early in this process, so, at this point, we’re trying to build an ad server. We can see on the horizon tons of work to do. We have an amazing roadmap in front of us, and again, back to the culture memo, we’re very much comfortable starting from a position of “we’re at zero, and we will learn to be better at this.” That’s definitely the process that we’re at in terms of building ads is iterating toward that model.

Do you think you’re at the end of price hikes? I have a graph in front of me of Netflix price hikes, and the verticality of this line, especially for premium with no ads, is pretty remarkable. It’s just gone up. And are you sending more people to the ads tier, is that the goal of that price increase? Are we done with this yet?

It’s definitely not the goal. What I would say is that our job is to add more value to the entertainment service that we are offering. We see consistently, again and again, that our members want more entertainment, they want a higher diversity of shows, they want more quality shows, so that’s our job. If we do that well, we’ll go back and occasionally ask members to pay a little bit more to keep that flywheel running. Back to the ads component of this, it’s really just how we then offer a wide range, and our goal will be to have an even wider range over time — to have a set of prices with the right features so that basically we can attract more members around the world and they can enjoy the incredible stories that we have available.

I’m a huge TV nerd. I’ve got a giant Atmos and Vision setup in my house, and then all the streaming services ask me to pay extra for 4K. It seems like everyone’s defaulted to, “You’re going to be fine with 1080p. That’s what we’re going to deliver to most people.” Is that because most consumers don’t care? Is it because you can set the price higher and people do care so they will pay more? How is that playing out?

It goes back to what I said, which is that we’re trying to essentially find a range of prices that have appropriate features connected to them so that people who value those features more have an opportunity to get them.

But I guess my question is: is it that people don’t value the features, or they value them so much that they will pay extra?

What’s the difference, really?

My theory is that most people don’t value the features, so there’s a tiny subset of people who will pay, but most people don’t care. Whereas, if everyone cared, it would just become the default and everyone would price it in.

That’s probably a fair statement. Yeah. I’m not sure. There’s clearly a community of folks that care.

They’re my people.

Yeah, and they’re not a small minority. So we want to serve them well, too.

I feel like I have to ask you a question about AI before we end here. We’ve talked a lot about personalization, about modifying the UI, and making it more responsive. There’s a huge talent war for AI. That’s a huge amount of cost just to get, I’m assuming, Nvidia GPUs. That’s what everybody wants. They’re very expensive, and Nvidia seems to be the only company that’s actually making money with this stuff right now. The talent is expensive. What’s your approach to this? Is it “we’ll wait for the models to come and be good for us”? Is it “we have to make our own foundation model”? How are you thinking about it?

We think one of our competitive advantages is to bring two worlds together. So it’s amazing entertainment, the creativity required to produce that, as well as the technology component. We have a long history of using machine learning and artificial intelligence in our recommender systems. We’ve been doing that for 20-some years. Again, we think that our job is to be proactive about understanding where there’s technical innovation. How do we use that both to serve creators, allow them to tell their stories in more compelling ways, and also then to serve our members better user experiences?

I would also say that we’ve generally been careful about not getting caught up in the hype cycle that often informs Silicon Valley. I am so excited that we didn’t spend any energy on NFTs. If we go back three or four years ago, everybody told me, “My God, Netflix has to have an NFT strategy.” We didn’t need to have an NFT strategy. Now, I would say generative AI, that’s not an apt comparison because I do think there are some interesting things associated with it. It’s not just smoke and mirrors.

But I would also say we have to understand what our role is in that ecosystem, and we’re not going to build frontier models. That’s not what our job is, really. Our job is to think about, “Okay, there are a bunch of interesting technologies that are being developed. How do we give these to creators in ways that allow them to tell their stories in a more compelling fashion?” Just as entertainment has evolved with the evolution of new technology, you can go back from cave painting to where we are today. You mentioned 4K and Atmos. Creators will figure out a way to leverage the newest tools to tell their stories in a more compelling way. Our job is to enable that at the end of the day.

When I think about use cases you could use AI for right now, one is to make different thumbnails for all the shows. It seems very obvious. You’re going to take some creative, you’re going to plug in some targeting keywords or personalization keywords, and you’re going to make a new poster for the show that is just for me. That would, charitably I think, piss off every single one of your creators. People doing that right now to make movie posters face enormous amounts of fan and creator backlash.

This is the tension of the company. You’ve got the creator side and the tech side, and even obviously interesting ideas like “let’s improve personalization by making more custom art,” which you’re already automating in some way, seem very tense. How are you balancing out the tension with the creative side?

It’s interesting because we do that right now. We have a team of people that create a variety of assets that represent the titles so that we can find the biggest audience for that title. Now, we do that with a strong sense of authenticity, meaning that it has to be reflective of the title. So we’ve been constantly using tools, with the human in the loop that understands, is this an authentic representation of that title?

Now, if it’s authentic, it doesn’t misportray the title, which not only would capture some creator ire but is not good for members, either. You don’t want to tell them that it’s a story about something and have it be about something else. Clickbait is something that we completely reject in that regard. But if it is authentic, then it serves a purpose that creators love, which is it finds a bigger audience for that story. If you talk to creators, generally, what they want to do is be able to tell their story in a compelling way and find the largest audience possible. We’ve essentially been navigating that situation already.

Do you think that you’ll start using more AI in situations like that that bridge tech and creative because that’s where some of the opportunity here is? If you’re just using it for the personalization algorithm, I don’t think people would be so upset. Once you start touching on the core creative, that’s at least where our audience has been the most vocal with us. Do you think you’re going to venture into the more creative side with AI?

Creators have been using tools that, in my mind, serve some of the same purposes that the AI that we’re seeing, the capabilities that we’re seeing today, serve. If you think about pixel manipulation or VFX, a great example is previsualization. So having a creator be able to have a more efficient way of understanding the script and be able to manipulate that and understand that in a way that then informs a shooting schedule. It’s a great example of how they can use these technologies. Our creators are going to want to use those tools. It’s also important to recognize that a lot of the creators that we use are independent production companies. They decide what tools they want to use. We don’t tell them that.

I want to wrap up on the TV business as a whole because it is undergoing a lot of change. Netflix itself is changing in response to that. One of the big trends we’re seeing everywhere is bundles. We’re going to make a new sports company and bundle all of our sports together. The cable company, the ISP, is now going to bundle some streamer with it. You sign up for Verizon and you get some collection of things.

The famous line is, “There are only two media business models: bundling and unbundling.” We’re obviously in the bundling period. How is Netflix participating in that? Are we just headed back toward a cable bundle but with Netflix at the center?

I don’t think so, but let’s just step back. I think bundling works when it’s better for consumers: where they have an easy opportunity to buy a package of goods that make sense together, and then they see more value in the assemblage of those than they would with each discreet purchase. To be clear, we’ve done this with a variety of partners for a long time.

[We’ve done this for] six, seven, eight years — whether it’s a pay TV operator or an ISP mobile operator. In some countries, we do it with utilities. You can literally, when you buy your energy, your power, you get Netflix with that as an option. We’ve done that because it’s an efficient way for consumers to be able to decide to add Netflix, and it allows us to access an audience that maybe is less tech-forward, or we show up in a way that’s natural. Think about, “Hey, I’m watching TV on a set-top box. I can see that I can get Netflix as part of that.” It’s a model that’s good for consumers. We’ve been doing that for quite some time.

Some of the bundling behavior that you see today, it’s not entirely clear to me that it’s actually going to work, that it’s pro-consumer. To some degree [the mentality] we’re getting is “let’s throw a bunch of stuff together and see what sticks.” We’ll see how that plays out. But I do think that a bundling of services is a logical thing and it’s pro-consumer, so I don’t see that going away.

Would you ever bundle with a competitor the way that some of your competitors are bundling with each other?

I would never say never. I mean, we’ve done a bunch of things that we probably too strongly had an oppositional position to. We want to remain intellectually open and humble about that. We’re being bundled with our competitors by our bundling partners right now. You can go out and buy a mobile service that includes us and somebody else. So, again, we don’t think it changes what we do, which is we’ve got to provide an incredible entertainment experience for our members. If we do that well, then they’re going to see value, and whether people buy directly from us or in a bundle, that’ll work out.

The place where most consumers, including me, would like all these services to get bundled is at the device level, where I open my Roku or my Apple TV and I just see a whole bunch of content for a whole bunch of services, and it figures out what I’ve paid for and I can go watch some stuff. Netflix has historically resisted those integrations. On the Apple TV, you’re not in their app. On other devices, you’re not in their unified experiences. Why is that? Is that because you want people to come to Netflix and stay there? Is it because the business terms aren’t agreeable? What’s the reason?

We do those integrations all the time. Actually, I would’ve described it as much more balanced in terms of when we do it and when we don’t. Essentially, we think that we’re investing significantly in creating a better user experience for discovering Netflix content within our application. We want to make sure that we have a level competitive field on the device for users who want that differentiated experience, who see that value, to be able to get to Netflix in a very easy, efficient way. Mostly, the terms that we have are around “is that a balanced place to do this?” And if it is, then we participate in things like an aggregated service discovery. And where it’s not, then we typically don’t.

Netflix is on basically every platform that you can think of. Famously, there are Netflix buttons on remotes; it’s nonnegotiable for a lot of TV manufacturers. Most of those platforms want to take a cut of your revenue, either at signup on the subscription side and very much so on the advertising side. You cannot appear on a Roku device unless you give them a cut of your ad sales. Are you big enough now that you can negotiate better terms for revenue splits across TV and on mobile, relative to your competitors? Or are the the terms these platforms and TV makers demand pretty flat across the board?

I don’t know what our competitors get, so I can’t give you a definitive answer to that question. But I would expect that, relative to some other folks who would provide different value, we probably bring more to the platform, and that probably shows up in some way, shape, or form.

The reason I ask that question is because Netflix is one of the companies that’s always pushing back against app store rules from Apple and Google and the restrictions they’ve placed on signup flows, customer acquisitions, and, of course, revenue splits. There’s been a lot of regulatory activity in that area in recent years. Are you starting to see the benefits of that regulatory action showing up in how you acquire new customers, get them to sign up to Netflix, and collect the subscription fee every month?

TV is a highly competitive ecosystem, so I would say there are tons of manufacturers out there, and there’s not really a concentration of power. I’d say that’s a well-functioning market that allows everyone to bring their value to it.

On mobile, we’re seeing marginal benefits as a result of some of the more recent changes. We have slightly better user experiences for our members on those platforms. We can be a little bit more communicative about how they can sign up for Netflix, which is great. Generally, I would say, consumers figure this stuff out, too. It’s been amazing how many people say, “Oh, this doesn’t work so I go to Netflix.com” and it all sort of figures itself out from there.

The power of Bridgerton marketing the service for you. I have one last question. It’s a total wild card. I’ve been dying to ask it the entire time. Do you keep track of what TVs your subscribers are using?

Of course.

Okay. So I have a theory about the Samsung Frame TV, which is the most popular TV. The reason people like this TV is not because it is a great TV; it’s because it looks good when it’s off. But do you see the popularity of the Frame TV as any sort of symbol for the future of television viewing? Because I see a TV that looks good when it’s off as being a sign that soon there might not be a TV. Because if you value the off state more than the on state, something is happening.

I would totally disagree with you on that. There’s a whole bunch of design considerations that go into the things that we use and how we interact with them in our daily lives, and some of those design considerations are about functionality. How does it show up when I’m actually using it? But a lot of the design considerations are aesthetic considerations: it’s an object in my home and I care how it looks even when it’s just static and sitting there. In some cases, we obviously choose these things based on no functional goal and just having something that looks amazing and beautiful.

Yeah. I’m just saying my radar about this has been spinning for a while.

I think you’re reading into it considerably.

I mean, I’m sitting in front of a Frame TV. There are three of them in this house, don’t get me wrong. We buy the TVs, but I think about them as a cultural object quite often. I figured I had one shot to ask you that question.

And you still watch them?

No, we watch the beautiful OLED. That’s what I’m saying. We hang up these TVs, and after a while, I can just be like, “Why are these even here?”

Well, then the key is that a TV manufacturer is going to be able to provide you both, right?

I think that’s what they would prefer. Alright, Greg, you’ve given me so much time. I’m going to let you get back to partying in France.

Thank you so much for being on Decoder.

Thank you. Appreciate it.