The Second Order Network Effect

The web is all about building networks on top of networks. The Internet is the base network, a global connection to 825 million people (that’s comscore’s Jan ’08 number and it doesn’t include mobile devices).

We’ve been building networks on top of the Internet network since we first got the Internet, but the creation of social nets in this decade has been transformative. We now have “social” networks like myspace and Facebook which each reach over 100 million people a month. These networks have rich social graph databases that know who we care about most among the >1bn people hanging out on the Internet around the world.

So that’s why Facebook platform was such a big deal. They allowed developers to build apps that can tap into the Facebook social graph. The top three Facebook app developers, Slide, Rock You, and Zynga have built networks of Facebook apps that respectively touch 4mm, 2.2mm, and 1.3mm people each day. And they’ve done that in less than one year. That’s not yet the kind of numbers an event like the Oscars can produce, but these numbers rival many cable networks daily viewers.

What’s happening now is that these app networks are becoming networks themselves. They are allowing app developers to join their networks and tap into the people hanging out in them. Because these app networks were built on top of social graphs (first Facebook, then Bebo, soon myspace), these networks are able to leverage the social graph to provide social network utility to app developers.

I call this a second order network effect. Building networks on top of social networks produces even more utility to app developers.

I am an investor in Zynga and saw some data on their new social game network that I found very interesting. With their permission, I am going to show you how and why this is happening.

But first, a short explanation of what a social game network is. Zynga makes social games. Examples are the word game Scramble, the hugely popular Texas Hold’em, and the war game Triumph. They all are connected via the game bar at the top of the app which shows your friends (from the social graph data they have access to) and what game they are currently playing. The game bar is the “game lobby” of the social game network.

Last week Zynga launched their social game network via the Zynga API.  To date fourteen game apps have joined with another ten on deck to launch this week. This about triples the number of games in the Zynga Game Network.

Here’s a screen shot of one of the games that joined in the first week, called The Dot Game. It’s a fun social game where you connect dots in a race against a friend to cover the board.


You’ll see at the top of the screen shot, right below the Facebook toolbar and above the game itself is a section of screen real estate that shows my friends on Facebook who are playing games in the Zynga network and what games they are playing. If anyone is currently active in a Zynga network game, it will show that and what game they are playing.

In the first week of the Zynga Game Network, almost 75,000 clicks were sent to third party apps that have joined the network. And those third party apps sent back about 32,000 clicks to the Zynga Game Network. It’s important to note that the Zynga Game Network is currently sending more clicks out to its partners than it gets back. More on that in a minute.

Here’s a screen shot of the Zynga Game Network dashboard for game developers. It shows for each third party app in the network how many clicks they are getting and how many clicks they are giving.


You’ll see that on the first day the network generated a lot more clicks to The Dot Game, about 3,000. The next two days as the network attempted to balance out the clicks a bit, the network got back more than it gave. On Friday, the network started giving more again and has been doing that ever since.

Zynga is initially giving more clicks to its partner game developers than they get back to help jump start other games. Having more successful apps in the network means more clicks for everyone. That’s the second order network effect in action.

How valuable is a click? Well Zynga’s experience has shown that early in the launch of a new game, most clicks result in an install and a game play session. Of course not every game play session turns into a repeat player. That’s where game design and evolution comes into play.

Here’s some data on two of Zynga’s own games, Triumph a Zynga developed war game, and Scramble, a Zynga developed word game.



These charts show the daily active users for both games.  Zynga has had to take different tacks with each of these games. Triumph is a real time game that they continue to evolve and improve resulting in more repeat usage and growth. Scramble is primarily a “turn based” game like Scrabulous. The secret in a game like that is viral tuning to increase the number of people that you play the game with.

So if you have a social game that you’ve built, getting clicks is not the only secret to success. But it is a big part of it, particularly early on in the life of a game. Joining a game network like Zynga a free way to get clicks. The Facebook directory has over 17,000 apps in it. It’s very hard to get noticed there. You can buy installs and many apps do that. And with Facebook cracking down on invites and notifications, it’s even harder to build an audience for your app. So if you are a game developer you should also join a game network and though I am clearly biased, I think Zynga is by far the best choice.

Wikipedia defines “network effect” as:

a network effect is a characteristic that causes a good or service to have a value to a potential customer which depends on the number of other customers who own the good or are users of the service. In other words, the number of prior adopters is a term in the value available to the next adopter.

That last part is key. The “number of prior adopters” is the value to “the next adopter”. So if you are picking a game network to join, go to the one that has the most game players in it. You’ll get the most clicks that way.

By interconnecting many of the Facebook game apps, Zynga and other open app netwworks are creating an open channel for everyone to participate in, so that companies don’t have to compete to own and control separate networks. Like the Facebook itself, all networks have more value with more nodes.

The biggest winner in all of this is the user. When you play social games that are in an open social game network, you will see and can play with more of your friends. And playing games with your friends is what social gaming is all about.

#VC & Technology

Comments (Archived):

  1. NGutha

    Fred, what are your thoughts on monetizing this?I agree, n-order networks have a great deal of potential but most of the discussion on this subject reminds me of the late 90s when the metric of choice was “unique eyeballs”. Not saying that you are trying to justify some f-ed up valuations or even validating the business models but I think that the space is getting mature enough that this needs to be the next focus.On a related note, I can see a large amount of deal volume in this area right around the time that monetization starts to become a reality but before it is fully implemented. Personally am a little skeptical of this being a advertising-only revenue source medium for a few reasons, would be interested to see what the other options are out there.

  2. spraycode

    “The “number of prior adopters” is the value to “the next adopter”. So if you are picking a game network to join, go to the one that has the most game players in it. You’ll get the most clicks that way.”There has been a lot of talk about these social gaming networks over the last few weeks and the conversation seems to inevitably turn to how developers might choose between them. Wouldn’t the best strategy for a game developer be to join as many of these social game networks as possible? Wouldn’t it be best for Zynga, SGN, or whomever to have the games in their platform getting popular through as many channels as possible?I don’t think Zynga, SGN or any social gaming network are ever going to be able to “lock in” games to be exclusive in their network. In console gaming the vast majority of games are available on every platform. The exceptions are tend to be lucrative exclusive deals and first party development efforts. Another exception is when the technology of the platform is unique in a way that a game might not necessarily be easily transferable across systems. Many Wii games just would not make sense on Xbox 360 or PS3. Looking at the Zynga developers documentation it looks like the only thing you need to add to your app to join is a couple lines of code to display the toolbar. Hardly a barrier to joining multiple networks.Lets say a social gaming network has an exclusive game that a user gets hooked on. They are more likely to further engage that user by offering many of the same games that user may like to play within other social gaming networks. More importantly, for social games to grow beyond cute retreads of board games there needs to be incentive for developers to invest their time ad creativity in this market. Having as many channels of adoption as possible open to them is critical to accomplish this. The better these games get the more the social gaming networks benefit.

    1. Zach

      I have to agree with the above. As long as the game networks are not mutually exclusive, it is in the best interest of the developers to put their games on multiple networks. Unless the networks give an incentive to the developers to remain with their network (monetary…) the developer has no reason to stick with the network and the network will have trouble differentiating itself.

  3. Jeff Jarvis

    This might sound trivial but I think it’s not: It’s actually a third-order network effect. The internet is the base; facebook is the second order; Zynga is third. That’s important, I think, because it illustrates a fundamental architectural shift in media and the economy.

    1. fredwilson

      You are right jeff. My bad. I may do a follow up post to correct thatFred

    2. vruz

      oh right, but isn’t the web another abstraction layer too ? the web works atop the internet. that would make them four, wouldn’t it ?

  4. Dick Costolo

    Awesome. This has got so many monetization opportunities beyond advertising, it makes your head spin. There are so many different ways that you can harvest the influence created by the popular games within networks. The affiliate opportunities alone should be a home run. If the entertainment industry didn’t have its head stuck in the sand, they would read this post four times and go out and do this for media. It would be so easy for them to do it, but they just can’t get their brains around indirect business models (in which the end consumer isn’t paying an explicit amount of money for a piece of content).

    1. Mike Dever

      Dick, Specifically, what are the head spinning monetization opportunities you refer to?

  5. vruz

    I think I like the “Second Order” moniker better than “web 2.0”. It gives a better idea of layered abstractions.Now perhaps this is in actuality the Third Order network effect, because the web circa 1995 in the Yahoo incarnation of yesteryear was a big leap forward over what we had at the time, I think that would be the layer immediately above the Internet, and the social abstraction would be the layer on top of it.

  6. Tom Hughes

    Some of these games then become focal points for tribes, clans, etc; that seems to be the next wave. “Triumph” (which I hadn’t heard of) seems to attract a lot of “join us, we’re the best” traffic. I imagine someone is already working on clanware, but in case they’re not: clanware is/would be a system for clan members to collaborate and share information about game(s) they’re playing, independent of the game itself. You can see this being done inside Facebook, e.g., for World of Warcraft; but I don’t believe there’s an application yet to allow WoW players to collaborate on their WoW playing via FB.

  7. Steven Kane

    Fred, how does Zynga calculate that 1.3 million number? When I add up their various app user numbers, I don’t get that. And what about duplication? How many texas hold em polayers are also scramble players? etc….

    1. fredwilson

      steve – here’s a link to the source of the 1.3mm number…it is not unduplicated. i’d like to figure out what that number is for all three companiesfred

      1. Steven Kane

        thanks, that is very useful sitebut shame on adonomics: it is so so easy to de-duplicate, and that kind of data hygiene is so standard in the world, that not de-duplicating is almost like lying, i think.or at least, if not one of commission, its a wicked sin of omission.i’ll bet anything zynga and the other app publishers de-duplicate regularly for their own internal purposes. but in a world where the undupliacted numbers are being passed off as accurate by third parties, of course no individual publisher will issue the true numbers.the new media metrics ecosystem is so messed up it feels almost certain there is going to be a meltdown or scandal

  8. WayneMulligan

    I think what we’re finally seeing with these 2nd degree/3rd degree networks is the emergence of true virtual communities. Joining a web site and participating a forum was the first step towards that. Building a network of relationships on top of that web site is another. But getting to the point where we’re micro-focused on shared interests (e.g. gaming and playing a specific game, thus exerting choice as to who you interact with and when) is almost the final step. To me, the final step in the process of forming true virtual communities is when these communities blur between online and offline.One of the companies doing that extremely well right now is’s when you take the inherently anti-social act of using a web site (I don’t care if you’re making friends on it and conversing with other people, you’re still home, alone in your bathrobe typing into a box that converts 1’s and 0’s into pixels) and make it inherently social.-Wayne

    1. fredwilson

      Yes, meetup is doing great thingsFred

  9. RickyYean

    As we continue to build more layers on top of these networks, what’s going to happen? Third-order applications on second-order Facebook have created a lot of fatigue, and there are many more steps to get to these higher order networks. We have to bring the higher-order networks to easy-to-access distance.

  10. Hounar Malak

    I’m wondering with all the complaints, problems, glitches, and developer silence, what the Triumph user chart looks like now. The game is dying, and they don’t care.