Invest Like the Best with Patrick O'Shaughnessy - Shishir Mehrotra – The Art and Science of the Bundle - [Invest Like the Best, EP.175]

Episode Date: May 26, 2020

My guest today is Shishir Mehrotra and the topic of our conversation is the bundle: offering access to multiple products, services, or providers for a single bundled price. This topic is full of incorrect pre-conceived notions, and as it turns out, the bundle is one of the most powerful ideas in business. Properly harnessed it is good for everyone involved. Shishir explains the ins and outs of bundles in this conversation. Shishir ran product at YouTube for years and sits on the Spotify board of directors. He founded and now leads Coda (which is “A Doc” spelled backwards) in 2014, to bundle together productivity apps like docs, spreadsheets, databases, and applications. I love this wonky, detailed conversation which has me thinking differently about many businesses and business strategy. Please enjoy.   This episode is brought to by Koyfin.   For more episodes go to InvestorFieldGuide.com/podcast. Sign up for the book club, where you’ll get a full investor curriculum and then 3-4 suggestions every month at InvestorFieldGuide.com/bookclub. Follow Patrick on Twitter at @patrick_oshag   Show Notes (2:08) – (First question) – The arc of his career (3:32) – Why he has an interest in bundling (7:45) – The concepts of superfan, casualfan, and nonfan businesses (11:05) – Using Spotify as an example of bundles (13:24) – The first myth of bundling: Bundling is bad for consumers (17:53) – The second myth of bundling: 1st vs 3rd party providers and the bundlers (23:03) – Low usage but high Marginal Churn Contribution (MCC) business (24:26) – How insurance fits into these models   (26:37) – Myth 3 of bundling: How this impacts consumers (32:12) – How marginal costs play into the thinking of bundling (34:54) – Myth 4: Bundling things that have nothing to do with each other (39:51) – How bundling companies can apply this into their product development (43:21) – Strategic advice to companies building bundles (49:01) – How price and pricing power play into advantages for certain bundlers (54:16) – How does bundling play into his investing thesis (56:47) – Most interesting bundles he’s observed             (58:44) – Eigenquestions: The Art of Framing Problems (59:14) – What the future of this trend is (1:02:24) – What is an eigenquestion (1:06:29) – Kindest thing anyone has done for him   Learn More For more episodes go to InvestorFieldGuide.com/podcast.  Sign up for the book club, where you’ll get a full investor curriculum and then 3-4 suggestions every month at InvestorFieldGuide.com/bookclub Follow Patrick on Twitter at @patrick_oshag

This episode is brought to you by Coifin. I've become very interested in the best software tools in investing. And when I asked Twitter for the best Bloomberg alternative, the overwhelming winner was an excellent new product called Coifin. It's a web-based platform that lets you analyze stocks, ETFs, mutual funds, and other asset classes in one place. I've been using it every day to track what's going on in the market, and I think if you try it, you will too.

Coifin has a ton of high-quality data, powerful functionality, and a clean interface. The best part is that it's free. You can sign up at www.coiffin.com. A-O-Y-F-I-N.com. Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, methods, stories, and of strategies that will help you better invest both your time and your money.

You can learn more and stay up to date at investorfieldguide.com. Patrick O'Shaughnessy is the CEO of O'Shaunicee Asset Management. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of O'Shaughnessy asset management. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of Oshonosi asset management may maintain positions and the securities discussed in this podcast. My guest today is Shishir Marotra and the topic of our conversation is the business bundle, offering access to multiple products, services, or providers for a single bundled price.

This topic is full of incorrect preconceived notions and as it turns out, the bundle is one of the most powerful ideas in business. Properly harnessed, it is good for everyone involved. Shashir explains the ins and outs of bundles in this conversation. Shashir ran product at YouTube for years and sits on the Spotify board of directors. He founded and now leads Koda, which is a doc spelled backwards in 2014 to bundle together productivity apps like docs, spreadsheets, databases, and applications. I love this wonky detailed conversation which has me thinking differently about many businesses and business strategy more generally. Please enjoy. Shashir, thank you so much for joining me today. The topic of a lot of our discussion is going to be an obsession of mine since finding your

work and thinking and reading on this topic, which is bundling and unbundling of products. It is something that probably most people will not have heard a deep dive podcast on, let alone considered in the depth that will go on today's conversation. But I just think it's one of the most interesting business angles I've uncovered in years. So first of all, welcome. And thanks for your time today. Thanks for having, Patrick. So this topic of bundling is something that you certainly seem to have thought more about than anybody.

I thought a fun place to begin would be with what I would call lightning round context. If you wouldn't mind describing sort of the course or the arc of your career in like two minutes or less, it will provide useful framing for the conversation on bundling. Sure. So I'll do it going backwards. So I'm currently the founder and CEO of a company called Coda. We make an all in one doc, blends together, dot. documents, spreadsheets, presentations, and applications into a single surface.

Anyone can make a dock as powerful as an app. Before that, I spent about six years at Google. I was responsible for the YouTube products, joined in 2008, pretty soon after the acquisition, and was there until 2014. It saw massive growth and value creation in that process. Before that, I spent about six years at Microsoft, worked on Windows, Office, and SQL server, and started my first company in the middle of the dot-com boom called Centrata before that.

That's me, in a nutshell. Fantastic. So tell me how you can. came to this issue of bundling, the sort of structure of our discussion of bundling will be around four myths that you've identified. But first, I'd be curious just to hear why you care about this issue. How did you come to learn about it? For your listeners, I'm trying to figure out the best way to express the kind of hobby obsession relationship I have with this topic because it's not typical. It's maybe the easiest way to put it is my wife will catch me sometimes, especially nowadays,

where we're like all in one house and talking all times like, oh, he's talking about bundling again. So it's something that that is top of mind for me a lot and infects a lot of how I think about products. It's how I think about the very specific literal topic of bundling, but also I think as we'll talk about as we go through it, it's really a philosophy for how to think about anything that can be broken out into parts and anything that can be put back together. My journey with bundling, and there's this paper that I think we're going to publish alongside this podcast called Four Myths of Bundling. The reason it's called Mitz is the journey starts from me having just joined YouTube in 2008.

And in those early days, it was pretty clear that YouTube was going to be an ad-supported business. And where we put 98% of our focus was to go make advertising work. And we had a lot of work to do. When I joined, we were losing a ton of money. The product was pretty misunderstood. It was not at all obvious that it was going to be a big business. And we always had in the back of our minds this idea that maybe we will have a user-paid, offering at some point and we'll build up a subscription model. We ran lots of experiments. I mean,

dozens of them. But it was always kind of like priority number four, five, six, seven. It was never the top thing we were working on. And the experiments, none of them worked. Some of them were like famous failures. We had one where we worked on it. I put two engineers on it for almost six months. They built this thing. We shipped it and it made $100. Not like $100 per person, $100 per country. Oh, literally $100. What was that thing? I mean, in retrospect, is like one of the dovest ideas out there. It was, it allowed you to pay for a video that you've already watched in order to download it, just that one video.

And basically, at the time, like, there was just no reason to do it. Like, I just watched the video. I don't mean to belabor that one was a particularly bad experiment. We had some pretty good ones, but none of them worked. And in that process, what I discovered, sort of had this offsite, we set back and we talked about it. And what I discovered was every experiment we tried had this implicit assumption. And it was often, you had to, like, kind of uncover and unpack all the

different reasoning that went into it, but it had an implicit assumption that we should do something in a subscription or user paid model, but it should not repeat the mistakes of the industry. And in particular, the assumption was that bundling is bad. And it was so fundamental, it's hard to even start to question it because I sometimes get talks on this topic and I'll ask a room full of people. When I say the word bundling, what do you think of? And especially in American audiences, the thing that immediately comes to mind is cable. They generally say Comcast.

And people's impressions are definitely not positive. Their view of it all starts from the source of negativity. And at this time, recognizing none of the things we're doing are working, and I step back and thought, maybe we got this wrong. Maybe there's something more to this than might be visible. And I just started reading and talking to everyone and everything I could to learn more about it. And in that process, I gradually accumulated a list of, and sort of one by one, it's now a list of

four of things that I used to say that turned out to be not only not true, but almost the exact opposite was true. And so I call those the four myths of bundling. Since then, that affected a lot of our strategy at YouTube. Then when I left YouTube in 2014, this became a sort of ongoing hobby obsession for me. It forms in my sort of core job, Coda is in some sense as a bundle. It's a all-on-one document, pulls together different products. We can talk more about that as well. I'm also a pretty reactive investor, advisor of companies and so on. So I've invested a number of companies in the space, some that are well-known, quite obvious, some that are a little bit earlier. Probably the most famous of that set was Spotify. I ended up joining up with Daniel pretty early on, and sit on the board

there now. And then gradually took this idea and refined it and turned it into a little bit more of a thesis. I think it's fascinating. I think it's intriguing. I don't know if everyone else thinks that, but I certainly do. And we'll talk more about why I think some of the implications of it are so interesting as well. One of the most interesting concepts, which also is a necessary set of terms for people to understand before getting into the detail is this idea of a superfan business, a casual fan business, and a non-fan business. Can you describe what each of those things mean and maybe give an example or two of businesses that would fit in those three categories because those sort of form an important base of bundling? Yeah. So I'm going to ignore the word

business for a moment and focus on super fan, casual fan, non-fans. So let's imagine, imagine I have a product and I can offer it out to the world. And let's say I have three of them. And I have a choice between having them be a la carte, have people pay for each one, or having them be done in a bundle. So superfan, casual, fan, and non-fan is a way to describe three audiences for each product.

So superfans hold two characteristics. Number one, they would pay retail for the product. They think that's a fair price. And number two, they have the activation energy to go find it. The second criteria is really important. But that defines the super fan of a product. I'd pay for the retail price, and I would go out and find it.

casual fans lack one of those two criteria. They might not pay that retail price or they don't have the energy to go find it. And non-fans ascribe either zero or sometimes negative value to a product. Negative is a particularly interesting case we can talk about later. These are the three different audiences. And this is actually leaking a little bit into myth one. We'll come back to in a moment. But if I have these three products and I can decide do I want to sell them all a cart or sell them in a bundle,

if I sell them all a cart, then by definition, the set of customers I get are super fans. For every product, I get the super fans as customers, and I miss out on casual fans. If I sell them in a bundle and I take the three products and I put them together, then the value that's created is all casual fan value. And this is particularly interesting, and we'll talk about it in a moment with Myth One, but if you think about that from the perspective of your question, super fan, casual fan, non-fan businesses, superfan businesses focus on the first set. They focus on delivering products to people, directly getting payment from people who ascribe value to that product and seek it out.

And there's a lot of great examples of that. One, Uber is a pretty good example. I pay for Uber's when I get into Uber. If I stop going out, as I have been for the last two months, Uber makes no money from me. There's no way. There's not really a way to be a casual fan of Uber. Non-fan businesses on the other end of the spectrum tend to be ones where the way that they produce value and monetize is with people where the payment relationship is not direct.

The most obvious examples of these are the ad-supported businesses. Most of YouTube fit in this category back when I was there nowadays is a pretty good paid description there as well. But a lot of the social networks fall in this category as well where the way that the relationship with the user is monetized is not direct with the user. Most common form of that is to monetize their time, which is what advertising businesses, media businesses tend to do. Casual fan businesses sit in the middle. They are businesses that provide access of their products to a variety of users, some of whom are super fans and some of whom are not. and the thing you're paying for is access to that good. And there's tend to be subscription businesses.

Cable is a very commonly used example of that. You pay for a lot of things in your cable bundle that are things you might be a super fan of and you would have paid for even if they were separate. But much of what you're getting in that service are things where you lack either the willingness to pay retail or the activation energy to go find it. So these are sort of three different types of businesses. Could you just maybe use Spotify since you sit on the board there? And it's a departure from the example of, say, Comcast.

which you understand why everyone thinks of that as a cable bundle. Talk about how that manifests in a business like Spotify, where there may be a la carte buyers of one artist that you love, but that that is very different from the value proposition of that business. So just use that example maybe to finalize the fleshing out of the concept. Yeah, and this also teases a different concept, which is the often totals all the way down relationship of bundling, that many bundles are themselves bundles of bundles of bundles of bundles.

And that's certainly what we see in the case, of Spotify. So what Spotify displaced, then once I displaced piracy, and Daniel gives a lot of great talks about that, from an economic perspective, what it was displacing was direct purchasing of music, which was clearly a superfan business. And for a while, that was in a record store. You're going to buy a record, gradually moved to digital, and you paid 99 cents per song on iTunes. And what Spotify offered was a way to pay for all the music you might want to listen to, even if you're going to get access to some things that you might not have gone out and bought individually.

And that offering turned out to be incredibly popular. To use that terminology, what it allowed was it allowed for every consumer to all of a sudden get access to goods that they were a casual fan of. And that turns out to be very powerful because the set of products for which any individual would pay retail and have the activation energy to find it, that set of products tends to be very narrow. But the set of products that you might find utility in and you might find pleasure in or entertainment in or whatever it might be tends to be much wider than that.

So at one level, Spotify bundled together, 99 cent purchases of music into a $10 offering. At another level, if you watch what Spotify is doing now, especially with podcasting, is producing a bundle that is really the audio bundle. And Daniel gives a really good talk about the value of the eyes and the ears and so ony has a really good analogy for it. But core idea being you can apply the same principle one step up and say it's not just about all the music you might listen to. It might be all the audio you might listen to.

And then beyond that, Spotify is built a. a set of bundling relationships across other companies as well. And one of the most popular ones is the Spotify students here. It comes with Hulu and Showtime as well. There's a number of similar offerings that have been made across the portfolio. But those are sort of three different levels, bundling up music, bundling up audio, and then cross-bundling across products and services. Let's return now to the structure.

So introduce us to this first myth that you encountered and why almost the exact opposite is true. Yeah. So a reminder on the terminology myth. The way this paper is written, and I expect some of you, Your listeners will want to read the details. I'll try to cover it here, but there's a lot of, there's a lot of fun in the details. It's written as a conversation between me and what I call the myth maker.

And the myth maker, I think each reader could picture someone. Everybody knows someone who says bundling is bad. These are all the bad things that come out of it. For the purpose of when I was writing this paper, I mostly just pictured myself. So these are all things that I said 10 years earlier. That's the way to picture it. And the first one written in this paper is the most superficial and the highest level one is the blanket statement.

Bundling is bad for consumers as well as providers. And the easiest way to understand this one is if you just start with the scenario we're described before, I have these three products. I can sell them each out of cart or I can sell them in a bundle. What should I do? And what most people see in that, we sort of ran through the example. If I sell them all of cart, every product, every provider only has access to super fans and every consumer on the flip side can only get access to products that they are a super

fan of. If I have these products and I separate them out, that's the value created. If I put them in a bundle, then I produce value in two ways. For providers, I give them access to casual fans. And for consumers, I give them access to products that they might be a casual fan of. To use our music example earlier, what did Spotify do to iTunes? It allowed people to listen to music that they might not have gone out and paid 99 cents to buy. And that value turns out to be really important. And if you go look at the usage grass and so on for what people spend their time listening to, Some of it is things that they might have gone out and bought and bought it 99 cents or bought in a record store or so on, but a very large portion of what people spend time consuming are things that they value,

but they would not have met that superfan test for. So the first myth, and then each of these has a sort of thesis that goes around with it, is the way that bundling produces value is not by producing more super fans, it's by producing casual fans. One example I like to use for this is, and some people like this example, some people don't, is UFC, the Ultimate Fighting Championship. Are you a fan, the UFC? I'm not your description of it, so I'm aware of it, but I'm not an active watcher of it. Right, right.

So interestingly, I sometimes give this talk in a room of like 100 people, I ask this question. And generally, somewhere between two and five people raise their hand and say, I'm a fan of the UFC. And generally those people raise their hand, not a little bit in the air, but like straight in the air. Because I'm sure you know someone like this. This is for the people who are into Ultimate Fighting Championship,

it is a deep obsession. Some of it, there's lots of reasons why it's perhaps a narrower audience, but I think one of the reasons is the business model. You pay 50 bucks per a fight on Fridays. You invite your friends over and you watch. If you ask the same question of people who are into the NFL, say, are you a fan of the NFL? That's the same group of 100 people.

And you'll get hands raised at all different levels. Some people will say, yeah, I guess I watch a Super Bowl, so I'm a fan. Some people will say whenever my team's winning, I watch, all the way to the person at the end who says, I have four different TVs, and I watch every game simultaneously, every Sunday. My Fantasy League is my obsession and so on.

And one observation of that is the NFL is very widely distributed. And most of those fans can't really describe for you how they pay for the NFL. They don't actually know. It comes through my bill, my Verizon bill, my sum of it seems to be free. I'm not really sure why. And that reason is that that as a business focused heavily on casual fans. They focused on distribution to casual fans. So backing all the way up, the first myth is the sort of highest level one.

We'll get more practical in a moment is I think that people will say bundling is cheating, both consumers and providers. The piece they're generally missing is that one done well, the way bundling produces value is by giving access to and revenue from casual events. I guess where my mind would go would be to say I pay this however much per month for my cable bundle and I only watch three channels. This is nonsense. I wish I could just pay those three channels directly. That sort of thing is sort of the reason why people don't like this concept. But I think what you've made clear is, I think the NFL example is perfect, that if you take away, all experience of the NFL from everyone except for the four screen fan. That seems like an obvious

negative. Right. Right. Exactly. And the positive is because of the bundle. Yeah. And by the way, the example you just gave of cable and why can't I just pay for three channels, we should pause that one, because that observation is very sharp on the consumer side. And that's what myth three is all about. We'll get to that in the second. Awesome. Awesome. So the second myth is really about more the business model of this on the other side, not the consumer side so much, but what we'll call the provider side. So first, I'd love a distinction between if relevant, third party providers and first party providers. And then why there's this kind of weird myth around who should get what share of the pie of the revenue that's generated from selling the bundle to consumers. Yeah. Okay. So let's do the definition part first. So providers, consumers, bundlers, most of this theory will sort of divide the world and try to treat them neatly as divided between the provider, the consumer, and the bundler.

Provider provides the good. The consumer consumes it and the bundler puts them together. It's not always that clean, as we gave an example earlier, sometimes the providers are bundlers themselves, and so there's lots of different ways that that works, but let's pretend it's clean for a moment. The third party versus first party thing is interesting. And one of the conjectures of this paper is all the examples given tend to be, as you described, a third party bundles. The bundler and the provider are different corporate entities. There are many cases of what you might call first party bundles where I have multiple products but produced by the same company. say Amazon Prime is a pretty good example. Amazon has Prime, it has Amazon Music, and so on.

New York Times is a pretty good example, too. They have not only the New York Times subscription, but they have a crossword subscription and a food subscription that goes along with the two. They bundle us together. There's also just one level deeper, to this third party, first party. There's also inside of a single product

you might call intro product bundles. And from that perspective, like Coda is a pretty good example, where the bundle is a bundle of products. It's a document, a spreadsheet, an application, all in one. So multiple different ways that this can be applied I'll describe the terminology all in terms of third-party bundles and third-party providers, because it's easier to understand if the entities are separate. But I think we should come back to because I think for a lot of listeners, that may not be

the most practical application of this. It might actually be the inside a single company or inside a single product. But, okay, so myth two. So, and this one, like you said, the myth two focuses on providers, myth three focuses on consumers and myth four will come back around to the bundler. So myth two, this one may seem odd to put as the second one. As I've thought about the ordering, understanding this one is the sort of key to understanding the next set.

And the way this generally gets set is a small sentence. What does the myth maker say? The myth maker says, hey, okay, maybe I get your whole thing about casual fans creating value and the NFL example is a pretty good one. But come on, the bundler is clearly cheating everybody. They're paying everybody, all these arbitrary rates. They should just pay people fairly and bundling would work better. And again, I'm putting the consumer side aside for a moment and just focus on that.

That said all the time. People look at it and say, how does Comcast get away with paying ESPN $4.50 per month and history channel 20 cents per month doesn't seem fair. Now, when people use the word fair, they mean lots of different things, but I find that most times what they mean by fair is by usage, especially in Silicon Valley. That's the, it's almost a direct proxy for that term. The myth here, to give the full sentences, the myth is revenue from bundles should be allocated based on usage. I used to say this all the time. We're at YouTube, building up these different products and saying we're going to build this paid offering,

we're going to build a better bundling model, and it's going to be more fair, and what we meant by that was we're going to pay based on usage. And as I spent more time thinking about it, I realized not only is that incorrect, it actually can lead to almost the opposite behaviors of what you really want. So let me describe a little bit one way to think about it. So there's a little diagram in the paper that just plots out usage for some different properties. I'm going to use cable as an example, because it's very familiar. But if you look back and when this paper was written the first time, we pulled some stats on this, History Channel and ESPN.

So I'll use these two products as an example. History Channel and ESPN get about the same amount of usage. If you look at it in terms of time spent, rating points, so on. It's very close. And yet, History Channel makes about 20 cents per month for a subscriber and ESPN makes somewhere between four and five bucks per subscriber per month. So what is the second axis? What is that price axis correlated to?

Now, usually you would, if I go ask people in the industry, the term that generally gets handed back is anchor value. That axis corresponds to anchor value. Anchor value is a word and a term that I think you use when the economic phenomenon is not yet clear. It's kind of a made-up term. It's not very descriptive. And so instead, we started using this term called marginal turn contribution, MCC.

And it's a pretty simple idea. That axis, what anchor value represents marginal turn contribution, is if I were to remove this one product from the bundle, what percentage of my audience would churn? Pretty easy to understand. If I remove ESPN from the bundle, how many people would churn? And there's been a bunch of studies on this, but the reason why ESPN is paid 20 times more than History Channel is if you were to pull ESPN from the bundle, 20 times as many people would churn as if you were to remove history channel. That core idea that bundles should divide up payment, divide up the revenue based on MCC as opposed to usage leads to a totally different way of thinking about the bundle.

The paper actually even walks through some formulas for how to calculate this. You can formulaically arrive at what a fair price would be, what we call wholesale prices, is how much you pay out to each provider. And it is correlated mostly to this term, the marginal turn. I think this concept is really interesting like this two-axis chart where usage is on the X axis, MCC is on the Y axis. It lets you plot all sorts of different businesses. So what would be an example of, say, a very low usage but high MCC business?

You could draw a diagonal through this chart. We go back to your very first question, like what superfan businesses, casual fan businesses and so on. So, okay, so let's start where you start in the top left of this chart. Low usage, high MCC, high anchor value if you want to use that term. The typical example I'll give is a sporting event. You go to a sporting event, you pay hundreds or thousands of dollars for an hour or two of entertainment.

That's at one extreme of the usage is very low. but your sense of utility is very high. So that's in one corner. And the other corner tend to be, there's a lot of products where there's a lot of usage, but not a lot of MTC, not a lot of anchor value. Much of YouTube bit in this category, people would use it for hours and hours a day. But you push people on if you had to pay for it, what would you pay? If we took it away, how would you feel? And you get a wide spectrum, of course. It's a widely spread business, but a lot of it would fall in this category of very high usage and low MCC. If you draw a diagonal through this chart, what you'll find is sort of above that diagonal are businesses that naturally

lead towards casual fan or super fan business models. And below that diagonal are ones that lead to what I call non-fan business models, advertising being the most obvious of those. Where would something like insurance figure into this way of viewing the business world? Insurance is such an interesting example. So insurance is a bundle on many levels. And you didn't specify, but I'll assume health insurance for a moment. Health insurance added sort of basic offering, what is health insurance is the bundle between sick people and healthy people. And at the next level, what is insurance? Well, it tends to get bundled together with other forms of insurance. So it gets bundled with dental and disability and so on. At another level, health insurance gets bundled

with, depending on your country. In the U.S., it gets bundled with employment. And many countries, it gets bundled with citizenship. And you get it as part of that country. At a sort of core level, if you think about usage versus anchor value, health insurance is like the perfect example. It's a thing that you pay, you know, you don't want to use. Yeah. You don't. You don't. never use it. You pay for it and you hope you never get sick. And when you do get sick, you're up in the top left. The question gets asked a lot about why don't people just pay for what they need. And health insurance is like the perfect example because you can't. The thing that I think people miss about it, I'm going to try to offer very little political commentary. Like I don't

think this is really, I have opinions on it, but I'm going to ignore all that. The reason health insurance works is the cost for the superfan is ridiculously exorbitantly high. And so the only way for that product to exist for the whole healthcare industry to work is to bundle healthy people with sick people. People need to contribute even though they're not using it or the whole system doesn't work. When you have a catastrophic issue and you spend a few days in the hospital or a few weeks in the hospital, it's millions of dollars. And no practical person can put away enough money to deal with that contingency. This is not about bad planning. This is about bundling. That's why that industry works. And you can now ask the question of should it be bundled with employment or should it be

bundled with citizenship. Those are totally reasonable. But should you bundle healthy and sick people together? Go find any country where insurance is not prevalent and just look at all the rates of illness, of mortality, and so on. And it's an order of magnitude. Different health insurance works. That is a very good example of a product that sits in the top left of this quadrant. It's an interesting one because, again, since we first talked, you start to sort of see all of the world at some sort of bundle. And the notion of a business that you never want to use but definitely want to pay for is strange at first and then insurance fixes that strangeness very quickly. Right. Exactly. Right. Talk a little bit now from the consumer side. So that's sort of maybe

the more proper way to compensate providers, whether that's first party or third party, is via this turn metric. Talk now about the consumer side of things in Myth 3. Okay, myth three. So you already gave the myth. The typical way this gets framed is, I've got this bundle. I only want to pay for these products. And Cable is used as a common example for this. The myth maker in this case, maybe they've fallen along so far. Okay, you produce value through casual fans. Maybe I can buy this MCC thing that seems like inside baseball. I kind of cared, but whatever. However Comcast divides the money is up to them. But come on. A bundling is clearly bad because what it represents at its core is a lack of choice. That's the myth. What is a

bundler doing? They're ripping off consumers by presenting a lack of choice. In order to get product A, I have to buy product B. So just as a sort of simple example, to use is I'll use McDonald's as an example. And imagine you walk into McDonald's, you look up at the board, what goes through everybody's head when you look up at that board. So first you look at it and you say, okay, I kind of want a Big Mac. I definitely want some fries. And if I get those two things together, then the drink is free. And the person right behind them, what do they say? I definitely want a Big Mac. I really want a drink. I guess the fries are free. And the person right behind them has a crying kid, looks up at the board and says, ooh, there's this Happy Meal thing that comes with a toy, and then the whole damn meal is free.

The interesting thing about that, and everybody's seen that happen, looking up this board, and everybody's looking at the same package and looking at it differently. And to all of them, it looks like a good deal. There's one fundamental reason why people think bundles are rip-offs when they're not. And that reason is for a consumer, because the thesis in this section, is for a consumer to properly value a bundle, there must be a transparent and reasonable, a la carte price for every product in the bundle.

I get asked a lot, having been media industry and so on, what's my opinion on what's happening with cable industry. Cables going through unbundling and so on. My view is that's not actually what's happening. Through a variety of mostly non-intentional choices, over the course of the last 40 years of the evolution of cable, what we've landed on is an incredibly non-transparent a la-car pricing mechanism. Nobody has any idea what the cost of the individual offerings inside the cable bundle would be.

And so they misattribute it. And so they presume that if I were to rip out parts of the bundle, I'd get it at this cheaper price. So the reason I put these two in order, we talked about myth two and mid three, you would think these were very disconnected. In order for me to properly understand the value of a bundle, I need to get a la carte pricing. Why does McDonald's work and Comcast has trouble with this? Because Microsoft puts a price next to the Big Mac and next to the fries and next to the drink. And you can, they're transparent, they're reasonable. Some people buy them. But you now have an understanding

of what that bundle feels like. And when you say, what you just said is if you survey people about what they like and dislike about cable, it is easily the number one answer is I just want to pay for the channels I want. And then you'll ask them. You'll say, okay, let's take an example. I just want ESPN. Great. What would you pay for ESPN? You pay 50 bucks a month for cable. What would you pay for ESPN? And generally what they'll say is, well, I read some article that says that ESPN gets five bucks a month from Comcast. So how about I pay a little bit more than that? I'll pay like seven or eight bucks and I'll get just ESPN. Why can't I do that? That seems like really clear. Okay. So why do these two myths connect? And what's the problem with what's happening here? And by the way, the punchline of this is, I think what we're

seeing happening with cable is we're going to see this unbundling and what we're going to see is transparency on all of car pricing and then we're going to see a massive re-bundling. As soon as people understand what the components of this bundle cost, we're going to see them all put back together in this good way. And there's lots of reason why I think that the bundle today is broken. I do think because we lost track of this, we'll get to this in mid-four too. Because we lost track of this, the cable industry has not been able to realign their bundle the right way.

But I think what we're seeing is transparency of all the car pricing. Okay. So you go through the scenario interviewing this consumer and you say, hey, what do you want from cable? So I would just want to pay for ESPN. Okay, great. What would you pay for ESPN? It said, I'd pay a little markup over the five bucks that I think they're getting. I go back to Myth 2.

So how is this working in myth 2 is, well, the way I divide up money in a bundle is by marginal churn contribution. And I take my $50 bundle and I divide up the money by what percentage of people would churn if this product were removed from the bundle. And there's an equation in there that I won't go into detail here. But the net impact of it is that there's a relationship. between the wholesale price of a product and the retail price of a product, and it's proportional to what we call the superfan percentage. So the wholesale price of ESPN inside of the bundle is

five bucks a month. Let's say the superfan percentage for it is 10%. Let's say that it makes that very crisp. Imagine that if you were to remove ESPN from a bundle, 10% of people would churn. What that implies is the correct retail price for ESPN is $50 per month. That's what the math implies. So now you go back to that consumer and say, great, you're paying 50 bucks a month for Comcast, what would you like to pay for? Well, I guess I'd like to pay for ESPN. Great, you can pay 50 bucks for ESPN. Now all of a sudden cable looks like an amazing deal. I pay 50 bucks for ESPN and I get $299 other channels for free. It's better than McDonald's. This is a really good deal. But because we don't have transparency on each of these components, where it lands is that

people think they're getting ripped off. And so this myth is quite important. And when you think about it, so that's cable. If you think about it for people trying to apply this, I get asked a lot about how I'm going to produce this bundle. Should I offer the product separately too? And my answer is definitely yes. Even if you don't think that's going to be a primary thing people are going to buy, you're going to make the bundle such a good deal because you think that that's really what you want people to buy. People's ability to understand a bundle starts from their ability to understand the components of the bundle. That's what MET3 is all about. Can you say a little bit about how marginal cost or cost of goods figures into this whole framework? So in the examples of, say,

a Netflix or a Spotify or a cable bundle, it doesn't cost them anything extra to distribute some of the product from the providers to the marginal user. So enormously high gross margins. Obviously, some things are physical bundles like McDonald's food. There's other bundles where there's more marginal cost as a percentage. How does that figure in to the thinking here? For sure, there's a section in the paper in the end about that. And basically you just adjust to all the equation to put a minimum on it.

Your product is basically unviable if your wholesale price gets below the marginal cost of delivery of that product. And so if I have a $5 value meal at McDonald's, and the marginal contribution of the burger is less than what it costs to make the burger, then you should just not include it. I'm not going to work. And it sets a floor effectively on what is the tolerable MCC for the product. In practical terms, I'll say, it's clearly I used a lot of media examples that all have

what you described as zero delivery costs, or not zero, but near zero delivery costs, marginal delivery costs. When you look at other bundles, the interesting thing is I think many people don't perceive the amount of markup that goes into marketing a product to a super fan. Go back to the definition. Superfans are defined by number one, willingness to pay retail, and number two, the activation energy to find it. Now, the activation energy to find a product is a function, not, I mean, it's a bidirectional thing. For me, I have this product and I need people to find it. I have to find a way to reach them. I pay in marketing, I pay in packaging, I pay in distribution. So when you boil down even a physical product and you look at the amount of true cause and the true cost of goods sold for the product, and you strip out all of the costs that are associated with attractive. a consumer. Just to take an easy example, a hardware product that ships a Best Buy is probably

their cogs are probably somewhere between a quarter and a third of what the end retail prices. Once you take out the Best Buy margin, the delivery costs and so on. So even for that product, it may not be obvious that bundling it together can be hugely productive. And I think we start to see that. I mean, your cell phones are a great example of that where you see cell phones get bundled with a service a lot because, yes, it costs a lot for the phone, but not as much as you think if you account for the fact that the cost of acquiring the customer is part of that superfan price. But yeah, so the specific answer is you do have to consider it. And there is a way in the paper to talk about it. But in practical terms, I've seen even businesses that you would think

be all physical. As an example, I'm a investor in a beer delivery company, you would think that all the cost would be the actual beer. But it's not. There's a lot that goes into providing a service like that. And so subscriptions are still pretty valuable in those cases. Amazon is an obvious example of that too. is one of the most popular descriptions in the world and clearly a lot of cost and delivery. So now we get to, for me, what is maybe the most interesting part of all this, which is the bundler side. I've got lots of questions here, but maybe we'll just begin with your Myth 4, and then I'll take it from there. Here's a preface for Myth 4. I think one, two, and three are a little bit definitional and sort of easy to get through.

Myth 4 has in practice turned out to be the trickiest in many ways. It's the trickiest. It's easy to understand, but it's the trickiest to apply. it's the most counterintuitive, but it is also the most powerful. And here's how the story is generally told is that the way the myth maker says it is, okay, bundles, maybe I buy casual fans produce value in bundles. And I'll also buy MCC is a better way to distribute money. And I'll buy that being transparent about component prices is a good way to establish value.

But why do bundlers keep bundling together things that have nothing to do with each other? And I'll tell a quick story of when this became very sharp for me. And my first inkling of this was, In 2012, I was at YouTube. One of the experiments we ran and were thinking about was a paid offering, and we got the opportunity to license NFL Sunday ticket. And this is, for some of the listeners that might not be familiar, this is an offering that DirecTV has had for years.

They were paying about a billion dollars a year for it. And Sunday ticket, the basic offering is you get all of the out-of-market games on a Sunday. And so if you live in San Francisco, you get to watch every game except for the 49ers to this offering. It's incredibly popular. It sells for 400 bucks. year. And if you're an NFL super fan, you not only likely buy this, you're likely only subscribe to Directive E because of this, which is back to MCC and what its value is to that bundle and so on.

Okay, so we have an opportunity to bid on this. Just to like reset the time period, this is 2012. YouTube is a good property, but not yet anywhere close to as greater property as it is now. For us to even be in the negotiation for this was not obvious. And we certainly didn't have any meaningful paid offering. Now YouTube is a much better paid offering. So we get into a discussion about it. And the first meeting, this is Larry, had Larry just taken over to CEO of Google. And so a lot of these conversations went through Larry's staff. And the first meeting, I bring this proposal in and say, hi, this is what we want to do. We think we can justify paying X dollars for it, which is more than direct to you was paying. And I thought enough to win the deal. And the first

conversation was kind of hilarious because we had to describe coming his what Sunday ticket is, here's how the NFL works. And the first question was billions of dollars per year for that. Why don't you just invent a sport? And this is like a typical Google conversation. So I lost a whole meeting. And there's some other approach to this. Sounds entertaining. Gets a laugh now, but it was definitely not fun for me. And then I come back a few weeks later, and we start from the basics. Okay, here's the NFL.

Here's how all the rights are divided up. Here's Sunday ticket. Sunday ticket is an important part of the NFL licensing model, but a very small part. The vast majority of the licensing fees goes to the broadcast networks and the cable networks and not to this one offering from direct TV. And I say, hey, this is what we want to bid on. And the obvious question is, are you going to get the rest of it? How are you going to get the rest of the deals?

And I said, look, first of all, you can't. I mean, these deals are locked up for sometimes 15 years. There's no way to go get these deals. And secondly, I have no intention of being comprehensive. I don't want all of them. And this was kind of shocking. Why would you do that? Again, back in Google mindset, Google doesn't do halfway.

Google Maps doesn't map just parts of the world. Map the entire world. You license football, you get the whole damn thing. You don't get part of it. And so this was kind of a travesty in that room. Like, what do you mean? You're going to offer football and you're not going to offer all of it? That doesn't seem to make any sense.

And so we go through, and I was at the time, the thoughts on bundling were very nascent. I don't think it had good descriptions of it at that time, as we do now. And I said, actually, the thing we want to bundle with, and I said, no, no, we're going to bundle it with other things. And they said, okay, so we're going to bundle with other sports. You get soccer or basketball, so on. I said, actually, no, I don't think so.

I think what we're going to bundle it is we have these channels that are really exciting into food space, and we have these other ones in great knitting channel that people seem to really like. And I think we can pull together the right bundle for people. And I got laughed out of the room. The deal didn't happen. And I'm not blaming the room.

I'm blaming myself. I don't think we explained it well enough to understand it. And it took me a while, I mean, probably years later, before I could describe this well. And the myth is the best bundles are narrow and have very similar products so that they make sense to consumers. But the thesis is that the best bundle is one that minimizes superfan overlap and maximizes casual fan overlap. And there's a little picture in the paper of these Venn diagrams. That picture by it probably describes it better than my story did, which is basically if you think about a product as having a superfan core and a casual fan exterior, what you want as you line up the overlap between these products, you want to minimize the course.

You want to minimize the chances that super fans overlap. If I'm bundling three products together and every person who already pays for all three products, when I bundle together, that's a net win for the consumer, but a net loss for the bundle. Every time I see casual fan overlap, I'm seeing a net win. Not only for the consumer, but generally for the industry, because these are often cases where this product had no market before them. And that's the theory. is that the best bundles maximize casual fan overlap, minimize superfan overlap, super powerful and can be applied in lots of different ways.

Talk about how that concept, which is certainly the concept that stuck with me after our first conversation, because it is super counterintuitive, Sunday ticket and knitting channels. It's just hard to imagine that that would work. But I think when you read through this and think through it, it makes a lot more sense. And some of the examples bear that out. How do you think about this within a company or within, say, building a product roadmap? So let's take Coda as an example. So you are a bundle of lots of different, we'll call it productivity apps that people use from typically different providers.

You're putting them all in one spot. When you're making a roadmap product decision, are you literally applying this idea so that the next best thing you could build would have a lot of casual fan overlap with what you already have, but bring it a distinct set of new super fans? Yeah. As you be clear, let's get out of the third party provider world and all the way into inside a single product. clearly when you're building a product, there's a piece of it which is addressing your superfans and addressing your casual fans. And if you have multiple superfan bases, you want to understand them and then understand the casual fan links between them. And you can't be 100% focused on either. Like if you give up, if the NFL gave up marketing and providing service to the forced green

Sunday person, they will eventually lose their market. So you have to build for your super fan base too. It's not a no things in life are easy. You don't get to just go work on the fringes of your audience. That said, Coda is kind of an obvious one because we're a bundle of, we describe it as Docs, tables, apps, and certainly a bundle of those different things put together. And you'll see it. You'll ask people, if you go ask a Coda user, we now have tens of thousands of companies around the world using Coda, lots and lots of different users. If you ask them what they're replacing with Coda, each of them will say something a little bit different. I mean, there's a whole class of people that will talk about, oh, Coda is just a better document. It's better than Word and Google Docs and

It's a better place to write, to a better place to express myself. There's a set of them that will compare us to what we call the table products. They'll compare us to Excel and Google Sheets and Smart Sheet and Airtable and so on. Say that's a better spreadsheet, database, the better way to manage my data, make sense of it. And there's another group that'll compare it to applications. And they'll say, I use Kota instead of buying a packaged system for managing my customers or my inventory or my task or whatever it might be. And Koda is a more expressive way for me to match my actual process than buying any of those packaged tools. So if we think about those different audio,

is the interesting part in our thesis is that that's actually one offering. The Cota thesis is that people don't want to switch products as they go between these different dynamics. And we are constantly watching indicators of, using this definition, casual fan overlaps. People who started with a viewpoint of, oh, it's just a better doc, did they figure out over time that, oh, I could use this for structured data, I could use this to build interactive application, they could use this to publish the web, or the other extreme people that showed up and said, hey, I'm going to replace package software XYZ with CODA, did they then start to use it

as a document surface as a spreadsheet modeling tool and so on. And so we're always looking for overlaps between them. And then you're shoring them up and you're saying this is where each of them each of them. That's a fairly clear one for us. I mean, I've seen it and many other products as well, as you define these are different cores. The easiest way to generally do it is like the way I just described as by market. You could do it by persona as well. I do great with the product manager and I do okay with the engineer and then the designer gets pulled in the middle. Or I do great with the parents, but not with the kids. Or I do great with the teachers, but not with the students. And so you kind of start to build up these bundles of relationships. But yeah, I think the same philosophy applies

inside a product. Can we talk a bit about business strategy of a bundler? And so you're on the Spotify board. So in the business of providing strategic insight on something like this, it would seem to me like if you were a bundler, and again, it's a media example because it's convenient, but the same concepts should hold, that what you would want is to get a diverse enough set of providers on the platform that no one provider holds any power over you, whether that's market fragmentation, how you want to look at that, or if you use Spotify as an example, there's some huge usage. I don't know who they are, but there's a top five most listened to musical artists on Spotify that my strategic goal would be to keep adding stuff to the single bundle such that even if I

ripped one of those top five out, I wouldn't lose the big fans of that musician or something. So how do you think about this as building strategic advice for businesses thinking through bundles? So I'll give an analogy to this framework. So when people talk about supply and demand curves, and many people have gone and read Adam Smith's work and he talks about the butcher and the baker. And the way that supply and demand works is that the butcher and the baker actually, they don't have to ever have taken an economics class to have supply and demand work. They just just have to follow their natural incentives, and it just works. So the reason supply and demand curves cross, and that's where prices equalize and so on, is because everyone just does the best

thing by their business, and the right outcomes happen out of that. So in any bundler and provider relationship, what are the incentives? The incentive of the bundler is to minimize superfan overlap, maximize casualty fan overlap, and grow their bundle. So it's not really, fragmentation can help, but by itself, fragmentation may or may not lead to lack a superfan overlap. I I mean, the key is you want to be building a product that minimizes the cases where I'm paying twice for the same person. And so from that perspective, breadth might be a better description than fragmentation. And so if you look at, for example, what Spotify and Daniel have been doing with podcasting is a clear example of we went out and looked and said, people listen to things all the time.

But the percentage of time they spend on music is a lot, but not all of it. And how about the rest of it? And it turns out many of those people are busy listening to Patrick Podcast. They'd be really good for that base. And so in Spotify's case, I'm not sure fragmentation, the music space is so fragmented and is literally millions of artists, that the fragmentation piece of it is not quite that relevant, but what is really relevant is the way the business grows,

is that it starts to take on a value proposition that broadens its base. On the flip side, from the provider's perspective, what is your job? Your job as a provider is to maximize your MCC in this bundle. And that's not a bad incentive. Your job is, to put it really simple, your job is to deliver a good product. And if you think about cable, negotiations is a good example of this. There's an example, and they're about the Weather Channel and

DirecTV, and that was a very public, I mean, there's lots of them, but that one was a very public example of a thing we've all seen, right? What happens? The channel is renegotiating with the cable provider. They get notice, hey, we want to renegotiate, and then the public war happens. And the messages go out to everybody saying, in that case, it was, Weather Channel came back and said, they put out this press release that said, Direct TV will lose 1.6 million subscribers if they drop the weather channel. Keep in mind, I think DirectTV had 20 million subs at the time. They're going to lose 8% of subs. Seems like a lot. Direct TV then puts out a notice saying, hey, these guys are trying to overcharge us. They want so much more money and so on. And then what

do they both do? They sit and wait by the phone and they see how many people call. And it turns out if you run the math, like it didn't need to be anywhere close to 1.6 million subscribers lost. I think actually if you look at where the actual wholesale price ended up, all they needed was about 20,000. They to justify about 20,000 to 40,000 subscribers churning to agree to what the Weather Channel wanted in their demands. And so what was the Weather Channel's job? The Weather Channel's job in that relationship, Direct TV's job is not necessarily fragment or not fragment. It's to pay fairly. It's, hey, I want to pay as close as possible to the marginal turn contribution of this product inside this bundle. And then from Weather Channel's perspective, it's, hey, I need another 40,000 people, love my

product. And love it so much that if it didn't exist in this bundle that they would churn from the whole bundle. That's a very high bar. I got to cross the casual fan bar into the system. superfan bar in order to do that. That incentive, in my opinion, leads to very good products. The same way that capitalism leads to good products. Having to meet a supply and demand curve, what does it do? It forces you to drive cost down, drive value up, and you get better products to your consumers. The same thing here. The incentive to increase superfan percentage for your product for a provider is a great incentive. And from a bundler's perspective, it's an incentive you want to encourage. So I would say, what are the, so you ask the question, what is the strategic implication

for bundlers? And for bundlers, I think it's you want to assemble a bundle that minimize Superfan overlap. You want to keep increasing the scope of your bundle, in particular, into new areas of superfans. And you want to create the market where the set of people inside your bundle have an incentive and an ability to continue to increase their value to your users. You don't actually want it to decrease. You want them to build better and better product. You want musicians putting out better songs. You want podcasters putting out better and better podcasts and knowing that they'll get back a return and value for that. And that works. I mean, the cable industry is a perfect example that as much as people love to hate on the cable industry, I think generates hundreds of billions

of dollars of revenue. And interestingly, for businesses that, in my opinion, would not exist without cable. I think History Channel is a great example of a business that simply would not be possible. They get paid 20 cents per sub per month. It's not like you could pull them off and then ask everybody in the world, would you pay 20 cents per sub per month or 20 cents per household per month for History Channel. I think you get such a small fraction of the users because the casual fan base is such a huge portion of what they're getting out of that. So I think those are the incentives of the bundler and the provider. One of the things I was thinking about was that, well, a lot of people might complain about cable. seemingly nobody complains about Netflix. And so it makes me wonder about

price and pricing power. How do you think about that as an important thing for a bundler to consider strategically? So some of this is around Netflix has a lot of fixed cost and around originals and things like that. And as they get scale, it's kind of classic scale economies type thing, you can get what seems like an incredible deal on a monthly basis, say relative to cable for your usage. How do you think about that from the bundler's perspective, price and pricing power of a bundle? Netflix, one of the best case studies in bundling. And I think their playbook for doing it, they did it in such an open secret. With everybody watching, they built one of the best bundles on the planet. I think Reed made a choice, Reed Hastings made a choice very early on,

is very public about it. I'm not moving price. People saw it as a consumer forward message. was eight bucks a month for a very, very long time. And he's not moving price. What did he really do by that? I'm not totally sure if this was his reasoning, but this is my interpretation of his reasoning, is in doing so, like most products, you build a product, and then you add to that product and you increase price as your value to your base increases. Let's imagine I have a customer base of one. I can only sell a product to Patrick. And so Patrick is paying me 10 bucks a month. And in order for me to continue to expand my business, I have to add things that cause Patrick to pay me 20 bucks and 30 bucks and 40 bucks and 50 bucks and have to keep adding things that are meant for Patrick. In our previous

terminology, I have to keep adding to the value proposition for my super fans. But if I hold price constant and I say, that's all I'm ever going to get from Patrick. As long as I'm convinced Patrick isn't going to churn, all of my business opportunity is to go get Patrick's buddy to now subscribe. And now I have to find a way to create a product that Patrick's buddy would pay for. Let's imagine I had completely evenly priced products. And I have a product built just for Patrick and a product built just for Rachel. And Patrick's willing to pay 10 bucks for Patrick's product and Rachel's product. And I build these two products and I say, hey, I have a great idea. I know you guys don't really love each other's products. I don't really need it that much. I think you'll find a valuable.

I'll just charge you both 10 bucks and you both get Patrick and Rachel's products. And now all of a sudden, what did I do? For both of you, you know, get two products instead of one. I get from my side, I get the same amount of money early on. But over time, what I get is a much stickier base. You're much less likely to turn because now you're getting two products, not number one. And you could take this philosophy and keep applying it. Like there's a little thing, there's this kind of crazy analogy in the paper about imagine I just repeated that process for seven billion people on the planet. And I produced a perfect product for person one, two, three, four, person one million, person one billion. And I got seven billion unique products.

And I charge everybody 10 bucks a month for seven billion unique products. And then one day, I wake up, I have this great insight and say, why don't I just make one bundle? And I'm going to charge everybody, I'm just still charge everybody 10 bucks a month, but now everybody gets access to all seven billion products. The amount of revenue created is exactly the same, but the amount of value created is way, way more, because now you get all these products that you're a casual fan of. So what did Netflix do by holding price constant? They forced themselves to think about their only path to growth was subscriber growth. And what that meant was every incentive of the business was creating more subscribers. And it percolated all the way, all the way to the company.

And by the way, cable often does the opposite. The interesting thing about, the cable market is that they start with the geographic monopoly. Comcast has 50 million homes passed, and so that's kind of their captive base. So they're a little bit stuck. At some point, their superfan base, they've kind of penetrated. There are no casual fans to go after anymore. They have that base. So all they can do is keep adding value for superfans. And what happens, their bill keeps going up. That's the nature of it. What did Netflix get away with because it's an internet business is they have nowhere near reach saturation of a superfan base. They keep getting to add to it by holding their price constant.

So I think that's one of the main things about, I'm not necessarily saying that if you bundle, you can't change price, you can. When people bring bundling problems to me and I say, when people say I'm going to raise prices, generally that's a sign that you've run out of ways to think about the next frontier of super fans and casual fans for your business.

And holding price constant is a interesting trick for that purpose. Amazon's another great example, this Amazon Prime has done a, didn't raise prices for years and years and years. and managed to add so much value to that bundle, the reason they could add so much value to that bundle was because each thing they added to that bundle, people thought it was crazy.

I got free shipping for Amazon. I thought that was nuts to begin with. That seems like such an amazing deal. And then they had music and they had books and they had all these different things to it. And the thing I think that is not easy to see about that is what they were really doing was expanding the base. It means that you have to keep adding users.

That's always a good indicator to me. If you want to think like an investor about it, a lot of people look for price leverage as a positive sign as an investor. I look at it as a negative sign. When you start seeing a subscription service increased prices, people look at it as a willingness to pay indicator. I look at it as they have saturated my ability to continue growing my base. And so I now have to monetize my existing base more. It's not a terrible indicator. Like, it's good to have a product of people willing to pay more for it. But it's not as positive an indicator as the ability to continue expanding at the same price. I know you've mentioned

the investor had. I know you're an investor, an advisor, an investor. you mentioned at the beginning, how much of that thinking is this bundle lens? Are you proactively looking for companies that have some of these counterintuitive features that we've talked about when making investments? Or is this just a piece of a sort of general framework for how you think about investing your own funds? Oh, I mean, for sure. I mean, I think it's not really that complicated a thesis because if you have this view of bundling, which looks like you have a similar viewpoint, and I'm sure some of your listeners will as well. The heart of this is subscriptions have an economy of scale. And there's lots of principles underneath that. You can get to of what are the

indicators when price moves versus subscribers move and so on. But the sort of core idea is the reason bundles produce value, back to myth one, is that they produce value for casual fans. Bundling has a natural economy of scale. The reason why, to describe economy of scale, it's easier to get from 100 million subs to 1 million subs and it is to get from zero to one. Why is that? Lots of other practical reasons, but in bundling terms, why is that? You have a much bigger casual fan based to amortize a new cost over. And you've talked about Netflix's investment in originals. One of the reasons Netflix's investment originals can keep going up is that the level of hit required by each subsequent original goes down over time because the number of

subscribers they have to keep from churning is getting larger and larger and larger. So you can have a much lesser hit still justify the investment. So what does it mean from an investment perspective is subscriptions are like Rolling Stones. They tend to have natural economies of scale unless they have a cap. in their market. They have a natural, like this product is only applicable to diabetes patients. There's a fixed market. It's only applicable to people in Fort Lauderdale. There's a fixed market. And so you're looking for things that don't have a fixed market. You're looking for things that have that natural economies of scale. I also think you're looking for things where the leadership team

is open to that viewpoint on bundling. So that's what I'm looking for indicators. How do they think about price? How do they think about the next offering? One thing I'd say, that's a double-edged sword here. Because you can interpret myth four, address casual fan overlap as opposed super-fan overlap. you could address that lots of different ways because it could seem like distraction. It could seem like you're going to stop focusing on your core in order to go address another piece of it. And I think you have to find entrepreneurs and leadership teams that know how to make that balance, that know how to both feed their core as well as grow their next audience.

I mean, I think Spotify is a great example of that. It's podcasting and increasing the audio market is obviously a big effort for the company, but the company continues to innovate and add value in its core music offering at the same time. It's not something where it's either or as you get through that. We may have already mentioned the bundle, but I'd be curious what the most interesting bundle that you've ever seen is. I can probably give two answers. I mean, the Spotify student bundle was one of the most fun ones to work on because I think most people's perception was that it was a discounting program. That in that process, what we were doing by giving, the original version was Spotify plus Hulu for five bucks a month for students.

I think most people's view was, oh, this is just these two companies losing money. That was an immensely profitable effort for both companies. And the reason was quite simple. The superfan overlap between those two products in the student space was quite small. The number of students who paid for both products was very small. So everybody who signed up, and a lot of people signed up for this combined product, everyone who signed up was new to this offering. And what that led to was massive value creation, casual fan creation, across both sides of this.

I think that's the most interesting one I've worked directly on. I find the most intriguing one to think about to be health care. I think this idea sort of gave that analogy earlier, healthcare, if you think about health care, if you recast the government health care debate and you say, should the government pay for health care or not, and you recast it and I don't know how practical it is to educate the population on this, and you say, hey, you have a choice. I want to bundle health care. And I'll just name you a set of things. I can bundle health care with your employment. I can bundle it with your citizenship. I can bundle it with your, like, let me just work through the list. What do you want to bundle it with? What do you think is best for you?

And what do you think is best for the economy? Not rendering an opinion on which one it would be, but that sounds kind of nuts. What if Netflix offered healthcare? That seems kind of crazy. Amazon offering healthcare maybe doesn't sound so crazy. I think Bezos has a bunch of projects. They go after it. But I think that idea that if you cast every,

should my taxes pay for this and are you left leaning or right leaning and so on, as just a practical choice. Who do you want your bundler to be? Would you rather be Google? Would you rather be the U.S. government? And I've got some interesting choices. And I think that way of thinking about the world, it's super thought-provoking.

It's like now a lot of times when you have questions, one of my is another piece I wrote on this thing called eigen questions, which is all about finding the right question. Whenever I spot a question that is good versus evil and people have kind of aligned it that way, it's usually a sign that there's something wrong here. It's very rare that complicated questions actually have such an obvious good versus evil. And I think bundling is one of those things that can be a helpful clarifying factor in many life decisions that otherwise wouldn't have been thought of as, oh, that's just a question of who's the bundler and who's the provider.

So I think, you know, I think the healthcare one is really interesting. I want to ask one more question about bundling and then touch on this eigen question idea, because that sounds fascinating before we wrap up. Where does this go from here? What do you think the future of bundling is? What are you most excited about? What is the internet uniquely enable? Where else might this become an important concept?

I think the bundles of the future, I think we're headed towards much, much larger bundles than we have today. And sort of one important conjecture, I think the economy of scale is very large. And I think that perception right now is that, oh, the cable bundle seems too broad and so on. I think we're going to see some bundles that covered products that we never would have imagined putting together. It won't just be your media and your delivery services, but it might actually be things like your health care, your dry cleaning, so on all put together into one, which sounds a little crazy, but I think we'll see it. The more interesting prediction is what I like to call the third business model for the internet. So back to the very start of this call, you asked about superfan businesses, capital fan businesses, non-fan businesses.

is little history of non-fan businesses. Before the mid-2000s, these businesses were quite hard to create. And then there was an advent of something called the Ad Network. One of the most popular ones was one we made at Google called AdSense. It's now called the Google Display Network. And the idea was pretty simple. If you're a publisher, Google would provide an advertising unit and allow you to monetize your audience without having to go sell to advertisers individually.

In my opinion, that business created many other businesses, just this platform. platform. The ability to create this non-fan business model created many businesses. Most of the blogs in the world are supported this way. Many podcasts and so on are supported this way. Much of the media industry games, a lot of games are supported this way, a lot of social networks are supported this way. But it actually allowed businesses to be created by having this business model be possible. I think that we're watching a shift to an emergence of a subscription bundled business model, I think that is going to allow a new class of businesses to exist. I gave the analogy earlier that I think if cable didn't exist, history channel wouldn't exist. It's not that it wouldn't

be good, it wouldn't be bad. I just don't think the company would exist. I think that product would not be there. I think that the theories in this paper are mostly, they're becoming significantly easier to execute. As the world moves online, as products are bought and sold in a much less frictioned way, the ability to do bundles is much simpler. You can pass back referral code. You can do OAuth to be able to get people to sign into multiple things. The actual production of the bundle is very small. I mean, for Comcast, for History Channel, to do that, you put up satellite dishes, you ran extra wires. You had to do all sorts of things to actually be part of that bundle. This is now becoming a business conversation. You can go produce a bundle with a couple of meetings.

So I think we're going to see, not only will it become much bigger, I think we're going to see this third business model of the internet, and I think we're going to see businesses that are only viable as part of this business model. Their core offering is so valuable to casual fans, but not valuable enough to super fans, that their only real path to market is through bundles. And I think that's really exciting. I think that we saw the advertising business model produced some really good businesses, YouTube being one of them. And I think we're going to see a wave of new businesses get created as a third business model comes up for the internet. This topic of eigen questions, I want to make sure we sneak in here just because it's a powerful way of thinking about things.

And maybe you could start with this fun example of teleportation and why it's a good example and kind of what you mean by what an eigen question is and why they're useful. So eigen questions. So this is a made up word, first off. And it's loosely based on a mathematical term from linear algebra. Igen vectors are the most discriminating vector in a multidimensional space, heavily used to machine learning as well. Just as background, I'm a mathematician by background, so a lot of those terms make sense. Are the natural analogies for me, but it's unnecessary for understanding the idea. An eigen question is, in a space of questions, find the most discriminating question, and answer that

question first. And you can think about that if I have 10 questions, for which question, if I answered it first, would it answer the other nine questions? It's a very powerful way of thinking about framing problems, which is what the topic it's used for. And it's a cultural technique we use at YouTube, we use a Kota, it's something. we test for in our interviews. It's something we train for as people join. I think if you're good at framing problems, you tend to be very good at addressing new challenges as they come up. So one of the

test questions we use for it, it was interesting when I published this early this week, and we had a lot of debate about revealing this question because it's been a popular interview question for us for a long time. But we decided it was a good example, and we have another question we'll replace it with. This is one of my favorite interview questions. So I'll sit down with someone and say, okay, a group of inventors have invented a teleportation device. And it works kind of like the one from Star Trek. They've come to you, Patrick, and they've said, hey, you seem really smart. I'd like you to help me figure out, we're just the techie guys. I'd like you to figure out how to bring this to market. So how would you bring a teleportation device to market? And most candidates will start off this question,

and good ones will at least start by being curious. And they'll start asking questions. And tell me about the device, how big is it? Is it two-sided? Do I need one on each side? How fast is it? How safe is it? How much to cause, what color is it. You'll ask all sorts of different questions. And I'll slowly collect, I'll answer some of them, deflect some of them. Hey, we'll come back to that one, put it in your list, and so on. And then I'll pause.

And I'll say, hey, you know, these inventors, they decided that they don't actually like answering all these questions. They're kind of annoyed, actually. And so they've decided they're only going to answer two questions. And you can ask any two questions you want. But after that, they expect an answer. They expect to go-to-market model. What are your two questions for this group of inventors

of this teleportation device? And good candidates will quickly get to an answer. I put one in this write-up about an answer I got recently that I think is pretty good, is a candidate said, I'm going to ask two questions, how safe is the device and one question, and is the device more costly to purchase or more costly to operate? And so how safe is it has a range from safe enough for humans to not safe enough for humans to use sort of binary outcomes?

And the other axis is it costs a lot to make but not a lot to operate, or it costs a lot to operate, but not a lot to make. and you take these four quadrants and you lay them out and you say, okay, what can I do? And it actually turns out with just those two questions, you can come up with most of what you need for a business model. So, for example, if it is safe enough for humans and it is cheap to deploy but costly to operate, then you want to deploy it everywhere. You want it in every room, in every house. You want to treat it almost like human fax machines.

I'm going to put it everywhere and anytime you want, beam me up Scotty and beam me over to the other place. If on the other hand, let's imagine that it is costly to deploy, but cheap to operate, then you can't deploy them very, in many places. And you probably deploy it a little bit closer to like an airport network. And you have a set of hubs and you set that up. And if it's safe enough for humans, you treat it like an airport and you have humans go through it. If it's not safe enough for humans, you treat it like a freight network. And you can kind of like work your way through all four quadrants and you can see it.

This technique, this idea of eigen questions, can I split a space into clear separation? It's super powerful. Hopefully that example is helpful for your listeners. but it's something I think you can practice. You can get better at this. Take any issue in front of you and say, like, what is the question here? It's a fun story in the document about another YouTube one from when we talked about YouTube search.

People can get through as well. But I think it's a really powerful technique. It's kind of fun word, too. I can question. I love it. So my closing question for everybody is to ask for the kindest thing that anyone's ever done for you. I'm going to tell a story.

It was the first one that popped in my mind is a story about one of my favorite coaches. It's a guy named Bill Campbell. And Bill, some of you listeners may know, him. He was known as the coach. There's a book written about him recently called Trillion Dollar Coach. He was one of my very early coaches when I started my first company, Centrada, the primary investors, Vinod's Coastal, Kleiner Perkins, and Bill was coaching all of the CEOs from that network. And so I got to spend time with Bill well before he was primary mentor to Steve Jobs and Larry Page and so on. And little old me could get some useful advice from Bill. And he was super

helpful. He would show up at, and he's one of those coaches and mentors that would help at a moments notice and sometimes provide useful advice and sometimes just provide a little nudge. And he would just sort of orient me the right direction or he'd tell me like, you're already on the right track. You should have some confidence this year. That's the right way to do it. And then you had this one conversation that was really instructive to me. And it really formed how I think about people contributions in life. So we're about a year into this relationship. And I go to Bill and I said, hey, I just noticed I was looking through our records. I noticed we never formalized an advisory agreement. You've been advising the company. He was literally showing up every

week. And he was spending a lot of time on it. And I said, we never formalized agreement. You're not getting paid. I'd like to make sure that. I don't want to take advantage of your time. And he said, no, no, no, she's sure. I don't need any of that. Don't worry about it. And I said, really? Like, that's a little bit weird. And he says, yeah, look, I've been quite lucky in my career. I've made a lot of money, so on. If you really feel strongly about it, write whatever you want, but just donate it straight to charity because that's all it's going to do. Okay, that's interesting. I'm 21 years old at the time. Like, the idea of interacting with someone that was willing to put in their time and turned down any compensation, wasn't something I was that used to at the time. But then I asked him,

I said, okay, well, that's interesting. I feel like every other person I talked to has got something in mind. They're all motivated by something like, what is it for you? What's causing you to do this? How are you judging? And he gave me the most useful framework. And he said, look, the way I look at it is I look at the list of people I've mentored and helped. And I just count how many of them are executives and CEOs at Fortune 500 companies. And I look at my old team and I see where did they land. And I just think of that as my personal scoreboard. It was so powerful. And by the way, he passed away a few years back. And his funeral was one of the, like it was a who's who funeral. Everybody you can, important you can think of was there. And every single person got up and gave a

speech said almost the exact same thing. Some version of the story I just gave. Like Bill always seemed to have time for me. He always seemed to make me feel like it was only my problems were important to him. And yet he seemed to have helped everybody else around as well. And I don't know how he possibly did it. So this is sort of a little bit meta, but his personal time commitment to me was one of the most kind things I'd ever seen done for me. But probably the bigger gift was it just gave me a framework for thinking about helping people that was not so direct and not so transactional and not so if those people succeed. He felt success in other people's success. That was his bar. That was his way of judging. Did I have impact? And I think if you look at it that way, all of a sudden,

like your payback on helping someone is like way more directly correlated. And you don't have to worry about, will this favor get repaid and so on. And so I thought that was a really kind gift. And I try to pay that forward as much as I can. It reminds me so much. It's a wonderful story about an idea Sarah Tattle shared with me, which is this philosophy idea around the notion of friendship being about adopting another person's means or ends as your own, that basically you're just mapping onto what they're trying to accomplish and assuming it was what you were trying to accomplish and that's it and act as if. And it just seems like Bill embodied that. I've heard that about him like you said for multiple people. So great closing answer.

Such an interesting topic. I can't believe our hour and a half is done. I could go another hour and a half for sure. Maybe we'll do it another time soon. But I appreciate your time and all the interesting insight. It's a totally unique business exploration for me and that's rare. So thank you. All right. Thank you very much, Patrick. Great question in interview. And I'm glad I had a chance to walk through a topic that has been fun hobby for me for a long time. Hey everyone, Patrick here again. To find more episodes of Invest like the Best, go to investorfieldguide.com forward slash podcast. If you're a book lover, you can also sign up for my book club at investorfield guide.com

forward slash book club. After you sign up, you'll receive a full investor curriculum right away and then three to four suggestions of new books every month. You can also follow me on Twitter at Patrick underscore Oshag, Oshag. If you enjoy the show, please leave you. leave a quick review for us on iTunes, which will help more people discover invest like the best. Thanks so much for listening.