Lessons of Greatness: DESIGN for Network Effects
Anu Hariharan: How to think about network effects for your startup, from the beginnning
📣 Announcements
👋 Matt Mullenweg, founder and CEO of Automattic and WordPress, is joining me on Clubhouse Tuesday, 5/25 at 5:30pm PT. I interviewed him on the podcast two episodes back—I’m excited to continue the conversation with him and I would love to see you there.
✨ So you never miss a beat, subscribe to this newsletter and follow @m2jr (me) and @floodgatefund on Twitter.
Now, it’s go time with Anu Hariharan.
Introducing Anu Hariharan
Anu Hariharan is a network effects guru and partner at Y Combinator Continuity Fund. To this day, Anu has created some of the very best content I've seen on how to design networks. And today, I'm excited that founders throughout the world will get to learn from her wisdom and apply it to their own startups as they seek greatness.
Network effects. It's perhaps the most important value proposition and competitive advantage in business today. Understanding network effects not only helps you build better products, it helps a company build moats that protect its franchise often over the course of many decades. But what is a network effect? And do they happen by accident or can networks be designed proactively? The answer is yes, they can be designed. And one of the foremost experts in how to do this is Anu Hariharan.
If you've ever wondered how you can create a network that becomes a winner-take-all scenario, Anu will help you learn the state-of-the-art approaches to help you connect the dots.
✨ Follow Anu @anuhariharan on Twitter.
📚 Anu’s recommended readings:
🎙 The Full Episode
Listen on Spotify
Listen on Apple Podcasts
Read the full transcript of the episode here
Teaser below:
🌟 Highlights from the Episode
Golden Line
“If you, as a founder, pay attention to whether your product has the potential to have network effects, it can really give you a market leading position relative to all the competitors.” —Anu Hariharan
What is a Network Effect?
The simplest way to define a network effect is as more users join a network, the service becomes valuable to all the users, including the existing users.
Think of Airbnb, it's a two-sided marketplace. You have hosts on one side, you have guests on the other side. As more guests join the platform, more hosts are able to get their rooms booked and make money. And therefore, more hosts joined the platform. As there is more availability, more guests joined the platform. So it's literally as more users joined the platform it's useful to everyone.
Importance of Network Effects
It has the ability to tip the markets to a winner-take-all. And if you are able to get to winner-take-all, it translates also into software margins and pricing power in the market. Therefore, if you, as a founder, pay attention to whether your product has the potential to have network effects, it can really give you a market-leading position relative to all the competitors.
What can Entrepreneurs do to Develop a Network?
When we did the deck at Andreessen Horowitz, we spoke to a lot of founders to understand whether in the zero to one phase, did they realize that they have potential to network effects? And I'd say the answer was in the middle.
If you look at Facebook, I think the founders of Facebook knew they were building a network from day one. And you can see that by the way they launched the product. The first question you have to ask yourself is, does your product have a utility? In the case of Facebook, the utility was it's a class directory. And if you ask some of the Facebook users from 2004, which were predominantly university students, they would say, especially in freshman year, it was a great way to get to know the rest of the class, because it could be quite intimidating to know who are the other folks in your class. And Facebook just made it really easy.
They also paid attention to, do I need to be a cluster in each university? Or should there be a cross-collaboration. The reason they asked that question was not because they thought about network effect, simply because there were other competitors in other universities. And so, they asked themselves, my product has a utility, but how do I get it to scale? How do I get students from other universities who have a competitor product to use us?
📌 One of Anu’s recommended readings, Harvard Business Review’s Why some platforms thrive and others don't?, examines Facebook’s network and adds support to Anu’s insights:
Economists have long understood that digital platforms like Facebook enjoy same-side (“direct”) network effects: The more Facebook friends you have in your network, the more likely you are to attract additional friends through your friends’ connections. Facebook also leverages cross-side (“indirect”) network effects, in which two different groups of participants—users and app developers—attract each other.
Metrics for Network Design
A metric that is a measure of network effect is usage, not growth. This is the number one mistake most of them make, which is, they assume if I'm growing really fast and I'm growing organically, I have a network effect. You don't because that's just wild growth. Think of Angry Birds. You grow really fast doesn't mean if the 10th player joined and started playing Angry Birds, it doesn't make it more useful or effective for the first user that joined.
But in Facebook's case, the number one thing that Zuckerberg paid attention to was retention. So they tracked what percentage of the users in a university logged in daily. But I think Facebook was one of the early platforms to pioneer something called a smile curve: they literally plotted from day one through day 30 in a month. And they would say, what percentage of the users logged one day of the month, two day of the month, three day of the month? The goal that Facebook was trying to get to was how can I get a smile curve that is majority of the users were logging in 30 days of the month? So it was a product that would be used effectively every day.
I think the focus on the smile curve was not just to boost the numbers, but it helped them think through what product features do I need to build so that I can increase the utility of the platform to a daily use case. That is a very important distinction, and I think the Facebook founders really paid attention to that. They may not have thought of network effects, but the focus was on retention.
Network Effects vs. Virality vs. Economies of Scale
The definition [of Network Effects] is very simple, which is as more users join, the service is more valuable to everyone. And I think that everyone is important. People often forget that. It also has to be useful to also the existing users.
The other definition for network effects, a lot of research scientists refer to it and economists refer to it as demand-side economies of scale. So if you look at any of the papers on network effects, they don't call it network effects. They call it as demand-side economies of scale, whereas economies of scale is usually supply-side economies of scale.
People more often than not confuse vitality with network effects. Virality means there's massive adoption, rapid growth without spending anything for acquisition of users. Say when Angry Bird launched, they grew exponentially with literally no marketing dollars. And I would say the same thing was true for Pokemon GO recently for at least the first few months, a lot of users were attracted to the product.
You get word of mouth, that's virality, which a lot of people are referring because it's just cool or it's an excellent product to use or it's catchy, whatever it is. But even though more users are using it, it doesn't mean it's valuable to the existing users.
Advice for Founders to Design for Network Effects
I think that there are four to six basic questions any founder should ask when they're building something, even from day one.
One is, what is the product's utility? And does it have the opportunity to connect different groups of users to form a network?
In the case of Facebook, it was clear the utility itself was the social connection.
In the case of Airbnb, it was guests connecting to hosts, which is a traditional two-sided marketplace.
In the case of Skype, it was placing phone call to each other, so they were just homogeneous users.
In the case of Open Table, it was restaurants connecting to diners.
If you pay attention to [if you] have a network, then you have three other sets of questions you have to ask, [one of which is:] once you know you have a network, what is my growth tactic? Because you can't just build it and expect the users to come. Once you have a growth tactic, you need to figure [out how to] get to critical mass, because there's no point having potential for network effect and not getting the critical mass.
In the case of Zapier, what is the growth tactic? Well, what viral loop mechanisms can I put in place via the product to scale the number of users as well as apps that integrate with Zapier? What they did was amazingly brilliant. They had a landing page for pretty much everything. So if you were searching online for an app to integrate with, there would be a page about that. There would be a separate SEO page about what the integration is. And then there would be a separate page about the use case of that integration.
3 Laws of Network Effects: Sarnoff’s, Metcalfe’s, Reed’s
Sarnoff's Law is a broadcast. The value of the network increases proportionally to the number of viewers. Think of Yahoo, because there is no bidirectional connection. This broadcast is more like I'm broadcasting information for anyone who wants to see or listen. And so, the value of the network does exist, but it's limited in terms of, it's just proportional to the number of users.
Metcalfe's is a true network effect. The value of the platform increases non-linearly as more users join the platform. Facebook is literally one-on-one friend connection, and you can see that the value of the network increased as more users join because you made more friends.
Reed's Law is a group forming network. The value of the group forming network is like an exponential curve. It increases within both the number and the ease with which the groups are formed. WhatsApp is a great example of this because… it was this app that you could download anywhere, no login required, address book was hacked, and you suddenly launched the cluster of close connections in groups. That's why when you look at WhatsApp's growth curve, actually, if you look at the year one to year four versus Facebook, WhatsApp grew three times faster than Facebook.
📌 On the topic of group forming networks, Anu recommends reading The Power of Community Building which states:
As the Internet continues to expand, investments in Group-Forming networks are likely to produce the biggest returns.
Anu’s Entry Strategy
If you think you're building a network or you think you have the potential for a network, you have to find a beachhead. If you try to go all in across the US, you won't really be able to find that network. Here's why. Think of network as nodes and clusters. If you think of clusters, clusters are usually your closest friends, which more often than not tend to be geographically concentrated. If you think you're building a network and you just say, "I'm just going to do paid advertising on Facebook and try to get my first 1,000 users," you're doing a disservice to yourself, because you don't even know whether those 1,000 users have connections.
That's why it's very important to figure out where your cluster is. That's the entry strategy. How do you figure out what my cluster is? For Facebook, it was university. For Airbnb, it was events. If you're a founder in year one or year two, your job is to first get 100 users that love you, 1,000 users that love you. By year two, or by the time you have say three or five universities or 11 universities, you can figure out what drives retention.
I often tell founders, the retention metric is not something that an investor sets. The retention metric is for what you have designed and developed, how often do you think your user should use it?
Words of Wisdom
The more cash you have, it's a detriment to building a network effect. You'll end up spending on paid marketing because you think that's a way to build the network, and you're not using or improving. Also, I think more people distract you from building the network. If you hire very quickly a 10 or 15 people team, you might be doing a disservice, and I think it's super important.
If you are extremely lean, maybe you're just three to five people, but you're really tweaking the product and paying attention by what is it this network of users really need? How do I replicate this cluster? The better the quality of conversations and the better the rollout of the features and better the experimentation.
The other question I would ask founders to really pay attention to is, barrier to exit for the user. We often think about barrier to entry. Nobody talks about barrier to exit for the user. This is something I learned from Marc Andreessen. If you look at Facebook, the barrier to exit for the users really high. That's why when Google plus launched, nobody switched. It was a pretty good product, but if I had all my connections, my inventory of photographs, my history, it was really hard for me to switch, right?
Finally, I wouldn't rush it. People often think that when they see these stories about WhatsApp or Facebook, which probably was more true for Facebook, they think it happened very quickly and accidental. A lot of network effect businesses take a long time.
💌 Enjoyed this Episode?
If you enjoyed this episode, we’d be grateful if you could leave us a review on Apple podcasts and Spotify.
If you found this episode insightful, you might also enjoy my interview with Reed Hoffman, who I like to call the Network Philosopher King. In it, we talk about how Reed built LinkedIn is a network from the ground up by design. Here's a preview of Reed's insights:
What most matters to other people? People. Sure, they like places. Sure, they like things. But we're social animals, focus on people. How do you make people better with each other? That's the thing to do and that's part of where LinkedIn came from.
Questions / Comments / Have a favorite quote or moment from the episode? Send us an email at greatness@floodgate.com