Chapter 10 - The Strength of Natural Network Effects
To read:
On the lack of network effects in the EV market: Tesla won't dominate the auto industry, because no one will by Noah Smith (I’m a paid subscriber to his Substack and highly recommend)
Cricket networks: It all began with the chirpy synchronizers writes neuroscientist Mark Humphries in this essay on Duncan Watts’s famous journey into network theory and practice.
Human networks: In this Nature magazine essay, Collective dynamics of ‘small-world’ networks, Duncan Watts and Steven H. Strogatz explain their supercharged network theory and research.
The law of the pack: In this seminal 2001 essay in the Harvard Business Review, David P. Reed successfully predicted that the most successful businesses on the Internet would travel in packs.
Nature’s networks: Science Focus magazine shares five examples of nature networking in ways entrepreneurs should mimic to create platforms greater than the sum of their parts.
Success formula: Startup durability = Network Effects + Economies of Scale + Brand + Embedding + IP, argues serial entrepreneur, investor, and network theorist James Currier.
Networking to success: The key factor in the success of companies as different as Microsoft and Wal-Mart is the success of the networks with which they do business, argue authors Marco Iansiti and Roy Levien.
Networks of networks: In the digital economy, networks are the architecture of the new order, argues Competing in the Age of AI - Strategy and Leadership When Algorithms and Networks Run the World, a complex but critical book.
Chapter 10
The Strength of Natural Network Effects
“The richest people in the world look for and build networks, everyone else looks for work.”
— Robert Kiyosaki, entrepreneur and author
To effectively think about the power of natural networks, a good place to start is with a remarkable experiment by social psychologist Stanley Milgram more than half a century ago. Sometimes his study is called the “small world experiment,” and it not only built on a great deal of earlier research, but it also influenced the making of a 1990 play by John Guare, Six Degrees of Separation, and a 1993 movie of the same name. You may have seen the latter, with Stockard Channing, Will Smith, and Donald Sutherland.
Now, I bring up this old study, a play, and a movie because I’m sure readers are already familiar with the well-known utility of the “network effect.” This is often explained as “Metcalfe’s Law,” for my friend, the brilliant scientist, and professor at the University of Texas at Austin, Bob Metcalfe, who articulated this years ago for purposes of understanding telecommunications networks. Simply put, the “law” states that the value of a network is asymptotically proportional to the square of the number of connected users of the system. Simply put, one phone is useless because it won’t have any appliance to connect with. If there are two phones in existence, then there’s a “network,” but of just one connection. Increase the number of phones to five, and the value of the network is 10, meaning 10 connections. Add just two more phones, and you have 21 connections. Double the number again to 14 phones and the value of your network is 91 connections. And on from there… you can march the exponential logic out of a Harvard dorm room and on to Facebook’s over 3 billion users.
And it’s a prediction of that Meta scale by which Metcalfe’s law, conceived in 1983, got an intellectual booster some two decades later from another computer scientist, David P. Reed at MIT. “Reed’s Law” explained in a famous essay referenced on the Digital Companion, the distinction of networks that add self-reinforcing velocity. Anticipating the rise of social media in 2001, Reed argued that “the most valuable of all” networks would be the “many-to-many” —or “group-forming” — network, which allows network members to form and maintain communicating groups.
Reed certainly got it right, and the science of emergent networks is endlessly fascinating. But what I believe is less understood is that these intentional networks build on what are often natural networks, essentially grapevines or matrices that have not yet been wired. Think of the line from the Nobel-winning Irish poet Walter Butler Yeats: “There are no strangers here; only friends who haven’t yet met.”
This takes me back to Stanley Milgram’s “Small-world” experiment. Milgram is best known for his experiments on obedience to authority in the early 1960s, prompted by the trial of Nazi Adolph Eichmann who, like others implicated in the Holocaust, claimed he was “only following orders.” But Milgram later explored the nature of human connectivity, essentially that Yeats-like experience we’ve all had: You know, where, for example, you get to chatting with someone in an airport lounge in Zurich while you’re waiting for a flight to Miami. You bring up the fact you’re from, say, Austin, Texas. Well, it turns out the person with whom you’re chatting has never been to Texas but… in fact, did know a woman from his high school in Florida once upon a time who moved to Austin to get married. Lo and behold, she turns out to be your sister-in-law. Or an employee. Or next-door neighbor.
“Wow,” you both exclaim simultaneously, “What a small world!” An amazing bit of serendipity? Maybe. Milgram wanted to find out if there was something more than coincidence at work in such seemingly random encounters.
In his 1967 experiment, Milgram wanted to test the connections humans share — what I call the networks of which we’re not fully aware. Or, natural networks.
“Milgram was teaching at Harvard at the time, so naturally he regarded greater Boston as the center of the universe,” wrote Wharton School Professor Duncan J. Watts, in his groundbreaking book on networks, Six Degrees — The Science of a Connected Age. “And what could be farther from it than Nebraska?”
So Milgram sent packages to 160 randomly chosen people living in Omaha. He asked this group, which he called “starters,” to forward the package to a friend or acquaintance whom they judged the most likely to bring the package closer to the “target,” a stockbroker in Sharon, Massachusetts. But there was a rule: The "starters" could only mail the package to someone they actually knew personally on a first-name basis. And that person, in turn, would repeat the process until at some point the mail would finally reach the intended destination. Hopefully. There were some other elements in the experiment, including “tracer postcards” that each person linked in the emerging chain would mail back to Milgram so he could map the movements of the 160 packages. Remember, this is 1967, long before anything like the FedEx or Amazon tracking that we all now take for granted.
At the outset, Milgram and his colleagues speculated it might take even hundreds of steps for the 160 packages to journey from Omaha to a stranger in greater Boston. It took an average of just six steps — as few as two, and no more than 10. And thus the meme was born: six degrees of separation.
Now there are some problems with Milgram’s work, and it has been criticized. If you’re inclined to math, think it through. Simplistically, if you have 10 friends and they each have 10 friends, then in one degree of separation you should be able to reach 100 people. Right? But there’s the matter of “clustering.” Probably five of my 10 friends are friends in common. If so, my presumed network of 100 is suddenly just 50.
But whatever the criticism, a 2008 study by Microsoft of its Microsoft Messenger service found that the average chain of contacts between users and the service was only 6.6 people. So many, including sociologist Watts, who is one of the great pioneers in network theory, conclude that Milgram’s work has stood the test of time. You’ll find links to some of Watts’ great work on the Digital Companion.
This now brings me to my own take on networks. In the past ten years, I’ve seen something north of 3,000 pitches from would-be entrepreneurs. The best among them talk about their business ideas and the network effects they are striving for in a hopeful way. After all, the subject of networks is a hot topic. But most of the time the discussion is just that — hope. And hope is not a strategy. Behind this, of course, is the fact that we often talk about building networks, as everyone is inspired by the growth of social networks and other digital endeavors like Uber or Airbnb that have scaled their networks geometrically and, in some periods, even exponentially. And that’s great. However, I want to encourage you to spend more time examining and identifying the networks that already exist. Networks abound in social interaction, in the organization and evolution of cities, and in nature... to the way starlings flock, the way ants build complex structures, and to the way — as we unfortunately now know so well — that viruses mutate and spread. Watts’ work, in fact, began with a study of the ways crickets coalesce into a vast ocean of synchronized chirping via natural network nodes.
This is why when we consider investing in startups at Hurt Family Investments (124 startups and counting in our portfolio as of this writing), we always look for natural network effects. It doesn’t mean we won’t invest when natural network effects aren’t present, but it is certainly more enticing when they are. I use the word natural deliberately because it is far easier to build solutions that will offer network effects if your market is indeed wired that way. Not long ago I attended a discussion at the University of Texas at Austin between computing pioneer Michael Dell and author Walter Isaacson, at the time CEO of the Aspen Institute. Isaacson’s many books include Steve Jobs, Einstein, and The Innovators: How a Group of Hackers, Geniuses, and Geeks Created the Digital Revolution, and he is now working on one about Elon Musk. As you might imagine, it was a broad and fascinating conversation. But the key takeaway for me was Dell’s remark that the level of difficulty in changing an industry’s behavior depends on how long that industry has been engaging in that behavior.
Now that might not seem a revolutionary insight when taken on its own. But it’s foundational when you pair it with natural network effects, which lead you to what Harvard Business School authors Marco Iansiti and Karim R. Lakhani have called “strategic collisions.” Their book on this and other topics, Competing in the Age of AI - Strategy and Leadership When Algorithms and Networks Run the World, is worth picking up and I’ve included parts of it on the Digital Companion. Its essence is all about networked companies colliding with those that are not — the confluence creating the power to accelerate change.
Think of how stodgy the hotel sector was from 1925 when Conrad Hilton innovated with on-premise retail shops and dining rooms at his first hotel in Dallas. Then Airbnb and Homeaway showed up eight decades later. The business of ferrying passengers by car didn’t change much beyond the trend of painting most of them yellow beginning in 1915. Then Uber and Lyft came along almost a century later. In both cases, industries almost defined by inertia were transformed overnight by the natural networks — constellations of spare rooms in one case and unused private car capacity in the other— that had been there all along, just waiting for the technology and insight to wire them together.
In my own entrepreneurial journey, I’ve experienced and utilized this power of pre-existing, natural networks in virtually all of my businesses, with the exception perhaps of the first which was in consulting (although I did leverage the Wharton alumni network well to identify and win new customers). And of course, this is a big part of the ongoing story at data.world where our data community now numbers nearly 1.6 million analysts and data scientists collaborating as a natural network to confront challenges from climate change to smart policing. But the most illustrative is the case of Bazaarvoice where the market was, fortunately, wired for brands and retailers to collaborate in a kind of grand bargain.
The retailers were to provide the audience — and therefore the sales — that the brands needed. The brands were to provide the co-op advertising to support the retailer in the endeavor through what the trade calls market-development funds (MDF). This is a major part of the commercial ecosystem as co-op advertising dollars are as high as $50 billion in the US and a staggering $520 billion worldwide, according to the Altimeter Group, a researcher of disruptive technologies. Brands would get higher margins but lower revenues, while retailers would get lower margins but higher revenues (generally speaking, with Apple being a notable exception in owning both sides of the deal for much of their sales). So it was natural for Bazaarvoice to tap into this network effect and provide solutions for not only retailers, but also the brands that sold through them.
The result is that Bazaarvoice actually had — and very much has today — a working network effect that benefits all participants: retailers, brands that sell through those retailers, consumers that shop at those brands and retailers, and of course Bazaarvoice itself and some of its partners. In other words, the more participants that are on the Bazaarvoice network, the greater the effect of that network to the benefit of all. They were just waiting for an introduction. What we did was bring them together and then allow them to amplify one another and grow together in this newer and increasingly more digital aisle.
Another way to consider the case of Bazaarvoice is as the “keystone” in the network we created. When defining and studying ecosystems, which are sets of complex interlocking natural networks, biologists look for the “keystone species.” These are the species that play an outsized role and sustain the overall system, even if their relative population is small. My favorite example is sea otters, which dine heavily on sea urchins. Because of the sea otters, urchins gather in protective crevices and consume smaller organisms, and leave the large kelp forests — depended on by many fish and invertebrates such as snails — untouched. Bazaarvoice became the “keystone species” in its particular corner of the retail ecosystem - again, benefiting retailers, brands, and consumers alike (the kelp forest of sorts), which is the role I encourage all entrepreneurs to seek for their startups.
Now it’s important to stress again that networks and the synergies they spawn are important across all facets of life and enterprise. One of my favorite entrepreneurial examples of leveraged network dynamics is that of my fellow Austinite Mike Rypka, who started a food truck selling tacos in 2006. Austin, with its university, tech, and music scenes is big enough to have plenty of taco demand. It’s also interconnected and small enough for word of tasty tacos to spread rapidly across Milgram’s six degrees of linkage. Today, Rypka’s pioneering of what has become known as “gourmet street food” has led to 75 brick-and-mortar outlets across Texas and four other states with revenues in 2019 of nearly $300 million. Although I’m more of a Tacodeli fan myself, Rypka’s Torchy’s Tacos is incredibly popular and successful (some inside baseball: this statement is like starting a war in Austin as you are either a Tacodeli or Torchy’s person).
Or consider the origin of the Beatles, whose two living members, and the estates of the two members no longer with us, made almost $70 million from their music and rights in 2019. “We were four guys who lived in this city in the North of England, but we didn’t know one another. Then, by chance, we did know one another,” Paul McCartney wrote in an essay for The New Yorker that caught my attention in late 2021. “To this day, it is still a complete mystery to me that it happened at all.”
A tale of a random, natural network indeed. Technology, however, is clearly where we see this “mystery” of surprising networks playing out most vividly. And in the business area with which I’m most familiar, software-as-a-service (SaaS), I see it all the time.
Most of the SaaS startups we’ve invested in (either as advisors or investors, or both), have natural network effects, with some inherently stronger in their industries than others. It is one of the key ingredients we look for, such as when we seed-backed ZenBusiness (Austin’s latest unicorn as of this writing). I think any SaaS startup would be wise to identify if its B2B market has a natural network effect to tap into. Not long ago, for example, I engaged in brainstorming with a SaaS startup team that wasn’t sure about this premise, but by the end of the session, we had all convinced ourselves (and not just hopefully, I might add) that indeed a natural network effect existed. That made all of us much more excited about the business than ever before.
A SaaS startup without a natural network effect can still be successful. For example, there are replacement-market SaaS businesses like Salesforce, which started out by disrupting a massive market (Siebel and CRM) with a better, cheaper, faster, and ultimately more function-rich solution. Salesforce succeeded because it built on a single platform versus spreading its R&D over multiple client computing platforms and armies of installation consultants. Workday is another example, running the same play as Salesforce did against Siebel, but in this case, it was Workday against PeopleSoft/Oracle. In fact, replacement-market SaaS businesses have grown much quicker than nascent-market SaaS businesses because the market already existed and can therefore be easily sized and disrupted (assuming great execution). Their success is about execution in sales, services, and feature parity … and then eventually superiority — again, due to that single platform to rapidly evolve on advantage. But now SaaS as a superior business model as compared to on-premise enterprise software secret is mostly out of the bag. Workday enjoyed a monster IPO and Salesforce.com is the most valuable SaaS business in the world. Snowflake is a more recent example, displacing on-premise Oracle installations left and right and having the largest SaaS (or software, period) IPO in history in 2020. But these replacement SaaS startup opportunities are few and far between. This means that most new SaaS startups we see are, in fact, nascent and therefore are trying to create their own demand, which requires both a keen understanding of natural networks and a lot of evangelism and education. Evangelizing in a nascent, or fledgling market is a topic I’ll get to later, in Chapter 16, Selling to the ‘Cool Kids’. I will also share ways to hire some great evangelists as full-time and part-time employees, and also as Advisory Board members, in Chapter 13, How to Leverage Advisors and Investors as Your Extended Team.
Back to the topic at hand, the bottom line is that when we see a SaaS startup pitch us, natural network effects in their industry (or the lack thereof) are among the first things on my mind. It doesn’t mean they will be successful if their industry has the embedded advantage of a natural network effect, but it will certainly help. In investing in B2C, we also look for network effects but by comparison, they are usually those created rather than those that naturally occur. For example, Apple created a network effect by the integration of iPhones and iPads to the App Store and to iTunes. Apple then tried to extend this network effect into social with the launch of iTunes Ping, which was one of their biggest failures in recent years. Amazon created a network effect through the launch of Prime and the subsequent launches of Prime Instant Video, Prime Photos, and Kindle Owners’ Lending Library. Facebook created a network effect by the launch of Facebook Login, allowing it to proliferate everywhere, and they also tapped into a natural network effect by launching at select universities (where students were naturally connected, the way it has always been) and expanding out from there. LinkedIn tapped into the natural network effect of professionals working together at companies (both as peers and as partners). These are just a few notable examples.
As a final thought, it’s also worth mentioning that networks, by their nature, are emergent. That’s where their mystery lies. And as such, they can be harvested to your advantage but also surprise you. Nokia, as many readers may recall, was riding on top of the world, with nearly 50 percent of the global smartphone market in 2007. That was the year that Apple launched the iPhone and we all know the rest of that story. Or consider the iconic Kodak founded in 1892. Although Kodak was successfully navigating the transition to digital photography in a plan that had begun in the 1990s, it was bankrupted in 2012. And not by a peer competitor like FujiFilm, but rather by camera-ready smartphones that became a networked and networking “black swan” that really ended the legendary firm as collateral damage.
Wherever you are on your entrepreneurial journey, think networks. They are all around us — even on your Peloton right now if you are riding it while reading this.
“Networks are present everywhere. All we need is an eye for them.”
— Albert-Laszlo Barabasi, physicist, and network theorist