You're listening to episode 3 mapping out the 16 network effects from the network effects master class Pete on the NFX podcast to watch the course, visit nfx.com/masterclass. So when people first encounter network effects, I think there's a network effect or there's not a network effect. That's like saying there's life and there's not life. There's actually different types of life, and there's different types of network effects. Okay? Not all network effects are created equal.
And we started studying and noticing these things back in the 2000s. And by the time we got to about 13 network effects, what we realized was we needed a map. And so we've come up with a map to help us visualize the different network effects and and where they all fit in in terms of how strong they are, what types they are, and then delineating how they're different from each other. And that was a few years back.
Now we're up to 16 network effects, and we're gonna walk through all of them with you right now. Go to nfx.com/masterclass to watch this course with a full video experience alongside transcripts, show notes, and additional reading. NFX Master class tracks your progress and allows you to move your own pace through the material. We will also be adding new seasons to the streaming platform, so be sure to register so you can gain early access new material from NFX. And now back to the episode.
Okay. So the first time we saw a commercial application network effects was way back, 1907. The chairman of the board of AT and T, who is deploying physical phones and copper wires to connect them all together so they could talk to each other.
He said in his report, his annual report, You know, you might wanna buy our stock because what we've noticed is that when we go into a town and we get a few nodes and a few of these copper wire links, nobody can compete with us, and we end up taking all the business in the town. So Beller time we get into, we own everything. We're probably gonna have a pretty good business here. He noticed it. He explained it, and he gave business reasons why it was Morgan. And boy, was he right?
Now this physical direct network effect, the physical being the copper wires, the physical being these telephones, these heavy black telephones that would sit in people's homes, This was the 1st commercial network effect discovered, and it turns out to be the strongest network that we know of. Okay. Now there's a whole type in the center of the network effects map of these blue network effects. And these are known as the direct network effects.
Because every node can touch every other node in the network and wants to. The physical is the most powerful because once you've actually embedded atoms in the ground, like if you're Comcast and you're putting cables to people's homes at your AT and T, once you have these physical things in the ground, it's very hard to rip them out. Okay?
And so I think Comcast has the worst NPS score of any major corporation in America, and they can afford to do that because we can't replace them because they have this incredible direct physical network effect. And so that's at the core of our network effects map physical to right now The second direct network effect we call protocol, And you can see this with faxes or with ethernet or with Bitcoin, where you develop a protocol, you publish it widely, Everyone can now attach to this protocol.
And the more people use that protocol, the more benefit everyone gets from that protocol. So even crappy protocols like fax, are still surviving in the 2020s, even though they were invented in the 70s. Because so many people are using that protocol network effect. So the 3rd most powerful network effect we've seen is the direct personal utility network effect, okay, where you have your identity, like WhatsApp or like Facebook Messenger or WeChat, where it's personal to you.
It directs because I can connect with anyone on WhatsApp. I can connect with anyone, on on Facebook Messenger on WeChat. And its utility, meaning I get some utility out of it, like making a payment or picking the kids up from school. I'll tell you, if my wife you know, hits me on WhatsApp and says, you need to pick the kids up from school, and I don't. I'm in deep doo doo. So I need to be on that network. I cannot leave that network. Right? That is a good defensibility.
That is a good, retention of me, and that's what network effects are there to do. 4th one is what we call the personal direct network effect, which is what Facebook has, which is my person, is my name, is my picture, And I've got all my friends. It's a direct network effect. I can see them all. They can see me. I can friend somebody. I can unfriend somebody. Oh, every lot works, but it's not as powerful as the one with utility. I could stop using Facebook next week.
Not a big problem, but if I stop using WhatsApp, big problem. I didn't pick up the kids. I'm a deep dude. But on the other hand, it's a very powerful network effect, this Facebook network effect, this personal direct one, because am I gonna spend a ton of time building a new friend group if I've already immersed myself in group. I'm really not. I might stop using Facebook for various reasons, but I'm probably not gonna spend a huge amount of time building a new friend group.
When I get sick of being away from Facebook, I'll probably just go back to it. And we see this people do this on. They quit Facebook, and then 9 months later, they say, I'm back. Why? Because you have this retention, this defensibility, this network effect that adds value to every node on the network, and no one wants to leave it. The 5th network effect is the market network. Okay. And this is also known as an N Sided marketplace, but it's still a direct network effect.
And, basically, how it works is, if you have a wedding, let's say, you're gonna need a florist. You're gonna need a venue. You're gonna need a wedding planner. You need a a caterer. These are all nodes on a network collaborating on a project. And in the real world, they know each other. They connect with each other. But we haven't seen it put in digital format Pete. So a market network puts it in digital. Now they can see each other. They can transact together. They can plan on their calendar.
There's typically a workflow software, which allows them to coordinate in this inside of that marketplace. Okay. And once you're on there, once you've got your name and your reputation, your transaction history there, you're not gonna wanna leave because it's your source of income. Just like a marketplace would be. But in this case, it's inside that everybody in 360 degrees can transact with each other. Okay. So the 6 network effect is a 2 sided marketplace network effect.
Very common, whether it's craigslistor monster.com, or you know, Uber. You've got a supply side and a demand side. The more supply you have, the more value the demand gets and vice versa. It's called indirect network effects or cross side network effects. And the marketplace is 1 is now in red because it's got a very different shape because supply and demand. It's not an inside a type of situation like a direct network effect is that's why we changed the color and made them red.
These are very durable, as you know, because so many of the marketplaces online are still around years later. Now they can be attacked. Right? I mean, eBay's been attacked by Amazon Marketplace very successfully by building pro seller tools that were basically better than what eBay had. Pete people multitenanted and moved off, but still, these are very durable network effects eBay still with us. So is Amazon marketplace. Now another red one is what we call the platform network effect.
Now There are some Pete, again, who have used this word platform to mean all sorts of stuff, but it's really not helping you if you use it that way. The way we use it here, and I think would be useful for you to adopt, is to understand that a platform allows businesses to be built on top, particularly software, businesses to be built on top. The best example would be Microsoft OS. Which in 1977 came out, he sort of stumbled into it.
And as long as that OS was operating all those IBM computers, and had all those people on the keyboards, software developers wanted to build their software products to work with that OS to get access to that distribution to get access to those customers. Okay? And you see this now in sort of developer programs that you know, even these web 3 companies wanna get developers developing on top of Ethereum or on top of avalanche, these are trying to make the network effects be a platform.
Same thing happened with Salesforce. They had a SaaS tool. They launched Morgan and said, this is a platform that you can build your software business on top of, please developers come develop on our platform. You'll be able to get access to distribution access to your customers through us And Salesforce is done incredibly well by building in this OS. IOS, right? Apple was worth 40,000,000,000. Then they got iOS. Now they're worth a trillion or 2.
Because they have this beautiful platform network effect that they've developed. So if you understand the word platform to mean this sort of development environment, and understand that this is a really great flavor to build your company with a great type of network effect to use rather than saying, oh, any business with a network effect platform business, which is what some other people say. I think you're gonna get a lot further. So I I would encourage you to think about it that way.
So the 8th network effect is the asymptotic marketplace network effect. Let me explain it in the context of the 3 basic types of marketplaces. You can have a a regular marketplace where the long tail is very long, like an eBay or a Craigslist Flint, where it just keeps going up and to the right for for forever.
The other one is what we call a delayed valve you marketplace where you're basically embedding software onto the demand side or onto the supply side for a long time, like an open Beller, who provided SAS software to restaurants for 6 or 7 years before they opened up the marketplace, and it suddenly became a marketplace where people could book their restaurant reservations through OpenTable. That wasn't always the case. They had to delay for 6 or 7 years before they could create that.
Then you get the asymptoting marketplaces. Where like we described before, Uber And Lyft, their value of their networks grow very quickly at the beginning. But then as the network size grows, the value doesn't increase that much because of the way you consume the value of the marketplace. Meaning you don't want the car coming in more than 4 minutes doesn't really help. You simply need to go to the bathroom before you get in. Okay?
We've seen this again, with with data network effects, where it asymptotes. These marketplaces for data asymptote because the increasing the the value doesn't keep increasing as the network size increases, and you see this quite a lot. So that's why we broke it out as its own separate marketplace time. Right. So the 9th network effect is what we call the expertise network effect. It's another form of 2 sided marketplace between employers and employees or you know, labor and employers.
So let's say we're using intuit. We're using QuickBooks. Once you get used to as a employee using QuickBooks, you get good at it. You know where the buttons are. You've kind of grooved into the software. You become an expert at that software. You now wanna go work for a place that uses QuickBooks, so you don't they'll use another accounting software.
Now the employer, they know that their accounts might be interchangeable between different Pete, So they wanna use the piece of software that most accountants use. So because QuickBooks got a dominant market share, they were able to then create this network effect so that employers are looking for people on QuickBooks, and they wanna use QuickBooks so that they can always hire new people to work on QuickBooks to augment whoever else they have.
So they have a more liquid pool of labor to work on their accounting. So everyone's incentivized to get good at QuickBooks, to use QuickBooks, and to grow that ecosystem, which creates great defensibility and retention, which is the definition of network effects. Other examples would include Figma. As people get used to managing that piece of software, they wanna go to a place that uses it.
And, of course, the companies wanna now use Figma, so they're gonna attract the talent who likes to use Figma. We see this, across all sorts of different pieces of software now that you've seen the pattern. So now we come to data network effects, which is every founder's favorite idea for the last 30 years. Why? Because software and the Internet makes it so easy to capture data. If you can pretend that it's worth a lot, then you feel like you're making progress.
But let's break down really how data network effects work. The idea would be that the more data you get into the system, the more value your system is gonna provide to everyone using the product. Now people think this might be true, and I'll give you a great example of where this does actually work. Which is ways. Right? So you get your ways app on your phone. It tracks where you're driving.
It can tell the system how fast you're driving, where you're driving, Pete can note if there's a accident or if there's a police officer or whatever. It's all in the system, and that data is changing all the time.
So once they've established 10 or 30 or 50000 people using it in a particular city area, you really don't wanna go to a different ways app, a competitor, because you're not gonna get the same granularity because every 10 minutes or every 20 minutes, the data is worthless because the conditions are changing all the time. So in this case, the data network effect of traffic conditions is a really good data network effect and created a $1,000,000,000 company that Google acquired.
But generally, it's hard to use your data to great effect. People talked about, oh, you know, Netflix knows what you like. Netflix knows what people like. We're gonna have these matching algorithms. We're gonna have these suggestions algorithms. How much value did that really add in Netflix? It didn't add that much even though they made a big deal of it. We see this in Beller care.
We see this in all sorts of, algorithms around logistics where the data is gonna make our product that much Morgan superior. It can. It can Beller. It can improve the product experience, but typically these data network effects aren't as strong as we want to believe. Okay. Nevertheless, they are valuable Nevertheless, you should strive to get them. And on top of them, you can build other product features, which might be more valuable than the data underneath them.
And so tech performance is the next network effect. And think of a bit torrent or a Skype where the idea is the more people have downloaded the software to their computer and is running it on the network. If it's close to your home, you're gonna be able to download a movie faster or a photo faster or have better connections to a conversation you're having because that bandwidth will be more available to you. So your tech performance actually increases.
This occurs in other mesh networks where people are doing communications, phone calls. You could even use a mesh network for, let's say, solar battery charging. So if you were to have lots of Tesla batteries on your block and they had the same protocol, you could actually share and the tech performance would increase of all the different batteries and all the different solar panels on your house. This is in theory.
It hasn't happened yet, but this is the idea of using a network effect to improve the technical performance of the system and thus adding more value to the users. So now we've gotten to the network effects that we consider social or human, and they're at the outside of the network map because we consider them less measurable, less defensible, less strong. Alright? But they're interesting, and you will see them play out in various ways in your product and in your competitors' products.
So let's walk through them. The first one would be language. If you can get someone to say, I'm gonna Google something, it's hard for them to go off and use Bing. If you say I'm gonna pick up an or I'm gonna grab an Uber, it's harder for them to open up the Flint app. Right? So language can have an impact on retaining people into a network business that you're running. Because we, as a network, I'm a node. You're a node. And that piece of language is the link between us.
And that can have a powerful effect on people's behavior. The next one is a belief network effect, and we see this with gold. We see this with Bitcoin. We see this with religions. If I believe that gold is valuable, even though I can't sleep on it, I can't eat it. I can't build it. I can't really do much with it. Doesn't have that many industrial uses, honestly. I believe it's valuable, then it's valuable.
And we've been believing it's valuable since recorded history, so we continue to believe it. There's reasons to believe gold become valuable in the future and stay valuable. Because it's been happening a lot, but it's just belief. The same thing is true of Bitcoin. Why is Bitcoin valuable? Because some people believe it will be valuable in the future. I certainly believe it will be, but there's really nothing other than that belief.
So you can create a lot of value with just a belief network effect. And in fact, with Bitcoin and religions and other things, that is the foundational network effect on top of which all the other network effects are built. So it can be really critical. And then there's the tribal network effect, another network effect that's very human or social.
And the way this one works is that you create an in group and an out group with a word, typically, like, I like the celtics, Now I'm in the tribe of people who likes the celtics. I cheer for them. I cheer against the other tribes because they are the outgroup. And if more people like the celtics and more people wear the t shirts, I feel better. I have more fun walking to the park.
I have more fun running into a Celtics fan when I'm in London or in California, even though that's a Boston based team. So these tribes grow, the more people in those tribes, the more value there is, and people feel very strong about the fact they defend them. That's the really unique thing about travel network effects is that if someone says I'm not using Eban using Amazon, I'm like, meh. But if someone says, I don't like the celtics. I like the New York Knicks. I get upset. I wanna fight them.
Okay? Or the same thing for sorortities, fraternities, right? There's lots of tribes that we as humans create that become very emotional for us and will actually cause us to stay inside a network and add more value to that network, maybe even detract value from another network.
If you have a network on network violence, and notice that what people are doing is they're saying I am a Beller fan, or I am a Stanford graduate, or I am a member of this tribe, and they take a word, and it becomes part of their identity that they will defend. So again, a very human, very social activity, but one that affects business and could affect yours if you design it properly. And so the 15th network effect, which is also social, is the bandwagon effect.
And this is the sense that humans have about not wanting to be left out. So Apple is the master at creating this fear of being left out. People are waiting in line to get the latest iPhone or whatever because they wanna get on the bandwagon. They don't wanna be left behind. Slack was very much like this. You know, there was hip chat well before Slack, but Slack was Beller. And the companies didn't wanna be left behind.
You couldn't be a modern software company if you didn't have your employees using slack. That's how cool they were. Just like Apple is cool. And so most companies had to adopt slack Beller they needed it or not because they needed to be on the bandwagon. And so that network formed, that network stuck, and a lot of value was created in Slack and other businesses that create this effect to get people to jump on, don't be left behind.
Now an example you might also know is if you go into Silicon Valley and you've got a Dell as your laptop versus an Apple, you know, you're just not the cool person. You're not on the bandwagon, and people look at you kind of funny around here. It's kind of an interesting cultural affectation. Now Is that enough to get you to move off of your Beller onto Apple? I don't know. It might be. If you're in the middle of fundraising, it might be.
So there's a 16th network effect calling it hub and spoke. And here's how it works. Think of medium in the old days. You would be a node on the medium network by signing up. You would publish into the hub of medium where everyone can see it. And then once a day, medium would send out an email with the best publications of the prior day or the prior few days, which means that Instead of your friends seeing what you posted, you might be sent out to everyone Right?
So there would be this posting, and then there would be this pushing of of an email or a notification out to all the nodes so that you might get a 100,000 visitors if you won the lottery. And then your node, what you would get more attention. You Pete more status. You get more money. You get more notoriety, which is what you Pete trying to get out of writing your post in the first place.
Now through a design of the software in the system and how the algorithm works, a company like medium can adjust how powerful this network effect can be. Now this was used by academic journals for 100 years.
Where academics would seek status and notoriety from their peers and maybe an increase in salary by publishing into the hub They would look at the hundreds of applications, and they would pick 2 or 3 or 4 to publish and then send out to everyone getting this node a lot of notoriety if again they won the lottery. And so you might submit your you might do research for 30 years.
And submit 28 times in an attempt to get 1 lottery win so that you get all the attention and and increase your increase your career. You can also see this in effect with TikTok where people put up their material. They hope that it gets brought into the algorithms and that they get a lot of likes and a lot of followers. And that they will again keep growing their own network within this hub and spoke system.
So now that you've seen the 16 network effects, what I wanna point out is a concept of reinforcement typically when we build these startups, we start by building a product and software and a network and clusters, right, the minimum viable cluster, and we build one network effect. And we get that going, and we Pete it going. Once you've got the first one going, the sky's the limit because you can reinforce that network effect with other network effects.
And part of the journey of the different seasons of running your business is to pick what is the next network effect I'm gonna add to the one I've already got. How does it interrelate? How does it easily flow? Do I take the people already have and move them into a new space? How do I bring in new nodes? Okay. That whole process is fun, and it's exciting, and it's amazing.
But it's it's what we call reinforcement, and you're gonna be hearing more about that in our case studies later about how companies have done that successfully in the past. So that completes the overview of the different types of network effects. Stay tuned to the NFX podcast. As we'll post 1 episode per week until we complete the course.
You can also watch this entire master class online at nfx.com/masterclass, where you can log in, track your progress, watch full videos, retranscripts, and find other related material. Thanks for listening to the NFX podcast.