How Does Netflix Get Us To Binge Watch?

Sam. The Science of Change podcast is presented by Set Sail, the world's number one revenue execution platform. My guest today is someone I'm excited to chat with. Gibson Biddle is a product leader who has worked or consulted with some of the biggest brands in the world. And for five years back in the 2000s, he was the VP of Product for a small DVD delivery startup called Netflix. Gibson was there to test and build the platform we all know and binge on today. So Netflix got literally millions of people through the pandemic in 2020. They added more than 36 million subscribers to pass the 200 million subscriber mark. Think about that for a minute. 200 million people is a lot. To give it even more context, it's 20% of the US population and 2.5% of the global population that subscribes to Netflix. Netflix was among the first in the streaming game, but now, with what seems like millions of competitors, they're still owning the game. They're obviously doing something right. But what is it? What have they learned about human behavior with all of these at bats? Gibson was at Netflix in its earliest days and he revealed quite a lot in our conversation, from what predicts our movie preferences to conversion rates of free trials. We also get into a light conversation on the ethics of autoplay and deep dive on the theory of revealed preferences. This is a super fun episode. So get ready to Netflix and chill. Well, sort of with me and Gibson.

So maybe we hop right into Netflix. What is your favorite Netflix test result?

Well, my favorite mainly was because it informed us so much and it helped us to create a model that we could apply to other questions and problems. So it would be early days, it would be during Netflix's DVD era at a time when it was maybe it had a million members. But the model we developed, we said our job is to delight customers in hard to copy margin enhancing ways and margin fancy way to save make money. The big test that we did was if you asked people in a survey or if you asked them in qualitative and focus groups what sucked about Netflix, they'd say the same thing. I don't get my DVDs fast enough. So in fact, in specific, I don't get my new release DVDs fast enough. So if you remember at the time, three months after a movie came out in the theater, the DVD was finally available. And if you added that DVD to the top of your movie list, some people would get it the next day in the mail, which is fast. And some people would have to wait two or three or four weeks because we just didn't have enough money to meet that initial demand. We knew after four or six weeks it would drop off. But we set up a test. Our metric was retention. So at the time, about 5% of our members would cancel every month. So the job was to see if you could lower that to 4.9 or 4.5% cancel. Only 4.5 canceling each month. And so we executed the test. And what would your guess be? So the structure of the test was every member in this special cell got their new release DVD substantially faster within a day or two, where people in the control might have to wait a week or two. So the question is, will it improve retention?

It's a good question. In a rational world, it would definitely improve retention.

If we value our time and we.

Value the movies, we'd want to watch them as soon as possible. But we don't live in a rational world, so it deserves more consideration. Imagine you could, by some weird twist of fate, kiss your favorite celebrity. Now, obviously consent is important. They agree. But imagine they consented to kissing you. When do you choose to kiss this celebrity? Do you choose tomorrow? Or do you choose to wait and do it in two weeks? One study posed this very question. And maybe surprisingly, more people chose to wait rather than to kiss the celebrity immediately. But is it so surprising? Wouldn't the anticipation over those two weeks be just as much a part of the experience as the kiss itself? Why is this? People tend to view waiting for experiential purchases to be much more pleasant than material ones. A movie is an experience and waiting adds to the excitement. Whereas you could be annoyed or frustrated waiting for your Amazon purchase to arrive. This means waiting and anticipating may be the key to maximizing happiness from an experience. For example, you should be booking a vacation way in advance. Book a restaurant rather than just show up and make plans to see friends two weeks before actually seeing them. Experiences can typically provide the most happiness.

Before they actually happen. I may go the counterintuitive answer and say that getting it right away may not be as as helpful to people or as enticing to them as they may think.

I see. So your consumer insight is people, they just need the moment to just oh, I just can't wait for that new release DVD to arrive. Well, the beauty of this one is if you answered yes, it will change customer behavior, we'll see an improvement in retention, or no, I don't think this really matters. You'd both be right because we changed it. But by a surprisingly small Amount. Our bias walking in was that this would be a game changer. And it turned out that people didn't value getting that new release DVD faster nearly as much as we thought. So the context here was we lowered a bit by a tenth. And that led us to do some math by the data. We said if we rolled this out to all, gave everybody a perfect new release experience, got it the next day in the mail of those million customers, if we wrote it out to all, we would save 5,000 customers. That's how the retention improvement looked. And the question was, how should we value 5,000 saved customers? And our math at the time, the lifetime value of a customer was about 100 bucks. So we took those 5,000 saved customers times 100 bucks and then we multiplied it by two. And the reason we multiplied it by two and this was deeply debated. We called that a word of mouth factor. But the idea was that the folks that had enjoyed this improvement, that stayed with the service, would always also rave about it and they would tell a friend about it. Now, of course, if you were at Amazon, that word of mouth factor might be 10x, but my CFO partner was apoplectic at valuing the word of mouth factor at two because it really encourages more investment. Oh, and it turns out that the cost, so say we were going to save a million bucks in this case, but the cost through the value of the customer, that stayed with us. The cost of the new release, getting lots of new releases in our library was about 5 million. So this is one of those things where it was a million of the good through a retention improvement, but 5 million to the bad through a bigger investment in DVDs. And by the way, in that case, Amazon would have done it with their 10x, but we were a poor startup, we couldn't do it. And it really got us thinking about how do we evaluate the trade off between delight and between margin. And the other thing it got us thinking about was building hard to copy advantage. That failure, if you will. What it did lead us to was thinking harder about the value of getting a DVD the next day in the mail, irrespective of whether it was a new release. And in that case we learned it actually was quite valuable. So we had a proxy metric. It was when we first looked at the data, about 50% of our customers were getting the DVD they wanted the next day in the mail. And we drove that up to about 90 or 95% eventually. And that improved retention. The way we did it was we had 20 hubs around the United States, we bumped it all the way up to about 100 hubs. And that was our hard to copy advantage. The operational capability, the magic of how we managed to get DVDs to customers the next day in the mail. And there were lots of secret magic tricks. Blockbuster spent years trying to figure it out. One of them was when we would merchandise your personalized homepage, Kristen. We would only merchandise the titles that were available in your local hub, which I presume is down in San Jose.

Yeah, actually, one of the things I was reading about, one of the findings that Netflix has had is that basically the idea that high quality movies, so basically, if people are rating them in high stars, you're giving something a five out of five that could drive retention. And it's a pretty logical hypothesis that if there's a really good movie that I just watched, then I'll come back to you for more movies. But it seems like there was another counterintuitive finding there, that maybe the high quality ratings didn't exactly correlate to retention, that people still found Jo in lower quality movies. Can you explain that finding a bit?

Yeah. So the hypothesis was just as you said, that if we were to drive the average movie ratings of our members from 4.0 to 4.4 on a five star basis, of course that's a better experience for customers and of course retention will improve. And this was within our personalization area. We spent a lot of time trying to magically guess what movies you would love. Our job was to connect you with movies you love. And so to your point, we drove it from 4 to 4.4. And then we didn't see any data that suggested that an improvement in movie ratings actually improve retention. So that's when we sort of went back to the drawing boards. We talked to customers and focus groups and we'd ask them questions like, what's your favorite movie? And in front of their 10 or 12 peers, they would say Hotel Rwanda or Schindler's List or some documentary about artificial intelligence stuff that would make them look smart. But then if you dug a little deeper, you'd find amazing excitement for, you know, like, think of any Adam Sandler movie. Okay. Or the favorite was Paul Blart's Mall Cop, which if you looked at the data, it was an average rating of 3. Right. But people really enjoyed those movies. And at some point we realized that our job was to provide movie enjoyment and ratings didn't really matter.

So this is the difference between what we say and what we do. We want to say we'd prefer to watch the art house film, but when given an option, we opt for Zoolander. There's a study that had one group of people select seven movies to watch over the course of the next week. One movie per night. And another group of people got to select the movie they watched each night right before they watched it. These movies were categorized between highbrow and lowbrow Films in the highbrow category were titles like Schindler's List and like Water for Chocolate. The lowbrow movies were things like My Cousin Vinnie, the Map or Mrs. Doubtfire. So what happened? Sure enough, nearly 60% of the movies people picked in advance were the highbrow titles. When we're picking for our future self and the week ahead, we pick what we think we should watch. Indie films, the classics, documentaries. But when we're picking for right now, only 43% of the movies picked were from the highbrow category. When we're picking for today, we go lowbrow. Gibson is right that it is really hard to predict your own movie tastes. And this doesn't only apply to movies or pop cultural tastes. Psychological research consistently shows that people themselves are notoriously bad at making predictions about what they'll like. Studies have shown that we tend to overestimate how unhappy we'll be after receiving bad test results or a divorce. On the flip, we also overestimate how happy we'll be when our favorite team wins a championship or when we start a new romantic relationship. When we're thinking about the happiness of our future selves, we make all kinds of prediction errors. One of the most common mistakes we make is simply thinking. Thinking that we'll be better people in the future. You know, the kind of people who like highbrow movies.

So Schindler's List, think about it. Sure, it's a five star movie, but it's kind of a downer. And then you got Paul Blart's Mall Cop, or even more excitingly, Paul Blart's Mall Cop 2. Those are both three star movies. But those are the things that perpetuate enjoyment. And it's actually movie enjoyment, not movie quality, that drives retention. But that was a super non intuitive learning for us. And our world was full of those surprising results. Largely, a lot of them were in the area of personalization. It's really hard to predict people's movie tastes and there were a lot of non intuitive results within that area for us.

Yep.

Wait, Kristin, you're not taking. You're not. I threw out the hook and you didn't grab it Gim. What are some of those other non intuitive results?

I'm trying sometimes I want to make sure we. I have so many questions for you, Gib. I have so many questions for you, so sometimes I want to keep moving. What are those non intuitive preferences that people have that they.

Oh, I'm glad you asked, Kristen.

Yeah.

Well, here's the question. If you're trying to predict people's movie taste, is understanding the age and gender or the demographics of a person, is it important in predicting a person's movie taste? That's a simple one. Yes or no. What do you think? Hmm.

So age basically. Is demographics good at predicting movie tapes?

Is it helpful?

Yeah, I would probably say no in most behavioral science studies, surprisingly, you don't find that demographics has as much of an impact as you would think. In some kind of financial experiments, you'd think that something like income has a bigger effect on results. But many times income doesn't matter. Gender matters sometimes, but not as much as you think.

You are correct. When we added that data, I mean our the main thing we were doing was try to predict your movie rating. So if you get a person's age and demographic data, it doesn't help you to predict how much they're going to like a movie. So that was surprising. At the end of the day, what best predicts a person's movie taste is just to know the three, four or five or six favorite movies or TV shows. That sort of seeds the engine, if you will. So if you go home tonight just for a test, set up another profile on your Netflix account and all it's going to ask you are, hey, do you like any of these movies or TV shows? It's just trying to get three, four or five to begin that that personalization journey. But again, it was not intuitive to us that demographics don't matter. That and the idea that better movie ratings doesn't actually predict enjoyment. So this is just the wild world of consumer science and why it's so freaking hard.

This is a good insight. Pretty much anything, not just movies. Past behavior, not demographics, is what typically predicts future choices. Look for simple creatures. What we do today is reflective of what we did yesterday. This is also important with habits. Frequency of past behavior reflects habit strength. What best predicts if you will exercise tomorrow and keep up that habit of running if you exercised in the last few weeks? Past behavior predicts future behavior. So it makes sense that the Netflix algorithm is based on which three to five films you've already watched and liked.

One Thing I have noticed is that there's a carousel on the homepage when you log into Netflix, and it usually highlights one movie. You know, Manifest was a popular one in the last year. It was highlighted on the carousel.

And.

And lo and behold, it also turns out to be the number one Netflix movie that week or series that week. How much are those connected by which? Whatever you show people on the carousel is actually what they watch. So basically, how much are my preferences determined by the algorithm versus my preferences determined by, you'd say, kind of personal agency. If you show me something in the very beginning, am I likely to watch it because it's what you show me, or is that actually my preference?

That's a great question. I mean, let's say it in a more foreboding way that Netflix is trying to make its business work, and because of that, it's taking the content it owns and it's putting it very aggressively on your homepage. That would be another way of spinning that question. Here's the way I think about it. In 2005, the average customer would look at about 100 titles before they chose something. And in 2010, it was probably 50. You know, you look at 50 things before you chose something, and now we're getting into the maybe 30 category. So you kind of lightly browse 30 things before you pick something. Netflix would like that to get to one eventually. And so it's. It's going to be magical. You turn on Netflix, it just starts playing something that you're just, like, totally jazzed for. That's the goal state. So back to your question. That hero that. That. That large placement for Manifest, I think you said, is really the thing that Netflix has super high confidence that it thinks you'll enjoy. You'll notice something else subtle. But this is, you know, this is a flash into that vision of just one thing. There's a little button that says play something. Right? And that's for a person who says, hey, finding stuff is way too hard. I just want to watch something. I mean, that's what you usually do when you're scanning all those TV channels on normal TV. And my guess is today that maybe 3% are hitting that button and continuing to watch. But think of it as a test ground for doing a great job at predicting what you'll love. If five years from now it's 15% are hitting the play something button, that's a sign that their intuition is radically better. So back to Manifest. It's there because Netflix thinks it will delight you, and the economics can Work. I'll give you one other past thing. When I was at Netflix, we actually developed a cost based algorithm. So we would look at how much a person would like a movie and if they would like it. He had two five star movies, but one of them was a little cheaper. We would push the five star, but we would never take a 4.9 movie and push it higher than the five star because it was cheaper to us. So the mindset is how do we delight customers? How do we delight customers? And I'm pretty sure that that mentality is still very much at work. So I'll tell you something else, like we put a lot of time and energy into that cost based algorithm and then we killed it. It just wasn't adding enough for the business. And that was one of those moments where I started focusing my energy and effort more on how do we delight customers, how do we delight customers? And not worrying as much about the economics.

That's a lot to think about.

So Netflix has a great algorithm of what they think we'll like and they've gotten really good at making it easy for us to find it and click on it and watch it. They decrease the time we spend on the couch with our loved ones agonizing over what to watch. That's not so bad, right? Right. Well, maybe the essence of what Gibson is talking about is what we call revealed preferences. We look at the things people are doing and we infer that those are their actual preferences. If you bought this widget, we can infer you have a preference for this widget and value its utility in your life. But what if what you choose in the moment conflicts with your long term preferences? Behavioral scientist and Caltech professor Colin Kammerer.

Reflected in a paper.

Suppose a clinically overweight person loves pizza delivered in large fatty portions. A revealed preference view would insist that since he ordered the pizzas, he must have higher utility for pizza than for something like a salad.

But suppose he goes to his doctor.

And says he wants to cut back on pizza. Showing up and saying this reveals a meta preference for changing his revealed preference. Should we use his revealed meta preference to override what we infer from his actual behavior eating pizza? This is the problem Netflix faces. It looks like people have a preference for the big video on the homepage or to binge watch three episodes in a row. But is this really the case? My meta preference may be to go to bed earlier and avoid watching that last episode, or to explore a new genre of movies or drama and broaden my horizons. But Netflix creates their algorithm. By looking at my main revealed preference, this looks like I love Manifest and want to keep watching it because autoplay is on and I just let it do its thing. But that's not really the case. It's actually that I have a preference for going to bed, but doing nothing and letting autoplay run is just easier and more fun in the moment. Sorry, Manifest. I'm not really your fangirl. It just looks that way to Netflix. What does Gibson have to say about this?

It's a hard problem. Let me try to tease through it. So first, Netflix invented binge watching. And we did that when we launched streaming in 2007. We didn't have a lot of content. It wasn't great content. But we actually slowly discovered the magic of episodic TV series. And they were magical for a few reasons. And then we made a very conscious decision to release all episodes to a customer at release, which nobody else is doing. And some of the other competitors still are not doing that. So just the idea of binge watching. And a lot of people thought, you know, their intuition was, well, that's bad because people are watching a lot of content. That's going to cost you. It's going to cost you Gibb. But the reality was binge watching was really delightful. And it's sort of one of the things that people understand about what Netflix does and how it does it.

Binge watching is delightful for Netflix because it helps them retain members. But as suggested by one comprehensive review, binge watching, like any other behavioral addiction is linked to higher levels of stress, anxiety, and depression. Binge watchers, on average, have lower sleep quality, more daytime fatigue, and a less healthy diet than people who don't. Binge watching can also lead to a sense of emptiness. One study found that binge watchers tend to experience a decrease in meaningful and positive affect after binge watching as a whole. Whether binge watching is delightful is debatable, but I'll let Gibson keep going.

So first, Netflix doesn't worry about the cost of your streaming more content. It just wants to make sure you enjoy it. The proxy metric there is the percentage of customers who watch at least. I'm going to make a kind of a educated guess. At least 50 hours in a month. That's the proxy for retention, because retention is really hard to move and really hard to measure. And so over the years, it's probably maybe 50% of Netflix members watch at least 50 hours in a month. And they're hoping to push that to 60%, that proxy metric, and that hopefully will get Netflix retention today. Only 2% cancel each month. That'll drive it down to. The hope is to only 1.9. Each of those tenths is a big win. So I think if Netflix thinks you're asleep, it wants you to stop because you're not enjoying that. And if you want to keep watching, it's going to let you and encourage you. And at that moment in time, it thinks that you're watching the next episode is better than trying to find something new to watch, you know, in that session, which is intuitive and going to sleep or going to sleep. Yeah.

Right.

Yeah.

It's made up.

It's kind of made the paternalistic decision for me that watching the next episode is better than the hour of sleep. And I think that's where it's tough.

Yeah. The other thing that comes up with people is. And it comes up personally for me, because you could hear, in my career, I go back and forth between entertainment and education. And so education, you know, I'm clearly doing good for the world and entertainment. My wife, who's trying to cure cancer, says, give your rotting out people's brains, binge watching is not a good thing. So Netflix is a little aware of that. But the way that they think about it is there are amazing stories from all over the world that they're bringing to life. Now. If you see, one of the cool things is they've married the idea. Documentaries used to be boring and you go to special kind of artistic theaters to watch them, and now they do these documentaries in the episodic TV format, which I find really, really interesting. And at heart, they're educating, and at the heart, they're helping me to appreciate the stories of people all over the world, which I think is way cool.

So I've heard, and this is a rumor you can verify or not, is that basically every time Netflix changes the algorithm, there's a small bump downwards in engagement. And the theory behind this would be that people don't like change. And so you can't actually measure if the algorithm is working or not right away. You have to wait until people adjust to it in order to see if the new algorithm has actually increased watch time, retention or engagement. Is that true or is that something that the Internet told us?

Well, I'll just. Nope. Well, here's what I know. Those five years I was in Netflix, most of our testing was with new members. So the idea was if you were. If you're trying to grow from 1 million to 200 million members, it's all about creating a simple Delightful experience for all of these new members. And I didn't worry very much about existing. I wouldn't test existing members because just to your point, they don't like change. It was all what I would nicely say, it's the new member stupid. And so I would do new member testing. Occasionally we would notice something that, well, this feels pretty big. So we would do a back test when we would roll it out to existing members. But to your point, and by the way, totally ignore what Twitter says and ignore what everybody's saying to your point, people do not like change. And all those Twitter people, they're kind of quasi experts and they are not normal human beings, right? So the way I sort of navigated this was it's the new members, stupid. All of our AB testing was with new members. From time to time, if we thought it was too big a deal, we might do a back test with existing members. But I don't think that ever changed us or caused us not to make a change. And then when we made big changes like price testing or price changes, you know, the main thing we learned was to over communicate. Like think about it. When your price has gone up, Netflix tells you six months in advance, we're going to raise your price. And then three months in advance we're going to raise your price. And then they finally do it, you're like, oh thank God, they finally raised my price. So that's one of the ways that we learned that basically through over communicate to handle this people hate change phenomenon.

Let's move on to other examples that you've seen in the wild.

Are there other A B tests that.

Stand out that you've heard from companies, either your companies or companies that you've worked with, that speak to basically either the power of a B testing or the non intuitive nature of the results?

Some other like from Netflix, the interesting ones are when the A B test results say you should do one thing, but you choose to do the other. So for me, a good example, we grew up on a free trial. So your first month it says watch anytime, anywhere and cancel anytime. So people, that first month when you sign up, you give your email your credit card information and then you're enjoying watching for free. And then at the end of the 30 days it switches you to pay. And there's this phenomenon where of course on day 31 or 32 or 33, people call in and say, hey, I forgot to cancel. And so there's like 20 million or 50 million in phone calls at the time from that phenomenon. And so A bright product manager said, hey, let's do a free trial reminder. So on day 28, let's text them, let's email them, let's put on their personalized homepage, hey, your free trial is about to end. So if you want to cancel, cancel now. And so you can a B test that. And just so you know, what would be your guess on actually two questions? What percentage of folks that hit that Netflix non member page actually sign up for the free trial? And I'll give you a clue. It's between 0 and 100.

What percentage of people are signing up for the free trial? I reflect on this idea of, you know, when you're in the airport line and in the airport line you have the option to go to clear, which is better. You can sign up for a free trial and you cut the whole line and you can pass people and make your plane on time. So why doesn't everybody do that? And one of the reflections is because you know that you will not remember to cancel the free trial. So that prevents you from signing up. So I would say maybe a lower number than expected. Because in theory a free trial is just good. There's no downside. But if we think that we're going to forget, there may be a downside. So I'll give you the guess of 30%.

Yeah. So substantially you're right to go low and it's substantially lower than you expected. And most people in the credit card industry will guess it's 2%. It's been 2%, like forever. And the real insight there is people are busy or maybe they're on their phone and it's kind of wonky to sign up on the phone. Or hey, we used to call it, honey, it's time for dinner. Right? They just broke the process anyway, so 2%. But here's the second question.

Sorry. And to clarify, is this a with.

Credit card or without?

So it's with credit card.

Okay.

Okay, there we go. So if it's with credit card, you've got a fair amount of friction there.

Totally, totally. Oh, an email. And by the way, Netflix is tested without email, without credit card, the results are horrible. Right. With anomalies. But anyways, my next question is what percentage of those folks that have signed up for free trial with their email, with their credit card information, what percent go on to become paying members, which is to say they retain after one month? What's your guess there?

Yeah, I'll still go around the 30%.

To say that if you're with the.

Trial, majority of People may not continue.

I hear you, I hear you. Yeah. And I'm bringing these up because the results are surprising. Right. So that one's 90%. Okay. So people are really enjoying their first months with the search service. It's pretty easy to find movies you love.

And then what happens in month two?

Because you. Because maybe the.

Okay, we'll get to that.

Okay, great.

Because I would imagine people forget and then of course you get them. And that's the essence of a free trial.

Okay. But here's the case. Here's the case. We're going to a B test whether or not we should do a free trial reminder. Should we tell people at day 28 that remind them, hey, we're about to hit your credit card cancel if you don't like it. So the key question is, what will happen to that 90% conversion? It is going to drop. And the question is how much will it drop? That's my question of you.

How much it drops if people know before the credit card hits that they.

Are going to be charged.

Yep, that's it. You're going to switch from free to paid unless you click this button to cancel. So the conversion is going to go from 90% down to what number?

Yeah, I've been wrong in the last two. So we'll go down to 75% that you get.

You get a quarter of people. Yeah.

Okay. Well, it's certainly directionally correct, by the way. This would be impossible to guess. Right. It goes to something like, I don't know, 88%. But congratulations, as the product manager for this feature, you just lost the company 150 million bucks. Except we're going to back out 50 million because we're not spending all that time on customer support calls with people calling on day 31, 32. So here's the deal. Does this delight customers in hard to copy margin enhancing ways? So is this a delightful thing?

It's definitely pro consumer. I mean, I think people want this.

Yes. And we already know on the margin side we're losing 100 million bucks. Is this creating hard to copy advantage? That's the key question.

Do more people sign up when they know they're going to get a reminder?

Good question. So yes, it says short answer is yes, but it's nominal. So let's not worry about that right now. But you're right. Based on this experiment, we should be aggressive about letting people know at the top of the funnel. Don't worry, you'll get a 28 day reminder. Okay, good insight. Right. But my question was we know this is going to delight. We know we're going to lose a lot of margin. Are we building hard to copy advantage by thinking about doing this? Because the AB test says you're going to lose 100 million bucks and you shouldn't do it. But why might you do it?

This is a question back to me. Why? Why might you do it? I think there is a trust issue, right, with companies.

There you go. And what is trust? What hard to copy advantage are you building with trust?

I actually don't know. I think trust may be overrated by many companies trying to build it.

But it's about building a brand, right? The hard to copy advantages that I focus on most are brand basically network effects, unique technology like personalization, and then economies scale. So brand's a big one of those. And by doing this, Netflix could build much more than $100 million worth of trust and perpetuate an even stronger brand than Netflix is today. So anyways, if you were the product manager, you're looking at these A B test results, what would you do? Would you choose to roll it out to all for the advantage of building hard to copy advantage in the brand and lose the 100 million box, or would you not roll it out? That's the question.

Why does sending reminders to cancel build trust and brand? When Netflix sends out reminders, the chances are more people are going to cancel and Netflix will make less money. And people know that. Reciprocity principles state that people tend to feel compelled to pay back a positive action from others with another positive action. This social norm of reciprocity is powerful and regulates behavior in almost all aspects of life. Waitresses who smile broadly receive more tips from customers than others with a minimal smile. A particularly powerful example of this is the use of free samples as a sales technique. In supermarkets. Customers are frequently given small amounts of a certain product for free. For many people, it seems to be very difficult to accept samples from a smiling attendant without actually buying anything. Netflix seems to understand the value of give before you take. By sending a kind reminder to people, people may actually be more likely to repay the actual by keeping the membership. This act can also build trust based on the idea of costly signaling. Plainly costly signaling is a signal to the market that the behavior is costly to perform. The act of Netflix telling people they will send reminders to cancel their subscription is a signal that they are willing to lose money in order to remind people this is costly. Research has shown that cost signaling and sacrificing one's own benefits tend to increase trust this is true across different contexts. People, regardless of their religious affiliation, judge Christians and Muslims who follow their own religion's strict dietary laws as more trustworthy than those who don't follow the dietary restrictions. By sending reminders that cost them, Netflix builds trust.

Well, let's end on one question we like to ask all of our guests, which is thinking about the future. So there's this analogy of what happened in the film industry where in one of the first feature films that happened, had a unique thing where the camera was placed just in one angle, which means that the actors were in front of it. In hindsight, you kind of want to go to the filmmakers and say you could have moved the camera. It would have been a much more interesting film had you just picked up the camera once in a while and caught something from a different angle. In hindsight, it's wildly obvious to us.

And so there's a general question when.

We think about today, what will our future selves say was obvious to us that we're just not understanding? And maybe for this question we'll think about Netflix and the movie TV watching streaming industry. What in the future will we look back on and say, well, I can't believe they didn't move the, move the camera a bit.

Yeah, I mean, for me personally, my career advice to myself is always the same, which is experiment more, experiment faster and experiment with bigger swings. Right. And generally be much more willing to take on risk. So in second half of my career I've understood that. And then the other thing is the balancing act between strategy, like we got to figure out what to do and just doing it. And for me, I've sort of learned this balancing act between discipline and chaos, where discipline is the stress. Strategic thinking. I sort of gave you one model, the delighting customers in hard to copy margin enhancing ways. But the chaos is you don't know what the results are going to be of the experiments and frankly some of the most interesting and impressive results, the innovations that come out of the ooze, you know, the, the mud in all of these things. It's really, I've been able to engage in this consumer science process with a B testing about the last 20 years. And it just keeps encouraging me to experiment more and faster, take bigger swings and take on more risk. And I think 20 years from now we'll see some of the results of that. Maybe we'll be vaguely smarter about consumer science and developing consumer insight. Maybe we won't be surprised as often, but I think that's going to be at the margin. And then I'm hoping that 20 years from now, when you stoke up Netflix, there's just going to be one thing watching Because Netflix has magically predicted to its personalization technology that just one thing, Kristen, that you want to watch tonight. That's my hope.

That was Gibson Biddle. This guy is speaking my language. Test and learn the power of seeing what people actually do versus what what they say they will do can be better for customers and for business. When Netflix switched from a five star rating system to a simpler thumbs up, thumbs down members grumbled about the change, but the new system doubled the number of ratings. They made it easy to do the behavior they wanted to encourage rate the things that we watch on their platform. Yes, this is genius. But this is also the main ethical dilemma facing Netflix right now. The easiest thing for us to do all day is to watch Netflix. Heck, I'd be home binging Bridgerton right now and not recording this show if that was possible. But is that what we really want to do in life? Of course not. Based on my long term ambitions, my desire for meaning and purpose, and a love of type 2 fun, I want to record this podcast. That's my passion. Not watching Bridgerton for three hours. But how will Netflix balance this? How will any social media company balance this? Will TikTok or Facebook put in stronger nudges that get you off their platform? And outside, will Netflix put in any friendly alert if you've been on the couch for multiple hours in a row that it may be time to at least stretch, if not turn off the tv? Maybe they will if they believe it is a costly signal that will increase overall brand trust. But so far it seems unlikely. Netflix is leaning into the autoplay of life. They once made special socks that could detect when a user fell asleep and paused the show so that the viewer didn't miss out on anything. Known as Netflix Socks, this pair of socks comes with a sewn in electronic device that detects your pulse. So Netflix is going hard on the binge. Beyond worrying about their investment in niche footwear, I am concerned that this difference between revealed preferences and meta ones will autoplay us into watching more TV and movies than we really want to. Don't forget to like and subscribe to the Science of Change on Apple podcasts or anywhere you listen to podcasts. And while you're at it, why not give us a review? Writing a review takes only seconds and is wildly helpful for the show, and we'd love to hear what you think.