In this podcast, I talk about network effects. And some of their strengths and weaknesses.

You can listen here or at iTunes, Google Podcasts and Himalaya.

I write, speak and consult about how to win (and not lose) in digital strategy and transformation.

I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.

My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.

Note: This content (articles, podcasts, website info) is not investment advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. Investing is risky. Do your own research.

——transcription below

:
Welcome, welcome everybody. My name is Jeff Towson and this is Tech Strategy. And the topic for today, why network effects suck and really why platforms fail often. You know, this is kind of, I’m filling in a bit of a gap here. I’ve been, you know, going through the content. We’ve got six to seven levels worth of contents. So those of you who are members and are sort of working your way through the various levels. We talk about network effects quite a bit, but I never really went through it in depth, just sort of tackled the subject. So that’s the goal for today, is just talk a lot about network effects, where they’re good, where they’re bad, how to think about them. You know, it’s a big topic. It’s generally uniformly thought as a good thing. Nothing is always a good thing. Things are good in some cases, and then they’re bad in others. So you wanna kind of know the limitations. So that’s what we’ll talk about today. So the key concept, obviously network effects, and for those of you who are members, this is going under learning goal 28, which is in level seven, which I haven’t really published yet. We’ve kind of gone through six established levels, so this is what’s coming next. And for those of you who aren’t members, please sign up, very easy to do so. Go over to jeffthousam.com, there’s a free 30-day trial. Try it out, join the group, get involved, see how you like it. Okay, let’s get into the subject. Network effects are by their very definition a type of externality. It’s a type of moving outside of your company, the idea of a firm, into the ecosystem and leveraging a network that exists in the ecosystem, maybe because it was there, like roads. Or maybe it’s something that you created by laying cable everywhere and suddenly you have a phone system in the network or railroad tracks or whatever. But it’s inherently out there in the ecosystem more than within a firm. And so before we get into that, I sort of kind of want to go up to 30,000 feet. And when we look at sort of the landscape, and we look at businesses operating in a landscape, but one of the things I just kind of keep an eye out for is what’s a powerful mechanism for economic profits? economic growth, which are not always the same thing. In some businesses, you really do see this. You just see a company take off very, very quickly and very powerfully. And you kind of want to ask yourself, well, what’s the mechanism behind that? Why is it moving so much faster than other companies, throwing off so much more cash, growing so rapidly? And there’s a handful of these things out there. You see them all the time. Gambling. Gambling is just really powerful. If you’re selling… Candy bars, that’s a nice business. If you’re selling casino services, you just take off faster. It has a lot of powerful psychology there. It’s just a very powerful mechanism in business, an economic engine, I would call it. McDonald’s, why is McDonald’s so successful? It really took off very quickly. And you could argue it’s about the hamburgers. I don’t think it’s about the hamburgers. I don’t think it’s about the soda. I think a lot of people sell sandwiches and I think a lot of people sell drinks. I think what they did, that they tapped into, was a franchise business model. That yeah, opening restaurants is a good business and it takes some time and effort, but if you franchise out with a popular product, you can grow very, very quickly. You can bring in a lot of partners that, you know, you have local managers, local franchisees that own their local restaurants. You can borrow money against those cash flows. It just had a powerful franchise business model that allowed, you know, McDonald’s to grow so quickly. And a lot of people have copied that. Franchising, you know, there’s some real power there. I don’t know, let’s say a non-business example, trees. I like nature examples, I do this a lot. Trees, trees are really powerful. You put a seed in the ground, it grows up to the sky, and the seed goes in the ground, the roots go down, the water comes in the bottom, the trees go up, the leaves pull the sunlight in, and the thing grows and expands, and then you have the root system that takes the water up, and they grow pretty darn fast. That’s just a mechanism. that exists out in the ecosystem and some businesses and or animals and or plants and other things tap into those sometimes. Now, I mean, the reason I bring up trees is because the second point would be none of these mechanisms go on forever. They do well for a while. And then ultimately there’s sort of a feedback loop that works in the other direction and it stops the mechanism. In the case of trees, You know, they grow and they grow, but trees don’t grow to the sky. There’s no mile high trees, right? They all go and then the gravity starts to become more and more of a factor. The bark and whatever can’t quite handle the weight. The root system can’t move enough water up that high so quickly. And eventually the powerful mechanism of growth is sort of counteracted by some other forces and it stops. And that’s kind of what we see. So whenever you see a powerful mechanism, all right, that’s good, study that. And then start to think about a sort of a negative feedback loop that’s going to go the other direction. And you see this all over the place. So casinos don’t sprawl forever. Well, I mean, casinos do seem to go pretty powerfully. You get some psychological impacts. You could argue that the government steps in at certain points. If the whole city’s gambling, they step in and start to become a counteracting force to slow that. Although often what they do is they take over the business. The largest gambling companies in the world are not Vegas and they’re not Wynn. It’s the state governments of the United States that run the lotteries. But usually you see any mechanism is going to be counteracted at a certain point. And we’re going to get to networks, because networks are an example of this. If you study medicine, really medicine is kind of interesting in this regard in that you ever wonder why your body is always at 98.6? Your temperature, it’s always at 98.6. It goes up a bit, down a bit, but how does it stay so stable? If you have a fever, it’s going to go up for a while, but it’s remarkably stable. Well, it’s because it’s kind of a dynamic stability. What the body has is it has a positive and a negative feedback loop. And if your temperature goes up a little bit, you go out in the sun, it’s Thailand, it’s really hot, immediately your blood vessels dilate, and there are all these mechanisms that kick in to cool your body, and it moves back down. If it goes too low, well then another set of mechanisms come online, and you start to constrict your blood vessels, and you shiver, and you do other things. But. The reason you sit at 98.6 is you’re sort of in this dynamic equilibrium where the positive and the negative feedback loops are in balance and it keeps you there. And the whole body is built like this. Literally every cell, all the channels that move potassium and chloride and things in and out of cells are all positive and negative feedback loops that counteract each other. So you sit in that sort of equilibrium points. Well, I think we see the same thing in business. And there’s one person you should follow in this regard. Well, there’s probably two. Two people who really think about these positive and negative feedback loops in business would be George Soros and Howard Marks. If you read Howard Marks, who’s Oak Tree Capital, unbelievably smart person, long time, mostly stock guy. He writes these newsletters. You can go over to Oak Tree Capital, sign up for a newsletter. You’ll get one about one every week. He’s incredibly smart. This is like, I’ve told you this before. I mean, this is, you know, Warren Buffett says when he gets his newsletter from Howard Marks, he stops what he’s doing to read it. He’s very, very smart. And what he talks about in business is cycles all the time. It’s kind of the same idea of temperature and cellular regulation. That the business world is always swinging back and forth. between positive and negative feedback mechanisms that keep most businesses, industries, economies, credit situations, cycling back and forth a little bit like a pendulum. And anytime it goes too far in one direction, other forces emerge that push it back. And so he talks about the economy all the time as a series of cycles, always going back and forth, always going back and forth. It’s a really good way to think about it. And within there, you can talk about you know, economic, you can talk about growth. But the one everyone pays the most attention to is credit cycles. Whether they’re short-term credit cycles or they’re sort of what they call the super bubbles, the super credit cycles, which go on for decades, and people talk about that a lot. But you know, I think it’s a lot easier to look at business, look at the ecosystems in terms of these cycles, and Howard Marks is the go-to person on that subject. The other person who’s always talked about this, really since Spike, 1980 ish is George Soros, who has, you know, he’s a political guy now, but he’s one of the best analysts you’ve ever seen in terms of macroeconomics. You know, the guy who broke the Bank of England going against the pound. And you know, he published a book, I think in 1983, called The Alchemy of Finance, that basically talks about boom bus cycles in mostly macro, mostly credit cycles, and things that have a lot of credit in them like real estate and banks. And that they form these sort of boom-bust cycles all the time and they’re very predictable. And he calls it reflexivity. That’s his theory of reflexivity. The idea is when real estate starts going up in value, often what happens is people, because the market is going up, going up, banks loosen their credit standards and you can borrow more because you can say, hey, look, my apartment’s worth more than it was, and you can borrow more money. So as the market goes up in terms of value of, let’s say real estate, credit becomes more accessible. Well, when there’s more credit in the market, that tends to boost real estate values even more. So they go up more and you get the sort of positive feedback loop between credit and real estate prices up to a point. And then the negative feedback loop happens and it crashes in the other direction. So suddenly housing prices start to fall. Banks pull back on their credit. That decreases the market perceived value of this real estate. The values fall anymore. Prices, they contract, they contract. So I mean, it kind of, you see these boom bust cycles based on what he calls reflexivity. It’s a great book, Alchemy of Finance. It’s terribly written. He’s such a smart man and he’s like the worst writer ever. It’s so verbose and the guy like never thought of putting in a chart. You know, it’s it’s really difficult to read, but thinking is pretty fantastic. And there’s not a big surprise as a billionaire if you read his analyst work. Anyway, so when you think about cycles, you know, it’s sort of I’ve sort of laid out three ideas here, the idea of powerful economic mechanisms, feedback loops that often go back and forth in nature and in business, and then cycles. which we see in credit. We also see these in industry. I mean, everybody knows this. Like when oil prices go up, people invest more money in drilling and doing refineries. So you get a big, you know, a lot of spending. And then when you get over capacity in your refining capacity, and then so prices fall, and then prices come down and then people pull back. So, you know, things like the oil industry and some of these classically cyclical industries, a real estate development. oil and gas production, things like that. I mean, they’re classically cyclical. And it’s the same idea that you see these mechanisms going back and forth. Okay, so powerful mechanisms, economic growth mechanisms, feedback loops, whether positive or negative, and then this idea of cycles. Okay, now that’s sort of standard business 101 stuff. Okay, what happens when we start applying that to networks? And networks are just things that exist out in the ecosystem of the, you know. of life, transportation networks, telephone networks. Now we’re seeing social networks, we’re seeing communication networks, collaboration networks. And sometimes these are created by companies, sometimes they already existed. Social networks have always existed, people have always had friends. Mark Zuckergerberg just moved it online and made it digital and then it became a lot more powerful, but it’s already there. So there’s this idea of building platform business models on networks and sometimes creating the network. in itself. When you start looking at mechanisms of economic power, economic growth, feedback loops, well it turns out when those things are based in networks, they can be particularly powerful and hence the idea of network effects, hence the idea of feedback loops within digital platforms. All of this just becomes really powerful when it’s no longer something happening within a company, but that’s something that’s going out across the ecosystem through a network. It’s a bigger phenomenon. Okay, let’s get back down to the micro level. I don’t really like macro very much. I don’t do it. I like to study companies, but that’s a little bit of high level stuff. Okay, so what is a network effect? The standard example I’ve been giving you over these many talks is it’s this idea of when more people use a product or service, the value of the product as perceived by the user goes up. So you go, you know, you go to KFC, I go to KFC. your chicken didn’t taste any better because I went there. The product was basically the same in terms of its perceived value to the customer. But if I’m on Zoom and then you join Zoom, well, my Zoom just got better because I can now call more people and that’s great. So the value of the service is actually higher. Now that’s the standard example I’ve been giving you, which is pretty simplistic. Basic thinking on this is once you start looking at network effects, people always say, well, there’s one-sided and two-sided network effects, also called direct and indirect network effects. The Zoom example I just gave you is a direct network effect. There’s one user group, in this case, users, customers, and if one customer uses the product, it’s better for the other customer. So that would be communications, Skype, WeChat. The more people that use it, the better it is. Telephone lines, the more better. Payment platforms, you know, if I can send you money, two-sided indirect network effects would be, you know, marketplace platforms like Alibaba, Amazon. The more merchants that are there, the better it is for all the consumers who wanna buy stuff. The more consumers there are, the better it is for all the merchants. But more merchants doesn’t help the other merchants, and more consumers doesn’t help the other consumers. So it’s kind of… across the platform, an indirect network effect. And that’s ride sharing, that’s eBay, that’s Amazon, I mean there’s a ton of those. That’s YouTube, where the more people that make videos, the better it is for people who watch videos. The more people that watch videos, the better it is for people who make videos. Content creators as one, user group viewers as another. Okay, that’s kind of the standard blurb on that. But if we were to look at one-sided network effects. Okay, let’s say we take the classic example I just gave you, which is communication network effect. That would be Zoom, that would be Skype, that would be WhatsApp, that would be WeChat, and that really would have been telephone landlines for a long time, mobile carriers, things like that. Okay, what value is being created here? Now, there’s a thing called Metcalfe’s Law, which has been floating around for 30 years. The idea that the value of a network goes up more or less exponentially with the number of users. So the more nodes in a network, the higher the value of the network, because the linkages between nodes goes up greater than the number of nodes. I think that’s a pretty bad idea. I don’t really like, one, it’s not a law. There are no laws in business. There are laws in science. So, you know, Newton has a law, Metcalf, there’s no law. It’s an idea. I don’t think it’s terribly good for a lot of reasons. I think if a bunch of people in Turkey join WhatsApp, that doesn’t increase the value to me. So it uses the word value very incorrectly. It says economic value. No, it’s not. The economic value comes from cashflow. Adding nodes to a network doesn’t necessarily impact cash flow. That’s bogus. Okay, then the other way to think about value is the value to the user, the perceived value. My chicken tasted better. I like this service. The value to me goes up. Well, if people in Turkey join WhatsApp, that doesn’t help me at all. So it doesn’t add the value to the user and it doesn’t add economic value. So I think the whole thing is not worth thinking about very much, but people cite it all the time. It turns out, you know… This is a powerful phenomenon. The more people that add a communication network, say one-sided direct network effect, increases the value to the user and probably the economic value. Okay, that’s a mechanism. And it is very, very powerful. Hence WhatsApp went up like a rocket ship. However, getting to my first point, it doesn’t go on forever. At a certain point, if another million people join WhatsApp from Turkey, it doesn’t help me. I don’t talk to a million people. I talked to a couple hundred people. I call a couple hundred people, but the increase in value doesn’t go on forever. At a certain point, it flatlines for each user. And so this network effect is powerful for a while, up to a while, and then it flatlines, and it doesn’t add any more value. And that’s where you want to think about it. So I think, OK, that’s one way to think about it. The other problem there is… the value doesn’t necessarily go up with the number of users, it actually goes up oftentimes with the participation. If I’m on YouTube and the number of videos, the number of people, content creators on the other side of the platform, this is two-sided, the more content creators doesn’t necessarily make more value for me. One, they may not be putting up many videos. What I really don’t care about is number of content creators, I care about number of videos. So it’s their activity is driving the value, not the number. And actually it’s a little more complicated than that because if they’re putting up a bunch of schlocky, stupid videos that I don’t like, that doesn’t add. So it’s actually quality activity that needs to go up for the value for me to go up. So it’s not just users, it can be participation, it can be engagement, it can be quality content. It can be quality products put on Alibaba. It can be successful purchases that increase the network effect to one user group or the other. So it’s a lot more complicated than, hey, there’s more people that your phone connects to. So communication network effect, that’s one type of one-sided network effect. Another type we’d call a standardization or collaboration network effect, that’s like PDFs or Word docs. The more people that save their documents under the PDF format, the better it is for all of us because we can all share documents. So we want a degree of standardization in document formatting. We want a type of standardization in documents or MP3s or MP4s or things like that because it allows easier collaboration and sharing of documents. So that would be a different type. And we see a lot of this in enterprise businesses where you really want a lots of types of businesses like accounting and video connections and document sharing and document editing. You want people using the same standards so it’s easier to connect things. That would be like if you had a bunch of railroads in your country, but they had different sizes of rails they were all using. It might be good within each railroad, but then they couldn’t connect with each other. So you kind of want the government to step in and say, look, all railroads must use these sides of rails so that the cars can all be used by everybody. So there’s a standardization one-sided network effect. Another type would be user-generated content. If you’re on Wikipedia, if you’re on Yelp, if you’re on TripAdvisor, You know, we could be a user on that as a consumer looking for a hotel, looking for a restaurant, whatever. And that’s a two-sided network effect, an indirect network effect. But if I’m also leaving content and I’m writing stuff, hey, I like this recommendation, that’s valuable to the other users on my side as well. The more reviews, assuming the reviews aren’t fake, which a lot of them are. So the more user-generated content generally also creates value on one or both sides of the network as well. people who leave comments in YouTube videos, Youku videos, Yelp reviews, and generally the value pretty much increases. If I’m going on Yelp looking for a restaurant or a Meituan or whatever, generally the more restaurants that list, which would be a two-sided, an indirect network effect, the value to me goes up linearly. One restaurant, two restaurant, three restaurant. the value to me in my neighborhood goes up linearly. So a local network effect that’s linearly increasing. And then the value by other consumers leaving reviews goes up for a while and then it kind of flatlines. I don’t need to read 100 reviews of the sushi restaurant down the street. If there’s 20 or 30, that’s enough. So that one kind of goes up and then it flatlines. So think about user generated content and. We could also talk about transportation networks. That’s a little more complicated. Okay, I think you get the idea. That’s a one-sided network effect. And then you kind of have to, okay, decide what, what is it that increases value and how does that value increase with number of users or with their level of activity or with their type of content or whatever. And it can be linear. It can keep going up and up and up. It can go for… a little while and then flatline pretty quick. It can go up exponentially for a while and then flatline kinda like an S-curve. And usually that’s pretty obvious. Uber tends to be a local network effect that goes up. Hey, I’m a rider, there’s more drivers in my area. The more drivers they are, the more available they are, the shorter my wait time it goes up, it goes up. But once there’s 50 drivers within five blocks of my home, then it flatlines. I don’t really care if there’s another 50, doesn’t help me. And you wanna look at it from both sides, the consumer side and from the merchant side. Okay, the other thing to think about is the value to each user side. Is the network effect being created by real time value or more cumulative value? Now real time value would be the Uber example. I want to know how many cars are on the street right now, 7 a.m., I’ve got to get to the airport, I look up, I open the app, how many cars are available. And if there’s 10 cars instead of five, the value to me goes up as a rider. If there’s 20 cars, it goes up. If it’s 50, because the wait time goes down. And it’s the same for the driver. They’re looking for people, they’re sitting there looking for a ride. You know, if there’s a lot of people looking for rides, that’s better. If there’s even more, that’s better. Easier to find them. but at a certain point it flatlines. So that’s sort of a real time, very local network effect that goes up quite sharply and then it flatlines. Okay. YouTube isn’t like that. YouTube is more of a cumulative network effect. If a bunch of people are making videos about travel to Thailand or whatever, something I’m interested, the more videos, the number of people making videos doesn’t help me. What I want is the number of videos posted and I want quality ones that are posted. If there’s a hundred bad ones that doesn’t help me. I want a hundred good ones. And the value keeps going up the more videos they are and it doesn’t have to be in real time. They could have been created two years ago. They could have been created five years ago. If it’s a history subject they could be created ten years ago. So the network effect for something like YouTube it goes up. with cumulative content, not how many people are making videos right now at 7 a.m. on a Monday morning. So there’s sort of cumulative value and real-time value, and you want to think about that. If you have a weaker network effect, the cumulative thing can really actually help you over time. Cumulative content also helps you if you have weak switching costs. If you’re in a real-time situation, you want a powerful network. Newspapers have this problem. Like the content they created yesterday, nobody wants to read it. Only want to read what they’ve created today. If you’re in a real time situation, creating real time value, you really want a powerful network effect. You want maybe unique data. You want a good switching cost. If you don’t have those things, often what you can have is you can have a cumulative effect where you can create collections of content over time, collections of feedback. you can build a reputation over time. Let’s say I’m a content creator, right? I’m creating content, but I’m not competing with the news to try and get your attention every day by coming up with some shocking contrarian hot take on whatever happens to be trending this morning. I don’t really like that game at all. I’m, you know, I don’t have that and I don’t have switching costs and I don’t have any of that. What I have is I’m building a library of content that is mostly what you’d call evergreen content where it’s gonna be more valuable. You know, if you sign up for this course today, you actually are getting a higher value than people who signed up nine months ago because you’re getting access to a dramatically more content because the library is all there. And you know, that’s what I’m building this course on. So I’m sort of using cumulative network effects to compensate for the fact that I don’t, you know, I don’t have switching costs and a hugely power. I don’t really have a network beyond that. So think about that. The other thing a guy like me would get as a content creator is I would get a reputation value. That is, as I create more and more content and do this longer and longer, I get sort of my reputation does grow. And this is what a lot of YouTube people are really doing is they’re building their reputation as opposed to driving traffic or making ad money. So you can get that cumulatively growing over time. Influence, Twitter followers. which is, you know, influence is not quite the same as reputation. If you’re doing something like insurance, where you’re looking at lots and lots of cases over time, you get sort of a cumulative effect in your ability to price risk, because you have this long history of looking at accidents and things. So as you get more and more interactions over time, you get better at that, they call that like a learning network effect or a data advantage, things like that. So you wanna think about real time versus cumulative, you wanna think about how it looks to each user group, and you wanna think about is it exponential, is it local, is it international? There’s a lot going on within network effects. And I always come back to what is the perceived value of each user group as there is more and more of what on the platform. Last bit on this is to think about what else should happen if you get a network effect? If my service or a Amazon or whatever is becoming more and more valuable to each user with activity or content or whatever, how should that show up? Should people be willing to pay more? What business metrics would you be looking for? You might be looking for willingness to pay because it’s more valuable. Netflix has so much content. Oh, they don’t really have a network effect. Let’s say YouTube. YouTube provides so much value. more and more and more every day with more videos that they could in theory charge more. So willingness to pay might be a measure of the value that is increasing over time. You might also see that maybe not as willingness to pay or pricing. You might see that as a falling customer acquisition cost. That it’s easier and easier to get users because the value is so overwhelming. and you don’t have to do as much marketing spend, maybe as you had to do when you were a much smaller platform. You might see rapid growth. These platforms do tend to take off as the market sort of as one player, either because it’s cool or because one platform, one player is so much more valuable than another smaller player that everyone just abandons the smaller one and moves to the big one. So you can see this in a lot of metrics and there’s ways to measure the effectiveness of your network effect over time. But those would be kind of three that a lot of people look at. And I guess the other question is, what do you mean by value? When we talk about these marketplaces, usually talking about transactions, hey, you’re buying a pair of shoes on Alibaba. So the more people that sell shoes, the more valuable it is to me as a consumer. the more consumers, the more valuable it is to the merchants. Okay, but does it just have to be about the transaction? Can’t a platform with a network effect create value beyond just a purchase? And this is what a lot of companies are doing now is they’re talking about, look, the goal is not just to try and sell someone sneakers anymore. Like I’m talking about Nike in China here. The goal is to provide your consumers, your community, which is the word everyone’s using, with just tons of value in their life. Make their lives better. Give them information, give them content, create somewhere they can hang out and meet other people who run. Provide them running clubs around town. Build them a free running app that they can download on their phone, which I have the Nike one, where you can train and be part of all this. And create all this value. that is much, it’s just beyond the idea of, oh, I’m gonna buy sneakers, so therefore five merchants that offer sneakers is better than one merchant, that’s the network effect. No, no, it’s creating value in their lives in an activity that is just far beyond that. And then along the way, people actually buy sneakers as well and that’s the part you monetize. Which is kind of how Facebook and a lot of these purely digital companies work. Like WeChat’s a great example of this. WeChat has a very powerful network effect in that they give so much to anyone who uses that app. Communication, payment, content, community, sharing, all of that, and they only monetize one part of it. So the transaction, you don’t wanna get stuck thinking about network effects just at the transaction level. You wanna think about it much more broadly. And the best companies are doing that. They’re creating communities and people have fun and the communities believe in things and they have political positions often and the company routinely tries to surprise their users with special gifts and special events and free live streams and all of this. They’re trying to make your life better. All that is value and you really want as much as possible for your network effect to be beyond just, hey, I’m buying shoes, here’s five merchants, five is better than one network effect, bam. It’s a very narrow version of all of this. Now let’s switch gears, because I’ve just told you this is all pretty cool, but the title of this podcast is Why Network Effects Suck. Yes, they can be powerful. They can be a very powerful mechanism. However, keep in mind feedback loops and cycles, which I mentioned, that they only reach to a certain point. There’s always a countervailing force that emerges at some point. So it’s within a certain type of situation, but it can also be really not a good thing to have. And the first sort of scenario you would want to avoid is anytime you’re sub-critical mass or where you’re still dealing with chicken and the egg problem. And this is basically like when you’re trying, you’re small. You’re trying to get a platform going, a network effect going between two user groups or within one user group. And generally, let’s say you’re trying to get merchants or you’re trying to get. Merchants to take your credit card the jeff credit card i’m rolling it out in shanghai why couldn’t do it in china. Let’s say new york i’m trying to get merchants to take the jeff credit card they ask me how many consumers do you have i say none i try to get the consumers to take it i sell by the way it’s only accepted two stores. They don’t want it so there’s a period where the thing doesn’t work. And i end up having to spend a lot of money cuz the perceived value to the consumer and the merchant. Is. less than it costs me to run the platform. So that’s when people start giving subsidies. You know, we’ll make it free for you. We’ll make it free for you. Because, look, it’s not worth that much as a service yet. You haven’t quite got the chicken and the egg problem worked out. And even if you have got it kind of functioning, you’re still small, you are generally still, as a service, operating below the, your value is still below the cost of running your service. You’ve built a platform, you’re doing all that stuff, you’ve got your IT cost, you’ve got your delivery people or whatever, and you’re not getting enough activity to cover your costs, you’re still negative. It hasn’t really clicked in where the network effect starts to take off. And that’s what they call critical mass. Critical mass is kind of when you’re still dealing with chicken and egg and your cost structure and your perceived value are out of whack. And then at a certain point you get to critical mass and everything starts to hum. Suddenly the value is perceived and or charged to, you know, it’s greater than your cost structure. And it starts accelerating away from that because usually this is a lot of fixed costs. And you know, you start making more and more profit and you start getting more and more users and your wait times for rides start going down and your wait times to get your food deliver, start going down and all the metrics that encapsulate the value start increasing sort of farther and further away usually from your cost structure. Platforms are not really profitable up until they get that point, so that’s bad. You don’t want to get stuck in that scenario. Your user acquisition costs are high, your service quality is probably not awesome, you’re not going to be profitable. So that’s all bad. On top of that, you are likely competing with someone and everybody is well-funded probably. and they are desperate to get to critical mass first because whoever gets to critical mass first, it’s gonna tip in their favor, the market’s gonna probably collapse to them. So it’s like a race, you know, it’s like a death match. You have to get there first. So one, it’s not a great place to be, but two, it’s a very desperate competitive situation. And this is where Reid Hoffman talks about blitz scaling. You know, it’s the analogy I like is the fast and the furious one. where you’re in the race and you’ve got to get to the finish line first and everyone who doesn’t get to the finish line first gets nothing. So that’s when you hit your afterburner, you flood money into your business, you give away subsidies and much again it’s a really difficult spot to be. And if you get to the finish line first, bam, you get the market. And everyone else dies pretty quick, you hit critical mass, your service levels start changing, blah, blah, blah. And hopefully your business is profitable, but it doesn’t necessarily mean it’s profitable, you’ve captured the market. Uber got there, they hit the finish line, they captured the market network effect, and then they realized the unit economics of their business was not awesome and maybe not profitable. So this is about competitive dynamics, not about profitability. So that kind of sucks. You don’t wanna be network effects subcritical mass. That’s very, very bad. The other problem they talk about, people talk about, is called the leaky bucket. None of this thinking is mine, by the way. I’m summarizing other people, and this is all kind of well-known. I don’t want you to come across like this is all me. People talk about the leaky bucket, which is let’s say, okay, you’ve got a network effect. Let’s say you’re Ctrip. Let’s say you’re Expedia. You’ve got your business going. You are mostly about your network effect. You’re bringing in consumers, and then you’re bringing in hotels, or you’re bringing in restaurants, and you have it, but… You know, you’re always, let’s say you don’t have a lot of repeat business. You’re always dependent on getting new people and bringing them to the platform. Maybe because you don’t have repeat business or maybe you just keep losing customers, hence the leaky bucket. You keep losing users. So therefore, you are incredibly dependent on bringing new people into the platform all the time because you’re always leaking. And you know, if your number of users goes down, well, the network effect starts to go the other direction. Suddenly you have fewer users, consumers, than last month, so you’re less valuable to the merchants. And then the merchants don’t like that and they start to leave, and then you have less value to the consumers, and then that’s less valuable. So this network effect goes in the other direction. If your activity levels, your user level, whatever your key metrics, if that’s decreasing, this thing can all go the other direction pretty fast. And that’s hard for businesses that have what we call the leaky bucket. Maybe it’s because they’re not well run, people aren’t having a good experience, or maybe it’s just the nature of the business. News companies don’t really have network effects, but they do have the problem of most of them have to keep getting users to come back every day and read what they’re writing about. That’s a very hard game to play, especially these days, because the amount of video sites, merchants, products, services is expanding over Every day there’s more and more abundance in our lives. So getting people’s attention is getting much harder every single month really And if you’re in this situation You see a lot of these platforms like I think hotels and I think tourism platforms have this problem like C trip and Expedia Where you know, you don’t buy a hotel or you don’t rent a hotel very often You don’t buy a plane ticket very often. So they are always trying to reacquire consumers because they don’t get repeat business, they don’t have a lot of attention. And they’re very dependent on this because they have a leaky bucket. And so they have to keep marketing on Google, on Facebook, to get people to come to their site. And their marketing costs keep going up. If you look at something like Expedia, look at their marketing costs as a percentage of sales. It was like 15%, 20%, 30%, it keeps going up because if they don’t get people there because of their leaky bucket, their network effect will start to go in reverse. So that you don’t wanna be in that scenario. It’s a really bad, bad scenario where you almost become desperate to keep people coming onto your platform. Most of the best businesses, I think, are not about acquisition anymore, they’re about retention, that they have very, very high retention rates. That’s sort of the key metric. Anyway, so leaky buckets a problem. Critical mass is a problem. Another term you’ll hear is the economics of participation. I wrote some articles about this a year or two ago called the Strange Economics of Participation. Everyone loves network effects because you have so much power in it and you don’t have to own anything. Hey, Uber doesn’t own any cars. Okay, the problem with that is if you don’t own stuff, you are very dependent on people participating. So your whole economics are dependent on other people, users, customers, whatever, participating in your site. So the economics of participation is another way people talk about this. One other concept that’s sort of related is multi-homing, which I brought up a week or so ago, was at the company level or at the product level, if you’re buying a can of Coke or signing up for… I don’t know an accounting firm to do your books. We would talk about switching costs. How easy is it for customers to switch? At the platform, at the network economics level, the phrase people use is multi-homing. Yeah, people can be on Uber, the drivers, but they can also be on Lyft very, very easily. They just have another app on their phone. If there’s no one to, if they can’t find a ride on Uber, they just hit the other button and they look on Lyft. And if Lyft decides to increase their take rate, So pay drivers a little less probably. Well, they don’t like that and they just switch back to Uber. And when it’s very easy to live on two platforms and hence have two homes, a multi-home, the platform doesn’t tend to collapse to, the market doesn’t tend to collapse to one player. It’s two players, three players, four players, and they’re all kind of equally strong. That’s also another scenario. That’s kind of where Uber is right now. Yes, they got their critical mass. Yes, they have their network effect. They don’t have a chicken and the egg problem anymore. They don’t really have a leaky bucket per se. What they have is a multi-homing problem that all the drivers and all the writers just live on two different platforms and they can switch back and forth like it’s nothing, depending on how well they’re being treated that particular day and that that’s a really, that problem is not going away for Uber or Lyft and it’s probably going to impact their profitability. DD doesn’t really have that problem because they don’t really have any competitors. The market collapsed to one player. And one sort of last why network effects suck example is what we call interaction failure. And I think I first heard this term in platform scale, which is the same gentleman who wrote platform revolution, which I’ve talked about many times. I think he came up with this term. That’s the first time I heard it. talked about was interaction fail, which is, okay, you’re on Amazon, let’s say Alibaba, I go on Alibaba, I’m looking for the product, and I look at, you know, I want Ugg boots or something, and I search and I find the Ugg boots and then I interact with the merchant, I chat a little bit, but for some reason, whatever happens at the end of this, the interaction just doesn’t go through. That’s an interaction failure. If you get a platform where, I mean all these network effects are based on interactions. If there is a high rate of failure of the interactions, people get very frustrated very quickly. If I click on a video I wanna see on YouTube and then I watch five, you know, two minutes of it, I’m like, this sucks. And I click over to another video about, let’s say I wanna learn about traveling in Bali and I click on a travel in Bali video. first one that comes on the search list. I watch it 30 seconds, I’m like, dude, this sucks. And I click on the next one, I watch that one. Dude, this sucks. Somewhere around three or four videos, I’m gonna be like, dude, YouTube sucks. Cause I’m getting, in that case, it would be low quality content. Or let’s say I search for, I wanna go to Warsaw, and I click, I search for travel Warsaw, and nothing comes up. So. Interaction failure because I don’t find what I’m looking for or I find an example of what I’m looking for, but I don’t like it Content the quality whether it’s a product or service or content is not good all of that can I either I don’t watch the video I don’t buy the product. I don’t get the food delivered if any of that happens That’s interaction failure and people will leave very quickly Dating sites are famous for this I mean, if people go on dating sites, especially women, and they don’t find people they like, or they have a couple bad experiences of chatting online, they exit that dating site super fast. So this is sort of a quality and a quantity of interaction that is required there. And the trick here, and this was Sanjit, who’s the author of Platforms Kill, his point here, which I thought was outstanding, was, This problem of interaction failure actually increases the larger you get. It’s actually not as hard when you’re a smaller platform or you’re doing a smaller number of interactions. It’s when you get bigger and bigger and bigger that curating your quality and customizing the interactions and making sure people are finding what they want and there’s not a lot of fake goods on your site or a lot of badly behaving people on the dating site or a lot of stupid videos. As you get bigger and bigger as a platform, this becomes a bigger and bigger problem. So fake news, misinformation, spam, hate speech, you know, that stuff that Facebook is always sort of dealing with. It’s all the same problems. It’s interaction failure. And I called this in a previous talk, mismatched and or crippled scale, where one side of your business scales beautifully like adding content. But your other side curating the content and keeping a certain level of quality doesn’t scale as easily, you get a mismatch. And that can lead to a reverse network effect. The dating site, all the women are like, we’re outta here, this place is toxic. And they all leave. And then, well, if it’s a straight site, then all the dudes leave and then the other women leave and you have a reverse network effect. I think Twitter has this. Like the number of people who have left Twitter is absolutely stunning. The number of people who have tried Twitter and left, I think it’s like 80 or 90% of their users have left cumulatively over time. It’s something like that because they found the environment toxic and the interaction’s so awful. And I think it’s a cesspool of a social media site. So you get a reverse network effect. This can be a big problem. goes down spam Unsolicited messages unsolicited connections on LinkedIn or whatever Poor customization, you know, lots of abandonment. We’re out of here and the people who tend to leave first are the most credible Like if you’re on YouTube or you’re making high quality content and you know, there’s a lot of bad behavior It’s the quality producers that leave fast It’s the quality content because they don’t want to be associated with any of this So you almost get the best people tend to live fast, leave early. Anyways, so think about that as a reverse network effect, and it actually becomes a bigger and bigger problem as you get bigger. Okay, that’s, I’ve been talking at you a lot today. No case for today, it’s just me going through a lot of theory. Didn’t even talk about any companies. Usually I try and talk mostly about companies because it’s a little more tangible. But I wanted to get through a couple ideas just how to think about network effects. That’s the key idea for today. Within that I’ve sort of mentioned a couple sub ideas of how you measure value in a network effect. We talked about critical mass, I talked about critical mass, talked about leaky bucket, talked about reverse network effects, and talked about interaction failure. Those are kind of the ideas for the day. All of this goes under learning goal 28, about which is network effects, and that’s in level seven for those of you who are members. If you want to know more about this stuff, I talked about this also in regards to Baidu in podcast 26, which was, Is Baidu the New AT&T? Not a great title, but it was the basics of physical versus virtual networks. So it was talking about railroads and telephone lines as ideas of networks. So that was kind of an intro and there was some network effects in there. That was podcast 26, if you want more on this. So that’s it for today. kind of a lot of theory. I’ll shift back to cases and more talking about companies next week. But I thought it was kind of worth going through network effects in some more depth. It is pretty important and people talk about it all the time. Anyways, I’m here in Bangkok. I had kind of a fun weekend. I taught a class at China Europe International Business School, which is based in Shanghai, but everything’s being done virtually by… professors and students. I’ve been teaching there for, I don’t know, eight years maybe, something like that, quite a long time. It’s really a great school. I mean, it’s a lot of business schools, a lot of educational institutions are not terribly well run. Siebs is really well run. The admin is just top-notch and the students are great. I mean really outstanding. I like, you know, they’re a little fiery. They test you, they call you if they think you’re full of it, which is great. So I really enjoyed that a couple days. Hello to everyone from Siebs who maybe was in the class who’s listening in. It was really a pleasure and sort of a shame that it’s being done at a distance this year. Hopefully I’m teaching there in the fall. So hopefully it’ll be back on the ground, which is a lot more fun. But we’ll see. But that was great. And I think I’m heading up to Chiang Mai in the next day. I’m getting stir crazy. I literally in in 10 years I have never been. based in a country for so long. Like I’ve been here five months in Thailand without leaving. I don’t think it’s must’ve been 10 or 15 years since I’ve ever stayed this long in one place. I’m going buggy. So I’m going up. I decided since I can’t fly out of the country very easily, I’m gonna take a trip every three weeks somewhere in Thailand. So a couple of weeks ago, it was down to the islands. I think this week I’ll go up to Chiang Mai and maybe ride some motorcycles up in the hills. around Chiang Mai, which gets you close to the border of Myanmar, and Pai, which is the little sort of farming, what you call it, almost hippie crafts village up in the mountains there, which is great. It’s an absolutely fantastic place for riding motorcycles. And you can go up into these little villages. And the funny thing is you get near the Myanmar border. And the last time I went up there, I was in some little town. I was just not too close to the border, but let’s say within an hour. And this little elderly woman came out of her hut, you know, in this little village, and she’s like waving at me, and I stopped the motorcycle, and she comes up to me, and I thought she was trying to sell me something, which she was, and she comes up, and she goes, oh, heroin, heroin? I didn’t understand her, I was like, what? Like, heroin? Like, are you selling me heroin? Turns out, yeah, she was. I’m like, no, thanks, though. Good to know. Apparently, that’s kind of a thing up there, and I guess it’s not unreasonable. I guess it’s not if you see a white dude on a motorcycle way up in the hills, sort of somewhat near the border. I guess assuming that’s what they’re looking for is not a crazy idea because I think it probably does happen. So I probably wasn’t going up there looking for milk tea, right? So it wasn’t totally crazy. I was like, no, thank you very much. I’ll keep that in mind. Anyways, it’s a great area, the whole Pai Cheng Mai. So I think I’m going to go up there for a week or so. see what’s what and maybe buzz around a little bit. So that’s great. Anyways, I think that’s more than enough for this week. Thank you everyone for listening. I hope everyone’s doing well. Everyone’s staying safe. For those of you who may be the CEEBS folks who are joining in for the first time, welcome. It was a real pleasure teaching the class these last couple of days. I had a great time. I hope that was helpful to you. And anyways, I will talk to everybody next week. Take care.