The Vergecast - How Planet Earth — and the Netflix homepage — get made

Episode Date: November 29, 2023

Today on the flagship podcast of multi-drone camera setups: 03:19 - The Verge's David Pierce chats with Planet Earth III producer Alex Walters and director Theo Webb about the gear used to make the latest nature documentary series. 30:13 - Netflix's Pat Flemming joins the show to discuss how Netflix figures out what to show when you open the app, and how to keep you coming back. 1:20:29 -The Verge's publisher Helen Havlak and editor-in-chief Nilay Patel join the show to answer this week's hotline question. Email us at [e-mail] or call us at 866-VERGE11, we love hearing from you. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Welcome to the Vergecast, the flagship podcast of multi-drone camera setups. I'm your friend David Pierce, and I am currently sitting here at my desk comparison shopping monitors. So I do this thing where I would say once every nine months or so, I get this idea in my head that what I need is a different number of screens of different size. Sometimes I'm like, I'm just going to be a laptop guy. I'm going to be more focused. I'm just going to sit here and look at this one small screen. It's going to keep me on task. It's going to be great.

That never works. So then I'm like, okay, I'm going to have a command station. I'm going to get a bunch of monitors. One of them's going to be vertical. One of them's going to be curved. One of them's going to be humongous. I'm going to get so much done. You're going to freak out. You don't even know. That doesn't work either. And then eventually I end up back at one monitor. Right now I have one 24 inch 4K monitor. I've had this for the last year or so. It works fine. But it's starting to feel kind of small. So what I'm going to do next is attempt to not swing one way or the other on this pendulum, but just add one more monitor, maybe 27 inches, maybe 24, everything's on sale, who knows. Anyway, we have an awesome show coming up for you today. Today turned out to be one of those episodes where we're kind of accidentally on a theme, and I would say the theme is how stuff appears on your television. We're going to talk to the folks who made Planet Earth, the documentary series from the BBC, all about the technology it takes to make that show and to get into the lives of animals

and the planet. Then we're going to talk to Pat Fleming, who is one of the heads of product at Netflix, all about recommendations and basically what it takes to show you exactly what you want to watch at exactly the right time. It's much harder and much higher stakes than you might think. All of that is coming up in just a second, but first I have to figure out if this Asus monitor I'm looking at is different from this other Asus monitor I'm looking at because I've been staring at these two pages for 10 minutes and I just realized they might be the same monitor. That's how this is going. This is the Vergecast. See in a sec. Support for the show comes from Retool. Too many companies run critical operations on duct taped

spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to, but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need. Prompt something like, build me a revenue dashboard on our Salesforce data. And Retool actually builds it on your company's data in your cloud with enterprise security built in. Go to Retool.com slash Verchcast. We all need to retool how we build software. What's up, y'all.

I'm Skyler Diggins, seven-time WNBA All-Star, Olympic gold medalist, and mom. And I'm Cassidy Hubbard, host and reporter for nearly 20 years covering the biggest names and stories in sports and mom. And this is Am Mom, a community for athletes, games, James of all kinds, dropping May 14th. Tap in with us. Welcome back. The thing you should know before we get into this segment is that I'm kind of obsessed with the TV series Planet Earth.

We can now show life on our planet in entirely new ways. Bring you closer to animals than ever before and reveal new wildlife dramas for the very first time. The truth is, I'll watch anything that's just like slow-moving drone shots of beautiful landscapes and then super tight shots of animals while someone talks in a cool accent about a whole part of the Earth I've never heard of before. And there are a lot of those shows out there. But Planet Earth does it better than everybody else. I mean, just like listen to this clip for a second from Planet Earth, too. He listens carefully to pinpoint his target.

It's moving. I mean, come on. That is the stuff right there. I love it. Anyway, the third Planet Earth series is out now. It's been seven years since season two aired, and now we get more Planet Earth. And I mean, this is not a spoiler and it's not surprising, but it's awesome. And it's actually a really fun snapshot of technology progress, too.

These folks spend years trying to access new places and get new kinds of shots and tell new stories about the world in different kinds of ways. And so much of that is actually about technology. So I called up two of the people who worked on Planet Earth 3, Theo Webb and Alex Walters, to get into the weeds a bit about the tech involved in making a series like this. I'm Theo Webb. I'm a producer director on the extremes episode of Planet Earth 3. My name's Alex Walters, and I'm an assistant producer on Planet Earth 3. Between them, they've basically seen how all of the sausage gets made.

I had this theory when I started watching Planet Earth 3, that the big difference is going to be how much less gear was involved in making every episode. Maybe they'd shoot everything on smart phones, and certainly the big cameras would be smaller and more maneuverable. Maybe you can shoot with like a phone, a DSLR, and a DJI drone, and boom, you're done. But then you watch the behind the scenes at the end of every episode, and no, it's still a ton of gear. There's one whole thing that's just a panning shot of all the batteries they had to charge, and it's terrifying to look at. And Theo said actually, if anything, it's more gear this time. I mean, look, 100% we need, we need this so much stuff.

We wanted to film macro. We wanted to kind of get as much moving imagery as possible, sliders, et cetera. We weren't quite sure what tool will be right. And actually, we took a lot of stuff, but almost all of it we used, actually, for at least one shot. Some of the times, yes, you take a bit of kit out to a location and, you know, it sits in a box and you're like,

well, that was a waste of time and effort. But on this occasion, actually, the poor people that carried it around the cave, They did not do it in vain. It was worthwhile. Like, there was so much reused. But the funny thing is, a lot of that behind the scenes segment, that was actually shot with iPhone. And we found that because there's that inbuilt software

that's processing the images, so much more clearly than, let's say, a camcorder, which is just basically built for outdoor use that we would normally take as a behind the scenes camera. The iPhone was processing it, and you could just get that live feedback of what you're filming. And it dealt with the noise and with the low light. well that that actually became the easiest thing. It meant you've got it in your pocket. And I dream of the day that we can just take that out as our main make your camera. So easy. Maybe by Planet Earth

4, when I guess we'll be on the iPhone 22, it'll be fully up for the task. Probably not, but we can hope. For now, though, making a show like this is still an incredible amount of work and an incredible amount of specialized equipment. Like the microphones that Alex and a crew use to capture these tiny things called treehoppers in an episode titled Forests. This tree hopper mother communicates with her young, encouraging them to stay together so that she can protect them. These things are the size of a pinhead, so they're not easy to find or capture at all. So what Alex and the other producers do in a situation like that is they talk to researchers

and scientists who know these animals really well and know where they'll be. and how to interact with them. They're actually, as a result, sometimes able to storyboard the shots they're going to get, knowing how these animals work. But to be fair, it really never seems to go to plan at all. I was talking to a scientist in America called Rex Kocroft, who is an absolute genius, and he spends his life listening in on conversations between animals, specifically treehoppers. So he was really key to this, and he helped us develop the technology to be able to listen in on them.

And then alongside that we had Javier, who was our specialist. He's a photographer, actually, and he was the one who brought this story to us about the tree hoppers interacting with this symbiotic relationship with bees. They're like bodyguards. These bees come in and protect the tree hoppers. So actually it was a combination of different scientists, and I spent a lot of time talking to them and planning in advance for the shoot. But the tree hoppers, it's inaudible to the human ear. So we used these Pizodisks that we would attach a microphone and we would put them on the branches

to be able to listen in on this world of these tiny little insects while they communicate to each other. They all have different language. There's over 3,000 species and they all communicate with different sounds and they mean different things. So some of them talk to each other about finding a mate. Some of them tell each other where food sources are.

Others will talk about where predators are, which is part of the sequence that we filmed. So they are absolutely fascinating. And there sounds are really bizarre. They sound like they're out of space or something. There are stories like this for basically every single frame and moment of planet Earth. Like there's this one shot in the extremes episode where we're inside this huge colony of butterflies, hundreds of thousands of them all hunkered down for the winter.

And if you disturb these butterflies, you're going to send thousands of them flying off, which A, ruins the shot. And B, also probably kills a lot of the butterflies. So you end up having to get creative trying to figure out how to get in there. The fun thing about this shoot is it relied on a very specific bit of technology, which is socks. Without socks, the whole shoe would have been ruined. So these butterflies, they can't really afford to lose too much energy over that winter period. Otherwise, they won't make it through.

So the crew had to be as quiet as possible, which literally meant taking off their shoes and walking around in socks. So they'd be setting up cable dollies on a tree without butterflies on, et cetera, you know, put one on it. And then they'd be kind of moving this cable dolly silently through. They'd have their socks on and they'd just be, they'd just be waiting. Meanwhile, it's not actually that warmer place. It looks kind of quite nice, warm and cozy. But it's actually quite a high altitude forest in Mexico. Drops below freezing easily at night, which is half the problems for the butterflies.

But the butterflies, they have this, you know, perfect little ecosystem in the canopy there, which keeps them nice and warm. But for the crew on the ground, that was less comfy. But sogs, who knew it? Some shots they know they can get. Animals have patterns. Big caves are probably still going to be there next year. One thing I've heard from a lot of nature filmmakers over the years is that if you just

show up and stay there, eventually most animals will get used to you and just carry on living their lives so you can get what you need as long as you're patient enough. That's why you see a lot of hides and camera traps. really a camera trap is actually a staple of planet Earth. You set up a camera in some slightly hidden but very central place, turn it on and just wait. And knowing where to put those cameras and actually how to place them and angle them and focus them is a real skill. Like when you're prepared to capture 8,000 hours of footage trying to catch a glimpse of a snow leopard in Mongolia. The guy who did that guy called Jake Davis, he's really specialized in these

camera traps. And he has this amazing ability to get into the head of the animal. And, you know, when you're setting up a camera trap, it's got to be the right height. If you set up, you know, the height of my head, you're never going to, you never film anything. It's got to be down at the right eye line. You know, you want to make it cinematic as well as functional. So you get it down, you get it set up, and then you have to crawl through the frame, right, pretending you're a snow leopard to kind of mimic the action. But also, and this is the thing you hear from nature filmmakers over and over is sometimes you just get lucky. But actually, my favorite shot in that sequence was one actually designed by the

snow leopard themselves. The camera trap was placed down, but the cat decided, I didn't really like the position of that. So they, like, rubbed themselves on it. They'd carried it. They'd moved it. And they just put it down, right? And then a couple of days later, I think it was.

The same cat walked down right in front of it. But what the frame had done was because it had been put a slight angle but also points to the ground. There was a super shallow depth of field, and the cat just placed his foot right in the middle of that depth of field. And it's like the most perfect snow leopard shot you could have imagined. But, you know, he can't take credit for that because it was a snow leopard that did it. There are a lot of moments in making something like this where you just prepare as best you can and then wait as long as possible. But sometimes you're literally only going to get one shot.

Like there's this one moment in the forest episode where a tree almost 200 feet tall goes crashing down in the middle of the, a forest. You know, if a tree falls in the forest and you're not there to film it, does it, no, mind, that's a terrible joke. Anyway, it's a really hard thing to film, it turns out. The tree fall was really complicated and took a huge amount of planning, as you can imagine, and a big risk assessment that I had to write. But yeah, so that one, we had quite a big crew. We had three drone operators who were flying at the same time. So two of the drones were static to keep it, you know, simple. And so they, you know, no disturbances within each other's shots. and then one that was moving and reacting with the tree as it fell.

So that's like a sweeping shot that pulls back. And then we had ground cameras. So myself and Yoris van Alfa and another camera operator, we were operating the ground cameras. We had six of them placed. We knew this tree was around 60 metres high. We'd measured it with a drone and sort of seen how high it was with that.

And then we'd measured on the ground how far we need the cameras to be a way to get this whoosh of all the leaves and everything breaking, coming up towards the camera, which is one of the shots you'll see in the forest episode, but not having the cameras completely smashed to Smithereens, which was a real risk. Some of the cameras were GoPro's and we were like, you know, if we lose a GoPro, then it's not the end of the world.

But, you know, some of them are these red cameras that are, you know, how much they're very expensive cameras with very expensive lenses. So they were based more at the base of the tree sort of where we knew it definitely wasn't going to fall in that direction. We then worked with some locals who allowed us to film them. felling, they selectively felled the tree. They told us we know that when we take this route off, it's going to fall in that direction. So we knew the direction it was going to fall. So there was a lot of planning, a lot of talking with the local people, and that was key. And none of the cameras

got smashed, which was absolutely amazing. I could do this forever, honestly. There are fun technology stories behind basically everything in the show. But let's just do two more here. First, there's this section of the extremes episode filmed in a cave in Vietnam that is thought to be the world's largest cave. It's called Hangsandong. The crew spent more than two weeks in this cave, filming it in all sorts of different ways.

But the most remarkable thing, I think, that they did is a two-drone setup to try and capture the true size and scope of this place. We had a team of fourth of cinematographers, but we chose one in particular because he'd been down inside Hansung-Dong cave. He's been down like almost 15 times or something like that. So he knows things.

the cave really well. He shot a lot for himself, but he'd never done what we wanted to do, which is take sort of cinema cameras down there and show the cave off in a different way. He'd also never spent, I don't think, more than three or four days there. So the fact that we were going to spend 18 days underground in that cave sort of blew his mind and also made him incredibly excited at the same time. He was like, oh my God, think about what I can do that here. So first of all, what he thought about was how to light it, right? Because as you say, say the paradox of filming underground is that unless you show light, there is nothing to see. And yet, you want to keep some of it dark to keep it like mysterious, but also it helps with

that sense of scale. If you like the whole thing, you lose that scale, you lose the texture, you lose everything. So what we decided to do, and with his assistance, we built these kind of four panel arrays of custom LEDs, which are super bright LED panel. And they have a cone attached to them and you can bolt them together, which is how we created this sort of four panel right we tucked that underneath this heavy lift drone so a dd i matri 600 which is you know 10 kilos heavy about a meter wide it's a huge beast we'd put that up in the air and the 15 minute timer would start ticking so the lights out of their own battery source and they were the weak link if you like they ran out about 15 minutes at the same time we'd send up a dj i mavc 3 pro and and that would

be our filming camera i should add the other kind of fun thing about filming inside a cave right is that obviously you've got no GPS on the drones, right? GPS is non-functional. And when you take that away from a lot of consumer drones, they freak out. They have this in-built cap, right? They only go 10 metres high. So you can't do anything crazy with them.

So we actually had to work with DJI to take those caps off for this very specific case. It didn't work with the DJI Mavic 3. So we had to hack it in the cave by doing these crazy flight stunts. So we'd worked out that if you used it to, of the cave, it would baffle the drone and it would think that it was above or below 10 meters, whatever. Anyway, so we'd fly over this bit of topography. It would think, oh, it's cool. I'm still within my 10 meter zone. Anyway, we could fly it 100 meters high, which is crazy in itself because we're underground in this, well, 100 meter plus cavern, right? Huge thing. So we'd take our fly around with that. And meanwhile, whilst we're filming with this Mavik, we would swing the lighting drone around in arcs, around the object that we were trying to film. What it would do, it was a move the pool of light to both reveal and hide certain areas of the cave, which would also enhance the shadows and the textures of the cave features. And it was a way that we could

introduce scale into the cave and also just show this underworld like has never been shown before. One cool thing, by the way, about that drone footage and about a lot of the stuff in the cave is that you see people in the shots a lot. The crew members are actually in this series much more than in previous planet Earth, which Alex and Theo both said was a deliberate choice. They both wanted to show scale, but also to remind people that humans are part of and change every single one of these stories. It works really well in actually both ways and is pretty wild in particular in the cave. So that's the big stuff.

And then all the way on the other end of the spectrum, there are two sequences in the episodes we're talking about where we actually get to see inside of a nest. One is inside a termite mound in the grasslands of Australia where parrots are hatching their eggs, and another with oriental pied hornbills burying into a tree to make a nest in Borneo. Once settled inside, the female does something truly bizarre. She pulls out her flight feathers. She is not going to need them because she will be staying in here for quite some time. Both times you're in this teeny tiny space that is somehow well lit, shot in high resolution, and doesn't drive the animals totally insane.

It's wild, and it's really hard to put together. These parrots, these golden-shouldered parrots, like beautiful. They live in these grasslands in Northern Australia. They're super rare. You've got to be super sensitive, right? You know, one failed nest equals a whole generation for that family, not pledging and replenishing the population. So we had to work with the landowner who has loved, who has studied, who has looked off and cared for her land and these parrots for like decades. She works with this amazing scientist Steve Murphy who allowed us to make a small hole in the back of the tomite mound once the chicks had hatched.

So they're much more relaxed. So they're feeding, they're feeding. So when the mum and dad are away, we need this small hole in. And using a probe lens, this lower probe lens, we are. able to get it through and it's super wide lens so you can see the whole nest. The nest is probably only the size of like a, I don't know, a great fruit, right? So it's actually pretty small. So you're getting this really intimate view in.

And then, of course, to light it, it's super important because if you just rely on the lighting from the hole, you just, everything's really backlit and dark and bright and it's awful. So you've got to make another tiny hole for the tiny little LED lights that we could shine through at creative angle and light it inside. And then we'd run a cable away to a hide, 20, 30 meters away. And then it would be a remote trigger. So once it's there, it was all set. But we wanted to film a different stage.

So we'd film for a day, go, yeah, pretty much we've got enough kind of feeding moments. We'd withdraw the camera. We thought, okay, two weeks later, we'll come back and stick it in and get the next part of the live stage. By which time, by which time, the termites, they're so industrious, right? Over that short period of time, they'd sealed back that hole. So they're like, oh, God. So I had to make that hole again. do this whole thing. But anyway, once we learned how to do that, the camera up Mitch decided to

put a cork in it. I think actually, just a plug, which meant the termite sealed around the plug. So it was with air tie and everything. But it meant they could just pop it out and put the lens back in. It was a super privilege look into these super rare parrots. And I'm really pleased to say that they all fledged absolutely beautifully. It was joyous to see them all fly away. But Alex, your story in the hornbill is amazing, right? Yeah. So we use a very similar setup. up to Theo's. We've actually used the same lighting and the same lens. So we sent all of that to Borneo for

our team in Borneo to use. And what they had to do, because it's such a long process, this nesting period is very long. First of all, we had to guess which nests that these hornbills would use because we didn't want to disturb them. So we chose five nests.

And the guys out there worked with scientists who studies the hornbills. So they went out and they set up tree platforms attached to the side of the trees. for the camera to go on eventually if that female chose that nest hole. And then they waited for them to sort of, for their behavior to start happening where they could see which nest they would be choosing. And then just like the termite mounts with Theo, we drilled very, very tiny holes and used

this almost endoscopic lower probe to sort of film inside the nest. So we've inserted, we've made a hole for that. And then we also made, I think, three holes for lighting. Because, you know, these endoscopic cameras, they need a lot of light. to get the light into these very dark cavities. And then what we had to do is, again, the cameras were tethered and the lights were tethered. So once she decided which nest she was in, we had to wait sort of a week or so for her to settle in. And then in the night time, well, in the very early morning, they then unplugged the hole,

which they'd plugged up with clay, because obviously it's a living tree and you don't want to add any bacteria or anything. So they've used clay. and then they would gently put in the prey blends and then they would also put in the lights and then Will Foster Grundy, our camera operator, would also gradually turn up the lights as the daylight started coming up. They've got quite big eyes, these birds,

they're very sensitive to their surroundings. And so there were times. She definitely muddied, you know, she's expert, you know, muddying up herself into this tree hole in creating this little gap for her to live in. but she would then muddy up the lights and the camera. It was frustrating, but they worked so delicately, and they managed to get the full behaviour of her,

and yeah, she accepted it eventually. And we got this absolutely incredible, intimate view of Inside the Nest, which, you know, without that, the story wouldn't have been half as good, you know, to be able to see her perspective, you know, almost as if she's in this prison. And it's just, yeah, I think it made for an excellent sequence. Ultimately, in talking to Theo and Alex, it didn't sound like there was much they couldn't do, technologically speaking. It's all a lot of work and a lot of gear.

It takes forever, and this show is expensive to make. But most of it is doable. Still, though, Alex and Theo said they both have some ideas about how tech could make life easier for, say, planet Earth four in a few years and even beyond that. I think personally, I would like the camera equipment to be a bit lighter. because, you know, it is getting smaller, but actually we end up with just as much clover. We have just as much here, even though it's becoming smarter because, you know, you have this specialized bit of cable that's used for this camera, but not for that camera. So you end up with loads of stuff.

So it would be nice if things actually just became a bit smaller and a bit lighter. But then sometimes things like tripods, you need them to be heavy for that stability. But for me, that would be great. But then for the making of, as Theo was saying earlier, having, an iPhone to film that has changed and I think that will continue changing. Okay, I've got two. One is drones with telephoto capability. That will change everything because as soon as you can start getting stabilization good

enough to get like 10 times zooms on plus, you know, you're starting to be able to film really, really cool behavior in a really cinematic way. You'll be able to get incredible low angle shots which are moving and yet compressed because you're at a 300-mill or whatever. That would be game-changing. The miniaturization of drones is a part of that as well, and that will definitely kind of fill that space, and it will be great.

My rogue prediction, maybe not for six years, but maybe slightly longer, is utilizing space imagery more. We had a great series a few years ago called Earth from Space, which utilized space imagery to get images of something happening on the ground at pretty high, super-high resolution,

which would then merge with drones, filming that same event, and you'd have this crazy interface, right, from outer space down and you could fly effectively straight into the action from space. And it was brilliant. It was super cool. I think that was way more innovative than the kind of time should allow for. And I think that imagery from space, the more resolution you can get, the more focus will allow us to do so much more, both in a filmmaking perspective, but also from a science perspective,

counting animals. You know, that's super important. What they do with satellite imagery now for like counting penguin colonies is crazy. You know, they look for shit, basically. They look for penguin shit on the ice. And they go, oh, look, there's a whole new colony there. We have no idea there was that many more penguins or that many less penguins. And they can tell how close they are to being like, well, wiped off by kind of by ice loss, basically.

And so that kind of technology will feed into science and it will feed into what we do as wildlife filmmakers so much more. And I'm super excited about that. And then filming on Mars. You heard it here first, friends. Planet Earth 4, the Mars edition. All right, we have to take a break. And then we are going to come back and talk about TV algorithms and how to get recommendations right.

We'll be right back. Support for this show comes from Shopify. Starting something new isn't just hard. It can be really scary, too. So much work goes into this thing that you're not entirely sure will even work. But here's a better thought. What if it did all work? What if your instincts were actually right all along?

Shopify wants to help you get there. They're the commerce platform behind millions of businesses worldwide and nearly 10% of all e-commerce in the U.S., from established brands like Allbirds and Heinz, to companies just getting started. Their design tools make it simple to create the exact online presence you're envisioning, with hundreds of ready-to-use templates available. And with built-in marketing tools,

you can launch full email and social campaigns in just a few clicks. so you can connect with customers wherever they are. It's time to turn those what-ifs into with Shopify today. You can sign up for your $1 per month trial today at Shopify.com slash vergecast. You can go to Shopify.com slash vergecast. That's Shopify.com slash vergecast.

Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers. That's where LinkedIn HiringPro comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process

from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screening. It's updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn. LinkedIn.com slash track.

Terms and conditions apply. Welcome back. We're going to stay on the theme of, like I said, how stuff appears on your television, but I want to take a look at it in a really different way. For years, I have been trying to figure out how streaming apps are supposed to work. The way they do work, where you open the app and you're presented with like 100 million tiles of shows and movies you might like, and then you spend forever scrolling through them and never find anything to watch, and you eventually just,

just give up and throw your TV into a river and move on with your life, that can't possibly be the best way. But what is the best way? A bunch of years ago, I asked somebody at Hulu what the goal was, like the North Star for how they think about this. And he said it was just like a play button. You open the app and it should just know what you want to watch and maybe even start playing it for you.

I'm not sure that's the answer either. So what is the answer? There are, in the world, few people better equipped to know the answer to that question than Pat Fleming. I am a senior director of product management here, so I lead the member experience PM team. Think of that as the group responsible for how profiles work, how the core browsing discovery experience works, how you watch something great tonight or play something great if you're playing games across iOS and Android, and that's across TV devices, the website, iOS and

Android phones and tablets. In case you're counting, that's a lot of jobs. But one of them is very much to figure out what should happen when I open the app. Netflix is famous for all the time it spends thinking about these recommendation systems. Do you remember a bunch of years ago when it offered a million dollars to any developer that could improve the system in a meaningful way? Even now, picture the app. There's that huge banner that shows up whenever you open Netflix.

And then below it, there's all those endless rows of increasingly specific genre. and what's popular and what you've watched before, and there are a million ways big and small that Netflix is choosing and influencing what you see and in what order. As a result, I guess, more than most companies, I found myself often sitting there wondering why Netflix is showing me what it's showing me. So that's what Pat and I talked about. I started by asking him that big picture question I've been asking people for years, the one that I asked that guy at Hulu all those years ago.

What is the best case scenario here? When I open Netflix, is the goal to show me a million options? Is it to show me three options? 300 options? Is it to get so good at recommendations that it does get everything right immediately every time and just starts playing the exact perfect thing? Is that the plan? That would be so cool, but it's quite hard.

So people have such diverse tastes and interest. So it's not just that you might love sports documentaries and dramas and baking shows, for example. It's also that tonight you might be interested in watching Beckham, but tomorrow you might be interested in watching the Crown. And it's quite difficult for us to know that when you sit down on the couch. So by definition, that hit the button on your remote and Beckham or the Crown starts playing. Like, that's a hard, that's a hard game for us to win at at very high frequency. And so I think the other thing to say is people like to browse. It's not that people don't like browsing.

It's the people don't like bad browsing experiences. So if we can deliver a great discovery experience, that's the word we've. would use internally and make that browsing experience feel great, feel the right amount fast, the right amount compelling, give you the right amount of options, then you're satisfied when you land on one that's perfect for you. The other thing to know is that you might not know what you love. So let's take those genre examples. So let's say I'm into sports docs and dramas. And let's say many folks who use Netflix are into sports docs, dramas, and baking shows. Let's say I haven't tried a baking show yet. The beauty of personalization is that it's

nearly a certainty that we would then try to recommend baking shows to me because there's lots of folks out there for whatever reason who happen to share many of the same similar taste and preferences I do and this one that I haven't demonstrated yet, but it's plausible that I might. And so we would make that recommendation, hopefully you find it. That's hard for us to do if we're just going to go for that most likely thing on any given night. Right. Yeah, it does seem like, A, if you were to do that, you would end up, just playing for me like the safest thing where it's like, If I just booted it up and you just showed me community every single time I turn on Netflix, I'd have an okay time with that, right?

Like, if it was just a community machine for me, like, I'm very rarely going to be mad at that. But there's also no way to ever find me anything new in that world. And it also seems like one theory I have that I can't prove because no one has ever tried it, but I hope somebody does so I can see if it's true, is that even if you showed me the right thing, like if I was deciding, do I want to watch The Crown or do I don't want to watch Beckham, I think I would reflexively not want to watch whatever you showed me. There's just something about being a sort of force-fed that makes people not want to do it. I kind of can't describe it, but there's something that feels bad about the UI of being told what I'm going to watch, even when it's the right thing.

I like that theory. I think it's very plausible. I think we probably hear that in research from time to time that giving me options and then the right amount of what we would call evidence, by the way, to help me make an informed decision feels more satisfying than us just pushing something to you. And that is the truth of it. members and our incentives are completely aligned. We want to find you something awesome to watch and also make it delightful for you to get there. The simplicity of the Netflix business model is its subscription. And so towards the end of the month, if you find something, the odds are you're going to subscribe again tomorrow for the next month. And so that's what we're after. And whatever way works best for you, we want to make that happen. So to the point of making that decision feel good to you. So to your question, I'm like, well, would it be better just to kind of push the

one thing or do people like the choices? Something you might not know is that the way the choices are presented to you or in any particular country might be a little bit different. It takes squid game, for example. We would try a variety of different ways to position squid game to you. And so we would have different sets of static artwork, slightly different synopsies, slightly different ways to promote via video. So like a trailer, for example. And then we would try those different ways, perhaps in different countries, to different groups, and they would all be reflective of what Squid Game is about, but they would be oriented to try to find what would be best for you to understand what a particular title is about and then make a very informed decision. So that's another element of the personalization,

again, to help you be super satisfied with whatever you click play on so that we're doing our best job we can being a matchmaker for you. That's really interesting. So yeah, let's actually talk about Squid Game, because I think the question I've always had about Netflix, right, is like how something goes from sort of appearing on Netflix to suddenly being everywhere. And I think we talk about it, like it's a certain kind of either accidental magic or like brute corporate force from Netflix. And I think the truth is probably somewhere in the middle. But I do think the last sort of big picture question I have, and then I want to kind of tell this good game story, one of the things I've always wondered about Netflix is it seems to me like you're talking about being a subscription business.

all that really matters is that I find something to watch. As long as I'm satisfied every time I open the app, everybody wins, odds are I'm going to keep subscribing. And it sounds like that's the case. So what is the upside for Netflix and having these kind of cultural moments? We talk a lot about sort of the death of the show everybody's watching. If you're Netflix, do you care?

As long as I'm finding something I like, what does it matter if it's the same thing that everybody else is watching? I mean, isn't it so fun when something like Squid Game or Wednesday or La Casse de Papel happens. I mean, it's fun. Is it required? No, probably not.

That's the beauty of personalization. So when you open the app and when I open the app, we're going to see completely different home experiences because we probably have completely different taste and preferences. But when it happens, it sure feels magical. And so take Squid Game as the example. The beginning of any of these sort of large phenomenons is, phenomenon is an incredible story.

So Squid Game, very Korean. You can define it as violent. as a commentary on society. It's funny at times. It's certainly dark. It's exciting. It's thrilling. It's a game show.

It's all these different things packed together. Created distinctly for a Korean audience. There was no illusions at the outset that that would necessarily be a huge global hit. But what happens is that starts to play incredibly well in Korea. So we start with an audience that is likely to be quite highly qualified in terms of delivering impressions, recommending Squid Game for that group. And then because Squid Game and other Netflix originals are available globally. We've got dubs and subtitles in often over the 30 languages.

So in the case of Squid Game, for Brazilian Portuguese and for Latin American Spanish, it started to take off in Latin America. And from there, it's sort of difficult to know the causality, because lots of things are happening outside of Netflix. So people started talking about Squid Game, and this became kind of a global thing on social. And then people are searching for Squid Game on Netflix, and that provides us another signal to further change how we're recommending Squid Game. And from there, the fire just continues to burn. And all of a sudden, everywhere, you're hearing about Squid Game, and you open Netflix and you find it happens to be there.

And so that's a pretty good example of how it would work. I think Lacoste de Papel would probably be another example. Our goal is not necessarily to have titles like Squid Game be the one thing everyone is talking about globally. We want to make titles that are authentically local and can potentially find great audiences all around the world, and that's incredible when they do. It's not always the case. In fact, it's more rare than not, but it's super exciting when it does happen. Okay, so you just breeze through a bunch of, I think, really interesting, like, personalization

and discovery things. So take the squid game thing and go, like, rewind all the way back to the beginning. So, like, not the why we think the show is interesting thing. Like, that I get. You make the show because you think it's going to be cool. At the very beginning, it seems like with something like that, like, I'm assuming sort of from your perch, the job is to figure out, like, who are we confident this is going to be interesting too. When you're at the point nobody has seen it, you can either just like beat people

over the head with it and make them see it, or you can try and guess who you think is most likely to like it and show it to them first. Is that, am I thinking about that the right way? Totally the right way. So like two good examples to give you on this one. One, let's keep it with Squid Game. You can go find out about titles that are going to launch on Netflix in an area on Netflix, either on mobile across iOS and Android or on TV. You can see a set of coming soon titles, right? Is that a popular thing? Like, do people interact with that a lot? People love watching trailers and finding out about what's coming, whether it's with us or with other services. So I'd say generally, yes, of course, I think people love to sort of anticipate what will be fun for them next

on Netflix. And so you can tell us directly, I'm interested in this thing. So you would, you would literally click, remind me. And then we would remind you, we would follow up on that obligation. So we'd send you a push notification or an email or remind you in the app when you open up Netflix. But in addition, that gives us this very explicit signal of the kinds of members who are likely to be interested. Now, it's not perfect because you could say, you know, remind me for, let's say, you know, the next season of the Crown, but you have no intention of watching it. You haven't watched the prior seasons. It was like an accident or like you think your partner might want to watch with you. So it's not a perfect correlation that when you hit, remind me,

you will definitely want to watch. But it's a pretty good signal for us to work from. And then we would use that to extrapolate if people like David, are hitting, remind me at very high rates and people like David watch these kinds of things, then we would probably expand that initial audience to reflect people like you, whether or not, again, people like you based on what they watch, we would then recommend. And then from there, it just builds. And the beauty, again, is that it's quite pure and that all we really care about is being a great matchmaker. We worship at the altar of customer satisfaction. So for us to deliver a bad recommendation is for us to have missed out on delivering you a great recommendation.

So at the beginning, we know very little. And so the other example I would give you here is if you're a brand new member to Netflix, we know almost nothing about you, right? We know where you are, the device, whether it's a phone or TV or laptop that you signed up on. And then all you have to do is tell us just a very little bit about your interest. So we ask you, tell us a few titles that you're interested in. And that is the initial spark for personalization where we'll say, okay, based on those expressed stated taste of preferences,

Here are the kinds of things that folks who watch those things also like. Right. And it seems like I was reading through the guidelines to the recommendation system that are on your website. And it seems like basically everything about the system is either metadata about the thing itself or your viewing activity. And that there's almost nothing else that you're thinking about in that. Like I expected my age and zip code and ethnicity and, I don't know, net worth, whatever. Like there are a million signals that we've been trained by the internet to believe are super. relevant and targeting to us. And at least from what the stuff I was reading says, you don't

care about most of that. That's the beauty of it, is that our goal is just to match you with something you're going to love watching tonight. And the best input for what you might love watching tonight or tomorrow or next week is what you've watched or spent time with before. And those other signals might be additive, but the thing that really matters is what have you demonstrated that you're really interested in via your behavior? And that's, you're a lot of you. And that's, That's sort of the beauty of it. So yes, as you point out, you could think of titles as being broken down into these constituent components, this metadata. Is it a movie or a film? What's its originating country? How long is it? What language is it available in? genre, perhaps some

tagging about attributes of the title. And those things all help to sort of round out what a title is like. And then you match that against member taste and preferences. And again, take that sports, docs, drama, baking shows. If you watch any two of those things and, lots of people watch all three of those things, then it's a pretty good chance that that third genre might be interesting to you. And so some point in time, you'll discover some genre that what wouldn't otherwise have been in your consciousness and try it on Netflix. Now, the thing that we also do is inject some randomness into these recommendations. Because again, to the point of tonight when you sit down on the couch, I don't know what mood you're in necessarily or who's on the

couch with you. And it could be just something different tonight would be better for you. And in that case, we inject just a little bit of serendipity because we know it's quite hard for us to get it right every time. And that also helps continue to broaden what could be recommended to you. Yeah, what is the right balance there? I think about like the great innovation of TikTok was that it made it really low stakes for TikTok to be wrong, which, because it's so easy to scroll, it's so easy to just flip past something you don't like, that TikTok can just show you stuff. Everything in there is just pure randomness. And you don't really penalize it because it's so easy to scroll. I don't think you have quite the same permission because by the time I'm like 20 minutes

into a show and decide that I hate it, that's a much bigger commitment than one and a half seconds on a TikTok thing. So I would think there's less room to just sort of throw something at me just on the off chance I might like it to see how I react. How do you think about that? Like you could make the safe recommendation. You could just throw spaghetti at the wall and see what sticks. But neither of those feels exactly right for Netflix. We're definitely trying to not throw spaghetti at the So we're trying to make recommendations. Think about it as we're doing our best to predict what might be interesting to you based on everything that's known about your taste and preferences before you open Netflix tonight.

But if you follow that all the way out, you just show me community every time, right? And then we're trying to help you find, always have something great to watch. So yes, to your point like that, you know, maybe you might generalize that community challenge. And for me, it might be like, let's see, what could I watch your pre-9? I mean, Seinfeld would be fun for me most nights, right? Yeah, I'm not mad at community. I'm just like, I'm just only going to, like, yeah. Or it's like, I think about the one that happens to us all the time as we watch one trashy

and I don't use that as an insult reality show. And then it's just infinite trashy reality shows. And that actually makes a lot of sense, but at the same time, there's a finite amount of trashy reality TV that I actually want to watch. And so, but then to just show up and be like, oh, you like trashy reality TV, would you like the Great British Baking Show? That seems like a hard leap to make. but when you eventually kind of have to make.

I think customers might say they would watch community every night, but when we, as I say, add a little bit of serendipity, what we find is customers are more satisfied when they have a little bit more breadth in terms of what they watch. And I think if intuitively when you run that by members and you talk to them and you ask qualitatively, like how does it feel when you've been watching a greater breadth of things versus more narrowly, say only a couple of genres,

I think intuitively it makes sense. And I think qualitatively, you would hear that from people as well if you just ask them. And so what we find in the data as well is that when we diversify slightly beyond your very narrow, tried and true, David always loves community types of recommendations. We find that you'll eventually try one or multiple. And by the way, those will be a great fit. And that's sort of the fun of it is that especially when we're introducing new stories to the world, like there's never been something like Squid Game before. no one would say, oh, I was hoping I could watch Squid Game tonight, but you find it and you get into it and your loss, whether it's for a handful of hours tonight or more than a handful of hours or over the next couple of weeks, that's kind of the magic of great storytelling. Yeah, so that's part of the Squid Game story that I think is really interesting. And you're talking about this and, you know, it starts in Korea. It's doing super well. It's doing, it sounds like better than expected, even in the place where you expected it to do very well. What happens next? Like, is there a meeting where somebody's like people love this show? This might be huge. We have to.

to like turn up the randomness knob on the wall and start showing it to more people. Like what, how do you take it from, it's doing well among the people we thought it would do well among to let's show this to people who have not necessarily shown us they would like something like this, but the early signals say that they might. The beauty of this approach to personalization

is no knobs, no buttons, no nothing needed to happen. I'm convinced there's a knob. You can't convince me there's not. a knob that just, it's the viewer's knob, that you're just like, show this to everybody. Sometimes it appears that way, but that's what's so sort of magical about it. And as I was saying, in the case of Squid Game, it's a little difficult to know for sure how much the conversation outside of Netflix drove what we saw sort of some of the behavior.

So as soon as people started to talk about Squid Game, then we start to see, as I say, some signals of increased demand that we wouldn't otherwise be able to observe. And so the- Meaning like people searching for it more. Exactly. A great example that would be searching. So if people are most members at that point, probably wouldn't, especially English primary members of folks in the U.S., for example, wouldn't necessarily have spent lots of time watching

original language Korean content before they were introduced to Squid Game. And so it would not necessarily have been something that would have been high up there list of, say, stated preferences or stated, I'm excited to watch a dark, gritty game show-oriented thriller series from Korea. But once people started talking about it, and we started to see some, again, for example, that search activity, that helps then reinforce for the algorithms that it's time to maybe open up the audience a little bit. And as I say, that's automatic in nature. There's no need for dials or buttons. And what we're doing all the time is trying to tune more generally. So we would like to always be improving the way we're making recommendations. So, for example, we have a view that maybe not necessarily every.

say half an hour you spend with Netflix is worth the same to you as a customer. So we would like to be able to learn a little bit more and be able to recommend, say, a half hour episode of something that you would love tonight versus a half hour of something you would be just okay with. So community for you might be a good example. Like you'd be satisfied with community tonight, but you would love if you found something new and helping to tune finding that what you might love and be able to make those recommendations with better precision as something that we're always after. Got it. Okay. So is there really, there's no, human interaction in this process where like something like squid game starts because one of the

funniest things about squid game to me is you could look at it at some point early in the process and think oh the algorithm has gone completely haywire here there's no way this many people in this many places want to watch this specific show that we never in a million years would have put on this particular trajectory and on the other hand it's obviously like a smashing success for the system that it essentially, I'm assuming at some point basically told every person on planet Earth, you should probably watch Squid Game. Everybody else is watching it and seems to like it. So is there not a moment where it's like jumping up the graph and you have to have a meeting where you're like, what is happening here? Is this what's supposed to be going on?

This is the beauty of it. It wasn't even necessary for Squid Game. It became big on its own. I think suits in the U.S. is another good example where you maybe wouldn't have necessarily expected in advance how big suits could be at this point in time. And it's, been huge. And so that's kind of the beauty of it, where there's no need to be in the background pulling a separate lever. It's just, again, try to think about from how do we be the best possible matchmaker between all the possible options you could have on Netflix and other services you might subscribe to and make the best most compelling recommendations and the easiest, most hopefully fun, to the extent browsing and finding something to watch tonight can be fun way

so that when you do settle in on what you're going to watch, you hit play. And it's like, like, ah, this was perfect for me. I would assume, though, that at some point, kind of popularity begets popularity, that, like, the bigger something gets, the more comfortable the system gets, recommending it more widely to people maybe even further out of sort of the traditional strike zone of a show like that. Well, I think, again, take, we can just focus on Squid Game there as an example where, yes, if you're thinking about that randomness or serendipity, imagine you're a member who's, again, never watched any Korean language content before, we can more confidently make that recommendation

to you, the more popular something is, especially among folks who are also members of Netflix who watch similar things to you. So yes, I'd say that's fair. Okay. And is there, is there any upside to having kind of human or editorial or curatorial input here? I mean, I think about, like, we talk to a lot of the folks at music services who talk a lot about, you have this giant corpus of stuff. We can do some really interesting personalization, but also, maybe what you want to know is what a person you think is smart likes, or we can recommend, I don't know, like, Halloween is a good example, right? It was just Halloween.

Let's put together a thing that is like some of our favorite old school Halloween movies. People love that kind of stuff. And if you do that at Netflix, you do it less obviously than some others. Is that kind of just not something you think is necessary at your scale? We do, from time to time, do things like that where we're offering collections that are, representative of a moment of time. So Halloween would be, you know, that would be a very good example or other collections seasonally. So those kinds of collections do exist on Netflix and you can either seek them out yourself or from time to time you might find them those kinds of recommendations

promoted either on TV or on other platforms. You should think of those collections as still personalized in nature. So imagine you've had, you say, 100 candidates for some kind of a seasonal collection. We would still want to present titles to you in a personalized way. So whether that's what order they would be presented in or how they're positioned to you, as I was mentioning earlier, we've got an incredible creative production team that's thinking about what are all the possible ways you could represent a title, whether that's with art or the synopsis or video promotional assets. And we would still want to do that in a way that positions any of those particular titles within a, say, more curated seasonal set in a way that would really hopefully resonate with you

individually. Yeah, one of my favorite things to do on Netflix is scroll around and try to guess why the system picked that particular poster to show me for any given thing, because I know there's a bunch of options. It shows them in lots of different ways. And I saw one that was just like, I forget what movie it was, but it was a movie in which Tom Cruise was like the fourth most important character in the movie. But the poster I saw was just Tom Cruise.

And I was like, you know what? Like, you're not wrong, Netflix. But I wonder what it says about me as a person that it was like, you know what David wants to watch this movie is Tom Cruise? You might be a huge Tom Cruise fan. You might have just revealed that. Oh, yeah, it's not wrong. To be clear, it's not wrong at all.

But it's, it was, yeah, it's just very funny thing. And I think, actually, to that point, one of the things that I think has always been tricky about personalization over the years is pairing that with transparency. Like, I think there's an interesting thing that we found, and this is especially true, I think, on social media, where we have these ranked feeds and people are like, why am I being shown this thing, this way, in this order?

And there's something about kind of understanding the order that is being presented to you that actually is really comforting and makes people feel better. And to some extent on Netflix, like you've done a bunch of work, it seems like making even just like the rows are fairly self-explanatory, like here's what's trending, here's what's this category.

But do you think about trying to sort of telegraph the reasons behind what you're showing people as you go? I love this question because I think we've got opportunity to do better here. So we would call all the ways that we would position a recommendation or set of recommendations to you evidence for why something might be compelling to you, right? And so explaining this,

that evidence, there's good thinking behind that that is helpful for folks. So why is something presented to me? So I think we've made some good progress there on broadening the candidate set of reasons why something could potentially be a good fit for you. So I've hit on a few of those, whether it's the art being tailored to you or the synopsis or other individual pieces of evidence, for example, the cast or other folks involved in bringing that title to life and how that might resonate with you. But I think we can do even more, which is to say make the connection between here's really specifically why we think this might be a good fit based on whatever the meaningful to you reason might be, whether it's, for example, in the Tom Cruise case,

you've binged everything we have on Tom Cruise and this is why we think this would be awesome for you, or it looks like you're ready to break out of your community rut. Here's what folks love. Here's Tom Cruise. Here's Tom Cruise, exactly. When you want to break out of your community rut, Tom Cruise is the solve. I tell people that all the time. That makes sense. And I think it's another a really interesting UI question over time. One of the things I love talking to people about on Netflix is the thumbs up, thumbs down system. I don't know if this is true.

I haven't quizzed random strangers on this in a while. But for a long time, I think the thing where it would say, you know, it would have the green thumbs up and it would be like 89% match. I think most people would read that as a sort of rotten tomato score that's like most people like this thing. That's not what it is, is my understanding. It is something closer to like the likelihood that you're going to like this thing. but it's just a really fascinating thing of trying to telegraph.

You have all this information about what I'm likely to like and why. And to some extent, all of that is useful for me to see as I'm browsing. Because by the time I've been doing this for a while, I have a fair amount of faith and Netflix's ability to show me something I'm going to like. But the more you can tell me about it, the faster you can help me make those decisions. But A, you don't want to overwhelm people, I wouldn't think. And B, it just gets confusing at some point. There's just only so much you can show me, especially if it's like,

you liked Beckham and the Crown, we think you'll love the Great British Baking Show. People like, what the hell are you talking about? So, like, I wonder even just from a UI perspective, how you think about putting that stuff in front of people. Well, you're absolutely right that you've got to make hard trade-offs, and that's a hard design challenge when you're constrained in the UI. And so for the many tens of possible reasons why a title could be a great fit for you, if you think about this in the weeds a little bit on how someone's actually experiencing Netflix, when you're browsing, you've got that first glimpse of a title and you're somewhat more restricted

and what can be shown to you there. We want to try to present the evidence that is most compelling individually for you. So imagine again, there's tens of possible options say we'd want to hopefully think about how we might rank those and then present the ones that are most compelling initially, which would then generate a little bit more interest from you. And then if you're one level deeper in that discovery journey, reveal a little bit more that would then be helpful. So as you're, for example, watching a trailer, you might learn a little bit more about the title just by watching the trailer, but there may be also additional longer-form synopsis evidence or a little bit more on the cast or a little bit more on why this title is similar to others you've watched and why that

might be a compelling match. So it's for sure a hard challenge. It tends to pay great dividends for members to know a little bit more because they're excited to know about, well, why would this be a good fit for me? And the way it manifests for us is hopefully we observe then better matches between our recommendations of what ultimately gets played and then ultimately higher member satisfaction. Got it. Okay. So again, not to keep coming back to Squid Game, but with something like Squid Game, how do you tell that story to people?

Do you eventually just turn it around and it's like, just everybody's watching this? Like, shut up. Hit play. Everybody's watching this show. On that one, it's a good example of the robot in the art became particularly compelling and visually interesting. I think there's just something about that that was intriguing to folks. And then for sure, popularity plays a role there.

So as something gets bigger and bigger and it's, for example, high up the top 10 list, also your friends are talking about it. That social proof is often super compelling. And again, that's happening. You know, we don't want to take too much credit in Netflix. That's happening outside with your friends when you're on a run, on a walk, on social media. You're hearing about things. And so that's happening just as much, if not more, off Netflix and the biggest versions of success when something is really blowing up. Yeah, where do you see signals of that? I mean, you mentioned search is increasing, but also this

elusive sort of sense of buzz is a very hard thing to quantify. But if you can, becomes really useful. So, like, how do you, especially in that sort of growth moment before a show, like Squid Game or Wednesday or whatever, kind of takes over the world, what do you look at as those moments that, like, somebody is telling me on a run that I should go watch this show? Like, you want to capture that, And I end the best case scenario would be when I go sit down on my couch, there it is, the thing that my friend just recommended to me, right? How do you bridge those two things? You should think of those things as like super indirect, right? And so I mentioned a couple ways that would manifest the first before something launches.

That would be you come to Netflix and maybe in the case before Squid Game launches, you say, oh, I heard someone told me about this and you search for it and then you get to watch the trailer in advance and you find out it's not available yet. You click remind me. That same would be true on something like the Netflix Cup next week. where not available yet. You can learn a little bit if you open the app and you hit Remind me. So that's one indirect way. And then the other one would be search. And that tends to be great at indicating what people are excited about.

And the great thing about the search results experience is it's also personalized and will continue to improve over time as behavior indicates what are the right responses to particular queries. So at the outside, for example, when someone first searches for Squidgame, maybe the the results would not have Squid Game in precisely the first results position. But over time, we would learn, as people are playing Squid Game after they type SQ and then immediately Squid Game gets played at very high rates, then we would change the ordering of those recommendations and search results and hopefully match you very quickly with the thing that you actually have in 10 for. Got it. Okay. One of the smartest things I think Netflix has always done is that when you search for something that isn't on Netflix, it tries to recommend you things like that thing, which is just

The most like subtle, devious stroke of genius. It's great. It's like, we don't have this movie. Don't worry about that. Watch these other movies. And I suspect that works. It's worked for me many times where it's like it picks something spiritually close enough that I'm like, oh, okay, I've been meaning to watch that anyway. I had not really thought of search results as a discovery engine, but it kind of is one.

Very meaningful source of discovery. And you're right that it can be frustrating if you're searching for that thing that is either not available yet or isn't available. on Netflix in your country. And we still really want to help you have an incredible night to night with Netflix. And so hopefully we've got something, if not exactly that thing, that would be a good fit. What do you call the thing at the top? The super big thing. Is it like the hero?

Is it the marquee? What do you call it? We call that internally the billboard. The billboard. Okay. Let's talk about the billboard. The theory I hear about the billboard most often is that, A, it is an unbelievably powerful piece of streaming real estate that whatever you put there, a lot of people,

people are going to interact with in some way, shape, or form. And that B, that is where Netflix brute forces things to be popular by force-feeding it to 100 million people whenever it feels like it just because it wants to make a show popular. I have questions about both parts of that. But start with the beginning. Is that as powerful a piece of screen real estate as it seems like it might be? That whatever you put there is going to be more noticed by people? Certainly meaningful. But if you think about how members are spending time with Netflix, they're coming in with something in mind and going to search, or they have a category in mind, whether it's a film or a genre,

or this is the kind of mood I'm in tonight. So it's one of many ways that folks do find what they want to watch. I won't disagree. It's important. It's the first thing you see after you select your profile. Is it the most important? There's lots of ways folks are expressing what they're interested in to watch watching tonight. Okay.

And do you think this is a total aside? I'm going to get to my second question in a second. But I think the way you just described that is really interesting, because it sort of assumes people come in with like one data point about what they're looking for, which sort of makes sense. But also I can imagine there's a set of people out there who just open Netflix. Like I think this is the unique thing about Netflix at this point is that you're big enough to essentially just be the thing. And I'm just going to open it and see what happens.

Does that behavior exist? And do you have to treat those people differently who sincerely have no idea what they want? They just want to watch something? That would be nirvana for us where you're so confident that we're, we're going to have both options and make great recommendations among those options that when you want to relax and you're ready to watch or play something, that you think of us first. And so you sit down on the couch and you just hit the Netflix button on your remote because you're so confident you're going to find something great on Netflix.

You are completely right that there's this spectrum where some nights you know exactly what you're excited about. So when you're right in the middle of the newest season of, for example, for me right now, it's the crown where I'm watching it on our internal employee preview. And I'm like, I am so excited tonight to sit down on the couch and watch the next episode. So that's going to be more likely that I'm going to continue watching something as opposed to be looking for something new. And then there's lots of nights where folks sit down and they have no idea and they just want to find something great. and they're willing to sit down and go on a discovery journey. And that's the fun, quite challenging problem we have,

is how do we make that the right amount fun, serendipitous, but also efficient so that it doesn't feel like I'm slogging through it. So we want to go from an experience that can at times feel overwhelming to an experience that feels most of the time incredibly delightful so that when you do finally land on a thing that's a perfect fit for you, you look back on that discovery journey and it felt great. Is that a thing you can discern? in real time that like if I open up the app and I scroll past continue watching and I'm now five

rows deep, you can suddenly like assume that I'm on a more open-ended discovery experience and show me different stuff. Like is that, can you be that sort of responsive to what people are doing even like button click to button click? We're humble and we we know we can always do better on that front. But I think you make a good, you know, your hunch there is probably a good one where if you're completely in the mode for the thing that you were watching last night and you're ready to watch the next episode, then it's a pretty good chance you might go straight to it. But if you skip over it and you start to demonstrate interest more broadly than that, a job we want to do much better is be able to pick up on what you might be in the mood for tonight more quickly and adapt our

recommendations accordingly right in the moment. So that like if you're seems like, oh, tonight might be a movie night for David. And based on his browsing right now, might be a Tom Cruise movie night. We want to accordingly adjust and make it easy for you to then satisfy that interest. And you might not have known that before you showed up, but it might have, as you sort of browsing, is like, oh, this is a Tom Cruise movie night for me. I really like the idea that the longer I spend browsing Netflix, like the weirder the recommendations would start to get, that it's eventually just like, I don't know, just watch this. You clearly don't know what you want. We don't know what you want. Like, here's some stuff. Let's see what happens.

I kind of enjoy that. The magic would be the longer that you spend, the better that that final recommendation would be, because we would then know so much more about the things you could have been interested in and might not have interested in. So one thing that's a little more challenging is extracting, say, negative feedback from you. So it would be nice for us to know with a little bit more precision when you skip over something, whether you spend a little bit of time with a trailer and you say, oh, that's not for me tonight.

Sometimes that's a little more challenging for us to know precisely in the moment, and that's an area where we want to spend a little more time. That's interesting. The question of like, is this something that's not interesting to me or just not what I want to watch tonight, but like show this to me again in a few days kind of thing. Exactly. What a fun, hard problem where it's like, it's very difficult to necessarily parse. I never want to see this again. So a thumbs down means please never show this thing to me again or just not in the mood for it tonight. Yeah. Okay. The second question is I think the overarching theory about the billboard is that that is where Netflix like flexes its originals and says, We spend a bunch money on this show. You're going to watch it, so we're going to show it to you every time you open the app, whether you want to or not. Here it is.

Watch it. We demand. Is that how it works? Fall is the same approach that we've been talking about for the last 40 minutes or so. Personalized to you. We want our goals every time you open the app or to match you with something compelling. And whether that's Netflix original or not something we licensed from a studio supplier doesn't matter to us at all. We want you to be incredibly satisfied.

I really want to believe you. Like, I really earnestly want to believe you, but I can also see a bunch of really good reasons it would be the other way, right? Like, it makes sense for Netflix as a service to prioritize the stuff that it owns for lots of reasons that it's not going to leave the service anytime soon. You have control over the IP. You can do all kinds of stuff with more popular shows that you make than something that you've licensed from somebody else. Also, there are costs associated with these things, and you want to recruit the costs. And you made them because you think people are going to like them. So you want to put them in front of people. So I guess why not put some of that in there? Like, is there an internal balance that you have to strike in how to get that stuff right, even if you're not showing it to people? If you again, come back to our business model. The nice thing is we have really good alignment here. So with the subscription business model, you've got a choice every month. And the trade is your hard-earned money for incredible entertainment every month.

And so our end of the bargain is match you up with hopefully something great among an increasingly broad catalog. And so if we have a choice between something that's less good for you but is a Netflix original or something that is better for you and not a Netflix original, we're going to take that better for you every single time because our goal is for you to be incredibly satisfied and to be entertained. And so at any given night, if that's less likely to be a Netflix original based on what you like to watch or what you're in the mood for tonight, that's okay with us. So do you think the people who feel like they're being sort of, you know, bludgeoned over the head with Squid Game or whatever else, is that mostly due to, having a hard time reading those negative signals where it's like, okay, you skipped past this. Is that because you hate it and never want to see it again? Or just because it's not what you're in the mood for right now? Like as you get better at understanding what people are doing sort of click to click, do you think people will feel less and less of that sense of like this thing is here every time I open the app even though I don't want to watch it?

Some of that could be popularity oriented, as you mentioned earlier. So if something is quite popular and it's not for you, then you might see that in something like the top 10 list, which is quite literally. what are the series and films in your country that are top 10? That one I think is fairly understandable. Yeah. So those sorts of popularity signals can be hard to get around. I think, as I mentioned earlier, I think we have more to do on incorporating both more major and then that more modest negative feedback. Like, please don't recommend this to me anymore.

And so I think there's more we can do there for sure. All right. I have two more questions for you. One question people ask a lot about kind of anything that is algorithmically curated like this is how do I exert some control over it? Right? I think, like, I want to teach Netflix what I like and don't like. And I think, I'm assuming the first thing you're going to say is, like, watch stuff you like, which, fair enough. But what else do you look at as sort of strong, like actual user action signals that can push Netflix in the direction of stuff that they like? Let us know when you really love something. And the best way you can do that is when you really love it, give it two thumbs up. When you love it, give it a thumbs up. And if you don't like it, that's okay. Let us know. Give it a thumbs down. So that's one. very concrete way. I would think the thumbs down is a very strong signal. That is generally a strong

signal, but that's okay. We would, as I said, we'd like to hear it, because that then helps us know what resonated versus what didn't. So that's probably the best place to do it. And then generally speaking, we don't want to force effort on members. We want to make it effortless for you to both find something to watch and then enjoy watching it. So that's have an incredible discovery experience, have a playback experience that never rebuffers and always feels wonderful and makes it easy for you to turn on subs for a night that you want subs on and have, you know, hopefully magically know that for this particular language and content type combination, you would never want subs so we wouldn't automatically do that. But that's it, literally. Just watch, as you guessed, watch things that you like

and then give us the feedback. And the best way you can do that is thumbs up, thumbs down, or double thumbs up. Okay. Why not give people tons more tools to show you what they do and don't like? I know there's the, I don't know if you still do it. I haven't signed up for Netflix in a long time. But the thing where you would sort of solve the cold start problem by saying just like, rate a few things that here are some things. Do you like these or not? And that's a decent way of the cold start problem. But I know there are a lot of tools where you can like make lists. And some people have sort of stolen the Tinder mechanic of like right and left swiping stuff you like and don't like, why not give people access to sort of all that kind of stuff that is just like,

tell us everything you want to tell us about what you like so that we can tune it for you? Did we still do that taste tuning when you first sign up for Netflix? So you tell us a little bit about a handful of series or films that you like so that we can do a little bit better with that cold start problem and get you started with good recommendations. I think it's a fair question. Could from time to time you be asked to maybe retune every now and then and members might have an interest in that? So we've tried things like that in the past.

We don't have anything running right now or that does exactly that. But you could imagine, for example, if you've got a list of titles you watched and you want to provide ratings on those titles, then that might be a way for us to improve your recommendations even more so we can get that explicit feedback. So you've got a queue of 10 things you've finished over the last handful of months. Why wouldn't we ask you a little bit more directly? Give us a thumbs up, thumbs down, et cetera. So possible, you know, nothing on the horizon there precisely, but I like the idea.

Okay, fair enough. And then the last thing I'm curious about is how specifically it is useful to personalize. Like we talk about personalization, but there's a whole spectrum of things that that means. So like when you think about Netflix personalization, is that to me, David? Is that to my house? Is that to people like me? Is that to, I don't know, are there a million people in the cohort that I'm part of? Like how specific is it even sort of useful and valuable to go as you're thinking about this stuff?

Incredibly individual. So it's to your profile. So if you've got multiple profiles set up on your Netflix account, each of those profiles has an individually tailored set of recommendations. And if you're not doing things like demographics and all that stuff, they should have essentially nothing to do with each other, right? All we know about each of those profiles is literally the pattern of watching or game playing. And that is the only input and the purest input that we can take to then make the next great, hopefully, recommendation for you. Got it. Okay. So, yeah, I mean, and that winds me all the way back to, like, how much data is it useful to have part of this? Because I think, I don't know, presumably there are things you could know about me as a person that might make some of it better. But then I guess at the scale you're working at, like, that eventually becomes untenable to try and, like, individually identify every single person based on, like, their eye color and what that might mean they'd like. Quite honestly, we've got more than we can handle more ideas than we can handle it with just how members are using Netflix today. So just, just.

by revealing their taste and preferences based on among the recommendations we show, what do you try, what do you not finish, what do you finish all the way to completion, what do you finish quickly? As I was mentioning earlier, if there are ways for us to have a little bit better sense for what you loved versus you just like. So one way you could tell us that would be two thumbs up versus one thumb up and doing that more frequently, for example. But if there might be other ways where we could extract a little more signal with respect to time even better spent and time you're super thrilled to have spent with Netflix, those would be great ways for us to continue to enrich the way we improve our recommendations for you.

Got it. Okay. All right. This is my actual last question. I swear. I've said this three times out. This is my actual last question. Tell me about thumbs versus stars.

I find this totally fascinating. And I think you've been at Netflix long enough to have seen both sides. So I will just lay my cards on the table. I think five star scales are the stupidest thing on the internet. And I think they essentially mean nothing. But I'm curious about how you think about like, the one thumb, two thumb, thumbs down thing versus a five-star recommendation and why the thumbs works better.

Thumbs is a simple way for you to provide feedback on whether you like something or didn't like something. The trick on stars, it sounds like you're not in the stars camp, so I don't have to persuade you. The trick on stars is sometimes you might rate something very highly on the stars because that's how something, you think something ought to be perceived. So it's critically acclaimed. And so you feel like it deserves many stars, as opposed to how did you actually feel about it, and that's what's more meaningful to improve your recommendations is the reflection of

how you actually felt about something and whether it was good for you. And so thumbs has the benefit of, one, being simpler, and two, hopefully being a better reflection of what you actually love or didn't love about what you're watching on Netflix. That's interesting. That's actually a more, like, emotional answer than I was expecting. Because I think, to me, the problem is always that if I like something, do I give it five stars or four stars or three stars, literally no one knows and everybody has different ideas. So you end up with this totally meaningless scale. So I've always liked the thumbs up, thumbs down, because did I like this or not is a much easier question to answer, I think. But you're describing something really different where it's like,

we want you to tell us how you feel, not is this good? And that's actually like a really different thing. Is it good for you? I mean, that's the response that we would like. And so that's the benefit of the simple two thumbs up, one thumb up or one thumb down is that it's pretty clear to most people. What does that mean to you? I love it. All right, I promise that was my last question. So I will let you go. But thank you for doing this. This is really fun. I really appreciate it. Very fun, David. Great to spend the time with you. And I look forward to catching up against it. All right. We got to take one more break and then we're going to get to the hotline. We'll be right back. Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts.

But time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time. you could be giving to your customers. That's where LinkedIn Hiring Pro comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process

from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With hiring pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire.

Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply. Support for the show comes from Anthropic. Not every question has an easy answer. And the ones that are really worth asking usually come with a healthy mix of inspiration and backpedaling, aha moments, and quiet meditation. When you're working through one of those problems,

you want a partner to bounce ideas off of and figure out where the deeper issue lies. That's where Claude can help. Claude is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you, whether you're debugging code at midnight or strategizing your next business move. Cloud extends your thinking to tackle the problems that matter. Plus, Claude's research capabilities go deeper than basic search. It can have comprehensive,

reliable analysis with proper citations, turning hours of research into minutes. Ready to tackle bigger problems? Get started with Claude today at cloud.aI slash vergecast. That's clod.aI slash Vergecast and check out Claude Pro, which includes access to all the features mentioned in today's episode. Claude.aI slash vergecast. We're back. Let's take a question from the Vergecast hotline. As a reminder, you can always email Vergecast at theverge.com or call 866 Verge1 to reach us. We try and answer at least one question on the show every Wednesday. Thank you, as always, to everybody who calls an emails. It's so fun.

Oh, and speaking of which, we're actually planning to do an all hotline episode next week, all about buying advice specifically. So if you're trying to figure out what to buy this holiday season, or you're deciding between two things, or you want to talk to me about monitors, get all your questions in now. We're going to take as many as we can next week. This week, though, no buying advice.

Instead, we have a super fun, super wonky question that we get. got in the Vergecast email, and I have the world's two foremost experts here with me to answer it. Helen Havlack is the Verge's publisher, and Nilai Patel is our editor-in-chief. Between them, they run everything. Helen Havelack, welcome to the Vergecast. Thank you. Excited to be here. Nialli Patel is here, too. Hi, Neelai. Hey, how's it going. You know, I dial up my enthusiasm based on how excited I am for the guest.

Okay, this question is kind of long, so bear with me, but then we have a bunch of stuff to dive into here. The question comes from Anthony, and here's what it says. I feel like the effects music streaming has had on income for artists is fairly well-defined, or at least better known, but as an avid reader of various journalists and publications, such as yourselves, I have similar questions about journalism on the internet. Maybe supporting journalism isn't talked about as much because of the overwhelming rise in news consumption on social media platforms, but for those of us who try to consume news from dedicated aggregators like Apple News, flipboard, etc., I find myself asking,

what are those platforms do for journalist readership and pay? From my user perspective, it's a good balance between the convenience of a single source for all my news and interests, but also intentional about seeking news and not just Instagram filtering my news based on my likes. For example, is there a significant difference in how well a Verge article does when I read it from Flipboard instead of going to the site directly? Some of these aggregators filter out ads, so that has to have an effect on the money the Verge makes, right? I also wonder if there's a cost benefit of having articles available on those aggregators, because it expands readership, again, in an age where fewer people are seeking out news and just

passively consuming it on social media. So the question if I'm going to boil this all the way down is basically how does a website like The Verge think about where you read our stuff and what that means to us as a publication? Is that a fair summary of all of that, would you say? Yeah, I think so. And maybe also how does our publication make money and support the work we make? True. Two good questions. Okay, so if I understand this correctly, you two have thought about this and talked about this every day forever for all of your adult lives. Helen, why don't you start at this from a sort of business perspective?

Obviously, this is the thing we think about a lot, where our stuff is who we work with, how we distribute things versus people coming to the site. Is there a good sort of simple rubric of how you think about all of that? Yeah, absolutely. I think it helps to start with the basic of how does the verge make money. The verge is a free website. So the single most important way we make money today is advertising. We also make money from other things like affiliate commerce on our excellent product reviews. if people subscribe to our paid newsletters like Commandline and Hot Pod, if people become loyal

subscribers to things like this, our podcasts. But a bit of nuance on that ads business. Ads are worth the most if you serve them inside your own ad technology. We serve many of our ads inside Vox Media's ad technology concert. And I'll also say a loyal reader who comes back many times is worth a lot more than someone who just comes by once via search, for example. To use the music metaphor from this question. Someone who consistently buys your entire album and then lines up for tickets to your concert and buys your merch, that person is worth a lot more money to you than someone who listens to your hit single on the radio or a Spotify playlist. It's kind of the same for a news publisher, where the single most valuable thing to a news publisher is a loyal

reader who comes back over and over again. No matter what kind of business you run, Is it an advertising business, a subscription business? So the kind of like guiding North Star of this conversation is, okay, to what extent do platforms help us build that loyal readership? Yeah, Nilai, this is something you and I talk a lot about, right? Like in just sort of day-to-day coverage, this question of like, what's the tradeoff of sort of trying to be everywhere that people are, understanding that the world in which we live is not one where people like spend a lot of time typing URLs into desktop browser bars and all of that?

versus like, we want the website to be the thing. Kind of even more so with this redesign we did a year ago. Like, how do you sort that out in your head at this point? So I do think about this all the time. Our entire redesign is basically a response to this. And it's why I lured David back to the verge. Because if you look at most publishers on the internet, they just give their stuff away all the time.

They're like, here's what we do. People type. And then we give their words away to other people's apps. and hopefully they'll pay us. And that hopefully is a huge problem, right? Because it has just never come true for anyone. And then you look at their websites, the products they actually make, and it's just amalgamations of other people's software.

Other people's ad software, very commonly Google and Facebook and other people's ad software, other people's tracking software, very commonly other people's video players, just glomming up pages. And you're like, why would anybody come back to this? Why would anybody choose to have this experience? Of course they're going to go to Flipboard or Apple News or Google Discover or any of these other platforms where you're giving your stuff away for pennies on the dollar. And of course that is a doom loop for the media. And the music industry and the movie industry have figured this out.

And they figured it out because they have collective leverage, right? There's only so many big labels. There's only so many big studios. They figured it out in different ways. The studios funded Spotify, basically. and they have a tremendous amount of leverage over Spotify. Taylor Swift is writing letters to the Wall Street Journal about streaming rates, you know, five or six years ago.

The studios just built their own apps. They're like, screw it. We're going away from Netflix. We're going to build their own apps. And that pendulum is swinging back and forth. And I'm not sure that they've landed on the right conclusion. But then you look at the news industry and no one's tried any of it, right? It's just been giving stuff away left and right.

Very few publishers have stood up and said, We need to make a product that competes on the merits for user experience to actually capture attention. The New York Times has done a good job with this. Everybody recognizes it. But we cannot live in a world where the New York Times makes news and everyone else reacts to the New York Times. I think that is an unhealthy place to be. So Helen and I have spent a lot of time thinking about one, we do have a loyal audience. The Verge audience is our great asset.

Very few publishers have an audience like we do. okay, we have to build that audience a product that is worth coming back to you several times a day so that even if you encounter us on all the other platforms that we distribute, we can make an offer to you that coming to us directly is a better and more rewarding experience for you, the consumer of news, not just for us, the collector of revenue. I'll get a little technical, though. We do participate in a lot of news readers, and there's two kinds of news readers. And I think it's helpful to understand the difference between the two. in terms of how we think about them driving our business. So one kind of newsreader is basically an RSS feed. We give them an RSS feed. It runs on their platform. If someone clicks to read that article, what they get is a VIRG page.

They get a verge page with verge advertising, adverge commerce links, and all of that. And ideally, what we're doing with that is either someone is hitting that page and they are deciding to read more VARGE content in that newsreader. And I think this audience question proves you can have a loyal verge reader who primarily reads us. somewhere like Flipboard. That is possible. That's valid. That's still pretty good for business. There is a second kind of newsreader, and I think we've talked probably quite a lot on the Vergecast about Google AMP or the now defunct Facebook instant articles, which rather than giving people content as an RSS feed, which serves your own page, you would put your content in some proprietary code format that then is native to that platform. And the promise of that was,

okay, you're going to make less money per article maybe because you don't have your fancy ad formats, all of the rest. But the product that then runs natively in Facebook is going to load so fast and it's going to be such a great experience that the kind of volume you will get by participating there will make up for lost revenue. That in particular turned out not to be a great business to be in. For example, many publishers got way too dependent on Facebook. Call it circa 2016, the height of Facebook traffic. Everyone was serving instant articles. Everyone was gaming the algorithm. And then Facebook decided basically to shift the algorithm, prioritize family and friends content, deprioritized news content. And a lot of publishers straight up died, like went out of business, died. So that kind of

dependence on a single platform to mediate your audience relationship, that is super dangerous. And that is part of why, as we evaluate, you know, what readers do we want to be on? I think a good rubric is one, does that send people back to our website or let us create a loyal relationship with that person inside that app? And if not, it better have some other kind of value for us. Here, we're primarily talking about text content, but we also make things like video. I think TikTok's a good analogy for the purpose of that platform's a little bit different. You know, I think about TikTok as primarily a marketing platform, a marketing platform to bring the Verge content and brand to younger audiences and initiate them to the verge.

And so it is valid to approach platforms from different perspectives, but I would say it either better convert on loyalty or better have a really different value for us or it's not going to be somewhere we're going to want to participate. That TikTok thing is really interesting because if you have a big TikTok channel and you're a creator, people come to you and say, hey, help me sell my product. Like do some ads for us.

If we have a big TikTok channel, what we should be saying is, hey, visit our website where we make money because we're, we can't really do the advertising that creators can do on TikTok because we have this irritating wall between journalists and sales. Creators don't. So that whole platform is built on like integrated advertising with the content. And that's fine.

More power to the people doing that work. That's not what we do because we are very precious about making journalism. So our product that we have to sell is our website, right? You're watching these TikToks. You like our people. Come find us on our, on our platform. Our platform had better be good enough to convert. right? If our product isn't any good, and we're saying come to our website that is loaded up with

junkie ads and SEO spam, the whole premise of doing TikTok falls apart. So I think there's a forcing function. Okay, we're going to do TikTok and Reels and YouTube and all these other things, but underneath it, there had better be a product that's really worth it that we are marketing the same way that anybody would go to any other creator and market their product. Right. Yeah, what you're describing in a funny way is not that different from what a lot of the folks who are building paywalls around stuff say, right? Which is that essentially it aligns incentives. We want you to engage with our product and we want you to support us directly.

And what we're staying instead of give us money for a subscription is like come hang out on our platform. And that is like ultimately that is how we win and how you win because it makes us give you the best experience. And like that's a good alignment of incentives. It seems like everybody wins if we get there. But I do. There is one more thing I want to talk about, which is the.

like Netflix for news kind of thing that has come and gone at various times. Apple News is probably the most obvious version of that right now. But there are others where it's like give us some money, take all the news and we'll figure out how to split the revenue. Is that a real thing that's going to work for anybody, Helen? Well, it's interesting to see what Apple is doing. If you are in Apple News today as someone who does not pay for Apple News Plus, what you will find is that Apple is doing an awful lot of work to promote the Apple News Plus subscription.

And that seems to be their primary focus because historically, Apple's not that great at advertising. And so Apple News is an interesting one because the business model there is a little bit different than the promise of just ads can make it. And what they seem to be doing is going after a subscription bundle, which your Netflix metaphor, I think, accurately describes. And then in that, you know, if you as a publisher can have a good piece of that, that's a good business to be in. It's trickier if you are an ad-supported publisher to figure out how to effectively participate in Apple News and make that work. To be clear, I think all these Netflix for News ideas

are going to run into the same exact trap as the music industry. And maybe this is a good place then because we started with a direct comparison to the music industry. The music labels are doing great. Artists are not. The music labels have a big share of Spotify. They own a big percentage of Spotify. They can negotiate rates as collectives. Their job is to basically churn artists up and out through the system, right? What is the shelf life of Jack Harlow? Who knows? But the music industry is good at finding the next Jack Harlow and spamming him onto your feeds.

That's great. And that is the history of music industry. I'm not even making a judgment there. That is just the whole history of the music industry. The news industry doesn't have any of the protections, doesn't have any of the cultural relevance of the music industry. if The Verge decided to pull a Metallica and start suing every TikTok creator that read one of our articles to camera for copyright infringement, this is a disaster. No one can do this.

So we're just in this less strong position as people who make text than the music or movies industries are. There's no content ID for text. There's no sense that reading an article out loud should pay us, whereas everybody knows if you make a famous cover of a famous song, the songwriter should get paid. None of that is there. None of it exists. So we have to come up with something else. And then on top of it, you add these aggregators to the mix.

And it's obvious they're going to extract more value than they pay back to the publishers because that's what they have done in music. That's what they have done in movies. And again, it's the middlemen are great. The record labels are doing great. And maybe some of these big holding companies will do great. But the idea that you're going to get a famous writer or a famous journalist. journalist or a famous author from an aggregation platform that is like well compensated and

feels like a star, I don't think that's remotely possible. I think the economics of that are totally stacked against the deck. I think all of these platforms would prefer to deal with an infinite supply of young individuals than companies that might exert leverage. And I think that alignment mismatch, like you said, David, is so real. Every millennial media company, what's called them, BuzzFeed, Vice, whatever, our own. all thought that they would become these like institutional brands. And Facebooks of the world would pay carriage fees,

like the cable companies paid ESPN carriage fees, because their brands are so strong and people would want them. And what the platforms figured out is they don't need to do that if there's always another 22-year-old creator on the come-up. And you can pay everyone the same rates. And all the creators figured out, there's no money in the platforms, and they need to go start telling energy drinks,

and sneakers, which is exactly what happened. And so, like, stuck in the middle of that is, like, how do you pay for the news? Right? Like, maybe we should tell Parker to start selling energy drinks and sneakers. We've already got great T-shirts. But, like, most publishers are not individual brands like this that can just hawk a product to you, and they probably shouldn't be. So, Helen runs the business.

She has to be nicer than I do about the platforms. I just see this dynamics in a very clear way for us, which is if we want to make the journalism we want to make. We want to hold power to account. We want to do the reporter. We want all the stuff. We want to have fun on our website and not write for algorithms, but instead write for our audience. We have to build a business that is independent of these platforms, especially weird for us, because we cover the platforms. Do you know how weird it would be if we had an existential financial dependency on meta? I just think like, we have to build something else over here. And so it's great. You should go read us wherever you want. The question is

like, how do we make money? We're on these places because we want to find our people. there. And these algorithms are good at matching us with audiences sometimes. But once you're there, and if you really like us, the thing to do to support us right now is come to our website three times a day. And the thing to do in the future is, like, pay for command line and take our survey about what other paid products, because that's going to help us continue to be independent. And honestly, make it easier for us to go report on the platforms without this existential dread that we're going to piss off Mark Zuckerberg and he's going to turn our traffic to zero, which has happened. to many publishers many times over the past few years.

Because I am the money, I will say you do not need to feel guilty for reading The Verge on Flipboard. Podcasting, what all of you are listening to right now, you are listening to in someone else's app that we happen to distribute to with our own advertising, and that's a great business. So there are a lot of these news readers where we can have loyal readers who are following the Verge, reading us, reading our page experience with our advertising. It is a great business to be in. But as Nielai says, the primary place we need to go is to have that direct audience relationship not to be too dependent on any one platform or algorithm.

And to offer things on our website that you won't get anywhere else. Notably, quick posts, our live streams, none of that goes to any of these distributed publishers. So if you want to hear like Alex Heath and Nelai's real time reporting on Open AI last week, you wouldn't get that if you were just reading us in Flipboard. you would only get that if you were on Theverge.com, looking at our homepage, looking at the story stream about the drama going down. And so that is a big part of it is, yeah, what can we offer in the website that you can't get anywhere else? I like it. So as always, the answer is download every episode of the Vergecast 60 or 70 times and everything will be fine. Awesome. Thank you both. Appreciate it. Theverge.com. It's a good website.

You should read it. The only website left. All right. That's it for the Vergecast today. thank you to everyone who's on the show, and thank you, as always, for listening. There is, as always, lots more on everything that we talked about at theverge.com. Go watch Planet Earth. We'll put some links into the show notes, and, you know, read theverge.com. There's a lot going on, even when OpenAI chaos isn't happening every minute of every day. And as always, if you have thoughts, questions, feelings, or just please want to

tell me which monitor to buy, you can always email us at Vergecast at theverge.com, or keep calling the hotline. 866, Verge11. I love hearing from you, and we're going to do. do that episode all about buying advice next week. So get your questions in now. This show is produced by Andrew Marino and Liam James. Vergecast is Verge production and part of the Vox Media Podcast Network. Milai, Alex, and I will be back on Friday to talk more about what's going on in OpenAI because that just won't stop happening. Plus AI sports reporters, Amazon computers, and lots more. We'll see you then. Rock and roll.