How Netflix Picked Your Personality
26:26 Share

Hello and welcome to the TLDR podcast, a show about the culture, gossip, and business of money. And this week, the algorithm is making you do it.

My name is Tevin Friedman. I am here with my co-host Matt Keres is the director of product for WellSimple, our sponsor. Matt, how are you this week? I'm doing great. Today was flower day. So I went to the flower market and I got to engage in an act of commerce and it made my day as it does every week. Very Eliza Doolittle. I mean, I felt like Mrs. Dalloway. That's the vibe you've always given me. Sarah Rieger, the markets and business correspondent for the TLDR newsletter.

How are you? What's going on this week? I'm great. I feel like I'm dressed up as Matthew today. I'm rocking your quarter zip because it's so cold. No. Yeah. Is it wool? I could not tell you what it's made of.

Okay, we're going to do something a little bit different today. We're going to spend the whole episode talking about one big subject, which is big tech, but more specifically, one of the primary tools that big tech uses, which is the algorithm. If you have been in such places as Google, Netflix, Spotify, you have either used the algorithm or been used by the algorithm, depending on your point of view.

We're going to have some conversations, including with the journalist Kyle Chayka. He's one of my favorite writers about tech and culture, and he's written a book about how algorithms have come to shape our culture and our minds in big and small ways. We are going to share a little bit of that conversation later. But in the meantime, to get to this big question, we're going to ask another question that we always ask.

Sarah Rieger, who is making and losing money right now that's interesting to you? Google is losing control of the search engine market, which feels really wild to say. Like, how does a $2 trillion company lose even, like, a tiny bit of grip on a market it has a monopoly over? But I think it's kind of shot itself in the foot in a way that's made the entire internet worse, in my opinion. And I want to illustrate, like, just how polluted the results are now by doing a Google search for Google+.

Best Printer for Home Use 2024. Devin, would you humor me and doodle that phrase? Hold on. I'm doing it right now. Best Printer for Home Use 2024. What kind of links are you seeing? Well, there are a bunch of sponsored links. There's a CNET link for Best Printer 2024. There's

A New York Times must be a Wirecutter link. There's a Reddit link. So that's a little bit different than mine, which makes sense because Google surfaces different results for people based on their profile and their location. But on mine, one of the top hits is this article from The Verge that I think is really funny. So the article is titled, Best Printer 2024, Best Printer for Home Use, Office Use, Printing Labels, Printer for School, Homework Printer, You Are a Printer, We Are All Printers.

That's a good headline. I feel like it started out as service journalism and ended up with existential questions. Yeah, it sounds like a poem kind of written by a robot slowly going insane to me. Or William Carlos Williams. Absolutely. Basically, for two years now, The Verge has published a version of this same best printer article that's filled with search engine optimized words, or SEO, in order to kind of draw attention to how Google's algorithm prioritizes search queries. And

And if you've ever searched for like any product online, I feel like you've probably read an article like the one that this is parodying. Like, you know, here's what's popular. Here's its best features. And then there's an affiliate link that means they will make money off of whatever you purchase online.

And The Verge has to randomly update this article every once in a while because Google is constantly prioritizing new content. So I just want to read a line from the article, which is written by Verge editor Nailai Patel. He says,

Having worked in journalism and media, you are constantly aware of what the search engines are going to optimize. And it leads to really weird stuff. Like my daughter...

when she wants to cook something, she is constantly complaining, like, why, when I pull up a recipe, does it have 18 paragraphs about really weird stuff before it gets to the recipe? And I have to explain to her, somehow, someone at Google made it so you had to do that in order to get your recipe noticed.

Essentially, what you have is like all these editors and writers writing for the robot. Yeah, totally. It's created this real disconnect between the content that websites are surfacing, which right now seems to be like aggressively monetized garbage and like what people are actually looking for. And it is because of like this constant changing algorithm that's just like, what are the whims of what the robot wants? Yeah.

And I think people are really increasingly looking for content elsewhere. So Google Trends data shows that an increasing number of users are looking to Reddit or TikTok for information that feels more authentic to them by tagging on that keyword to their searches. So you wouldn't go to just Google and search like best printer anymore. You're going to search best printer Reddit or best printer TikTok to see what it feels like real people are saying. And another competitor it's losing ground to is Amazon because shoppers are just so

searching directly for products on the platform. Yeah, yeah, I do that myself. Is it possible that the algorithm could get better and then correct some of its problems? Definitely. And I think one of the ways it's trying to do that right now is through AI. Like, Matthew, I know you use ChatGPT a lot. Do you use it for search? Yeah, I've also actually been using the

Google AI recommendations that come up with a lot of their searches. Initially, like, I thought they were really bad. There was all that public backlash. But slowly and steadily, they've actually gotten better. That's interesting. I haven't been using Google's AI much. Yeah, something I've noticed about the world over the last couple of years, or at least, like, the financial world, is that

is that periodically you get consensus that X company is bad and going to die. It happened with Meta. It happened to some extent with Apple. They're not innovating anymore. And Google is very clearly in that vortex. But when you actually look at the data, I don't know, it doesn't seem that far from savable to me. Yeah, I really doubt while Google is losing some market share, if it's actually going to be a serious problem.

hit to its company at all. In the last decade, its share of the market has fallen from 90% to 80% of search engines, which means it's still obviously like completely on top. I think it's more one of those things where I wonder how much all of us dependent on this product are going to lose out as like the quality just declines. And obviously this is happening on like a lot of other places on the internet too. Like many social networks are overrun with

bots and kind of like gamified nonsense content that's really directly targeted at you. So it's not just a Google problem. Sarah, we're about to have an algorithm guy on the show. What does all this have to do with like the nature of algorithms? I think it's that the Google search algorithm doesn't just affect what gets made, but how what's getting made connects with what we see or purchase. Right. I guess the idea is also like,

we may think we're choosing things that we never really were choosing in the first place because those choices had been made already. Choice is an illusion created between those with power and those without.

All right, Matthew, you are up. Who is making or losing money that is interesting to you right now? Well, Netflix is making like a pretty amazing comeback, one that did not look at all possible a couple of years ago, and one that's also making me think that it might be the only game in town in a couple of years. Yes, Netflix, the behemoth, the thing that's going to eat all entertainment.

I don't really think of it as a comeback story. What is it coming back from? Like, what happened? I mean, it's kind of crazy to think about at this point. In the 2010s, like, the big tech stocks that everyone was talking about and buying and worrying if they were, like, getting too big were the FAANGs, Facebook, Amazon, Google's at the end. And the N was not NVIDIA, but it was Netflix. And it was the company that all investors thought was going to, like, eat the media sector.

And that continued into the early 2020s. And then it completely collapsed. In 2022, the stock was down 70%. The company was facing an enormous amount of competition from everyone from Hulu to Disney+. We had Amazon and Apple spending ungodly amounts of money to try to win what investors were calling the streaming wars.

And it really looked like Netflix was down and out. But fast forward two years, the stock has rebounded. It's at all time highs. The company has been adding millions and millions of subscribers. And so all of that, you know, has added up to a picture for me that makes it seem like, you know, Netflix has made an improbable comeback. Why did it have a turnaround? So if you actually look at their profits, their profits didn't fall all that much.

You know, the way that I think about it is that you had these new competitors enter the arena. That was a real business threat and something that really needed them to like, you know, spend money they could have pocketed away as profits. But that like the stock moved like much, much more than the actual company fundamentals changed.

And so, you know, you saw an overreaction to the downside and then a massive, massive kickback. And what happened to change, you know, the investors' perceptions of Netflix? Was it like just investor psychology that changed? A lot of it probably is the narrative, but there was also like a...

pretty meaningful shift in improvement in the business over the last couple of years. You know, in April of 2022, when the stock was basically near its trough, Reed Hastings, the founder, came out and laid out a couple of big ideas.

And they were basically able to execute on them. Like the first thing you said they were going to do was that they were going to crack down on password sharing. So basically it forced people to actually pay for their content. And how that turned out is probably the most pivotal reason that the stock has recovered. Surprisingly to everyone on Wall Street,

It wasn't like they cracked down on password sharing and people immediately canceled their subscriptions and went to something else. Instead, you actually saw that coincide with a pickup in subscribers, suggesting that either it didn't lead to massive churn or people wanted to watch Netflix so badly that they went and signed up themselves. The second pillar of Reed Hastings' plan was that they were going to launch an ad-supported tier. They...

got that out and they managed to get 40 million people to sign up and start watching those ads in less than two years. And the third thing they did is they demonstrated their ability to move into extremely lucrative new parts of the media landscape. Yes, we need to talk about the fact that they just hosted the biggest ad

sporting events. The one where the world saw Mike Tyson's butt cheeks? Yes. That's one way to pitch it. And they also have an NFL game on Christmas. You know, this is like a big shift. But like, I guess what Netflix had to prove is people will pay to be here. People will stick with us. What we have created, the relationship we have with our customers is actually strong enough that if we make it more expensive or we make them watch ads, they will stick around. Yeah.

Yeah, and Netflix is just, like, unbelievably good at retaining their users. Like, the stat that keeps going around the internet is they only lose 2% of their subscribers per month, whereas, like, the best of their competitors loses double that, and others lose, like, many, many more times that. And so people realize that they, like, were a much less leaky bucket.

And the big question, and this is something that I haven't found a good answer to, is like, how are they able to do that? Because like they're doing a lot of stuff that like classical business thinking would mean like they should see even more users leave than other companies, like raising their prices, cutting down on free sharing. It's not the fact that they have what I want, because if I think pretty hard about it, they actually don't usually. It's that they just own so much content that

that you feel like you have to be there if you want to watch TV.

It is interesting, though, that we have this perception of Netflix having all the content because like Netflix now has fewer movies than like a large blockbuster would have had when I was growing up. But I think it just has so much churn, I guess, of big titles that we have this perception of it as being huge. Churn of big titles, but not churn of people. I kind of wonder if Netflix is more like cable now, like they're putting out all the reality shows. They're putting out like the Hallmark Christmas movie slasher.

Like they're putting out like really mass appeal stuff and just pumping it out at like a really fast pace. The stuff that you don't care about but is on all the time. That's their jam. Well, and I think it's like guilty pleasure stuff too. Like I've reactivated Netflix to watch shows that people are talking about that I don't think really stuck in my head.

Like what? That, what's it called? Love is Blind. My sister was talking about it all the time. So I activated it to watch a season. It was awful. But I felt like part of the conversation by having watched it. Like Apple has been spending so much money to make really, really good shows that basically nobody is watching. Yeah. The numbers at Apple are nuts. They're wild. But like Apple is like prestige though. It's like pure prestige. And you know,

That's good for their brand. Is it good for getting the most eyeballs all the time? I mean, it's possible that that's the diff, which is that Netflix's job, and you can hear their CEO, Ted Serino, talk about it like they were in the business of engagement. They were the first major streaming company. They have the most viewers of any other streaming company in the world, actually more than all the other ones combined. And that that

reach gives them a disproportionate data advantage to figure out what people actually want and that they have like honed the ability to figure that out over the years. I mean, look, they have records of what people have been watching going back to the 90s from when they were sending out DVDs. I have a friend who wrote on the show, Nobody Wants This. And he said that when they were in the meeting with Netflix, supposedly Netflix said to them,

We know that Kristen Bell has the most clickable face on our platform. If her face is on a tile, more people will click it than any other person. That is the kind of data that Netflix has right now. I love this idea of a clickable face.

It sounds bad. It does freak me out a little bit that we're making movie decisions based on who has the most clickable face. Like, I don't know what that's going to mean for, I don't know, the quality. Hasn't that always been the case, though? Do we really want to go back to a world where, like, what we watched was determined by, like, a set of people in Los Angeles and we had to pay through the nose to see things once?

Well, there's a great piece by Willie Staley in the New York Times Magazine about the advent of the era of meh TV and how like Netflix having one means that you're going to get stuff that's clickable and can go on in the background, but you're not going to get stuff that's very interesting anymore. Like that period of peak TV is over. I mean, people are voting with their feet like that and they're getting all this content for like much cheaper than the old legacy brands could have ever provided. Yeah.

I do feel like I keep reactivating my Netflix account lots because there's like one show everyone's talking about. But then once I'm actually on there, it seems like a lot of stuff that's like very targeted at the type of viewer I am. But then when I actually do I check it out, there's like very little on the platform that actually interests me. But that's the main point. It's enough to keep your... What's that line? You know, you think you're out and they pull you back in. Girl, just checking up on that...

So the Netflix story at this point is basically Netflix has won. It has essentially eaten a lot of Hollywood and a lot of entertainment. And in part, that is because of its mastery of algorithms.

And here's Kyle Chayka, who is a staff writer at The New Yorker and author of the book Filter World: How Algorithms Flatten Culture, talking to us about an example of one of the ways the Netflix algorithm manipulates us into watching stuff that we may not naturally be interested in. On the homepage, they change the thumbnails of the shows that they show you so that you are more likely to watch them.

It'll make an action movie look more like a rom-com by showing the one thumbnail with, you know, a couple on it. There was a really infamous example of this where they showed Love Actually with a thumbnail of the one Black actor in that movie, even though that actor shows up for maybe five minutes in the entire thing. And yet for Black audiences...

that is how the show is marketed in the interface. The way that they package these things, I think, is really distorting. It's kind of unbelievable. Okay, that was part of a conversation that I recently had with Kyle.

that we're going to share with you today. We talked about how we got here to a place where algorithms are feeding us stories and information that are going to get us to stay where we are. And the fact that it actually wasn't always this way.

So recommendation algorithms specifically were kind of invented in the 1990s, but they emerged more in the late 2000s and particularly in the 2010s, which I kind of see as the decade of the algorithm. And why they were so important is because there was just suddenly so much more content. Like we went from the DIY GeoCities page and like web rings and link directories to like

massive social networks where millions of people were posting all the time. Right. There's just a firehose of content and information out there, and you need some sorting mechanism. For sure. The problem comes in when the algorithmic recommendations are operating at too wide of a scale that it gets beyond just reflecting our interests.

So on Facebook or on TikTok, it's not just your individual data that's being used to predict what you should be shown. It's everyone's data across the entire platform. So really, it's sort of valuing the lowest common denominator. Whatever pushes the most people's buttons is going to be what you're exposed to. Yeah, exactly. Are there places that do it that we may not notice or think about?

Yeah, I think so. Like if you open DoorDash or whatever, Grubhub, like a food ordering app, it will literally just suggest to you, here's what you should order from where right now in convenient little buttons. It's like, oh, it's 1pm. Would you like a burrito from your favorite local Mexican place? It's like subconsciously like implanting the desire to consume this specific thing, just trying to shortcut your consumption. Yeah.

Right. I mean, in some way, it's the problem with the way Silicon Valley works in general, which is entirely data-driven, and data is backward-looking. So it only tells you what you've already liked, you know, so it's the opposite of discovery in a way. Yeah. It's literally judging patterns from your past habits of consumption and then reiterating those patterns to try to get you to consume more and more quickly. Yeah.

If the algorithm knows that you will definitely keep listening to a certain artist, then to keep you listening, it will play that particular artist. It doesn't want to disrupt your consumption flow. These companies have different objectives than making sure that you see the most interesting thing. They have business needs, right? For sure. But then I think in the same vein, like...

The fundamental economic functions of these platforms are not good for artists or for culture and is not incentivizing the production of good new culture or new innovation. And so I think that's partly why we feel so stuck. TikTok is a really interesting example for me. I'm not really on TikTok, but my daughter is.

And I think it's a good case study in not just how TikTok serves you content, but how it has changed a culture industry. For sure. I mean, music...

I think is one of the spaces that's been most impacted by TikTok. First, it dominated people's attention, like particularly young users. So TikTok is how they consume music. And the unit of music on TikTok is the five to 10 second clip.

It's like the micro clip of a part of a song that then everyone pastes over their own videos or does their own little choreography to. And that's how a song gets shared and becomes more popular. So I think that really encouraged musicians and producers to reduce their attention down to that level of like the five to ten second increment of

Where a hook has to come immediately and it has to change every 10 to 20 seconds. Is there an example of a song that feels created for that platform? Oh, man. I mean, there's this indie band called Caffoonay from New York. They released a song called Tech It, T-E-K-I-T-E.

They didn't mean it necessarily to spread over TikTok, but it became a huge hit on TikTok specifically because of one part of the song. And people on TikTok took that refrain and looped it or slowed it down or sped it up. And so TikTok kind of isolated the one memeable part of that piece of music and made it really, really popular.

And that's how that band's got a record deal. Right. It's like you take the

vast ecosystem of the world and repave it with like Scott's turbo lawn. And it all, every, every, it all is the same. Exactly. Okay. So, so far, you know, the tech culture industry has exploited a certain part of, of what being a person is, but there's still another part of what being a person is that longs for meaning and interestingness and novelty and,

Is that the hope? That as long as that's there, someone will come along and serve that part of us? I hope so. I mean, I think we lived through the peak algorithmic platform era. And I think the new stuff that's coming up and emerging is...

less about that scalable content mush and more about trying to serve people's individuality in a more authentic way. I think sub stack culture is a little bit like that. You are paying a person you like money to do what they do. Band camp is a place where you can still buy an album and like pay money directly to the artists.

I really like this classical music streaming service called Idagio, which is, you know, as frictionless as Spotify in some ways, but also gives you so much more information and density and ways to navigate this body of culture. How does it work? I mean, there's...

On Adagio, you can sort by the year something was actually recorded versus the year it was published. You can sort by orchestra. You can search by soloist or composer or style of music. There are videos in which classical musicians introduce their favorite orchestras. The information is just so much more rich and navigable.

It sounds like in order for the next generation of more interesting culture tech to emerge, founders have to be motivated more by enriching the lives of people than by having the most scale. I think you do have to care about what content exists on your platform and how you are changing user behaviors in a way that is not just addicting and frictionless.

Okay, Sarah, tell us what we learned. We learned that the best printer is still the same printer, even if the algorithm has changed. We learned that to succeed in TV now, you need to have a clickable face. And we learned that if you're craving a burrito, it might be because the algorithms are telling you to.

That is it for this week. This show is sponsored by Wealthsimple. It is made by me, Devin Friedman, Matt Keres, Sarah Rieger, with Mathilde Erfolino, Leah Fetter, Kat Angus, Jared Sullivan, with help from Tom Johnson and Allison Hopkins, fact-checking by Bernard Doherty, theme music by Andy Huckvale, and engineering by Emma Munger. Special thanks this week to Kyle Chayka.

The TLDR podcast is offered by Wealthsimple Media Incorporated and is for informational purposes only. The content in the TLDR podcast is not investment advice, a recommendation to buy or sell assets or securities, and does not represent the views of Wealthsimple Financial Corporation or any of its other subsidiaries or affiliates. Wealthsimple Media Incorporated does not endorse any third-party views referencing this content. More information at wealthsimple.com slash TLDR.