Network Stacking: The New Model for Growing Online Networks

Our co-Founder, Eric Jackson, wrote a guest post for Wedbush on the topic of “The Second Internet”

A decade ago, PayPal pioneered the concept of the “stackable network” by growing its network on top of eBay’s. Today, online networks are leveraging the giant social networking sites to follow in PayPal’s footsteps, making it easier and cheaper to scale.

by Eric M. Jackson, CEO of CapLinked and author of The PayPal Wars

*   *   *

Facebook might be The Social Network, but it’s not the only network on the web. To the contrary, online networks have proliferated in recent years. Yelp, YouTube, Zynga, FourSquare, Quora, Reddit, Yammer, BranchOut…the list goes on and on.

Unlike software as a service (SaaS) applications, which simply need to charge more in subscription or purchase fees than they do to acquire new users, online networks face an entirely different set of economics. Most users don’t pay to join networks, but their involvement makes the network itself more valuable. Revenue usually comes from several sources, such as advertising or premium features. But since revenue often cannot be generated on a significant scale until the network is quite large, user acquisition becomes a critical challenge for network sites.

Robert Metcalfe famously noted that the value (V) of a communications network is equal to the number of its users (n) squared (i.e. V = n²). In other words, the value of being on the network increases faster and faster as more users sign up. Telephones were the first example. As the number of telephones grew in the early 20^th century, the value of having a telephone also grew. Today, phones are so useful that most of us would never dream of leaving our house without a mobile one in our pocket.

This “can’t live without it” state is called the “network effect.” It means that a network has grown large enough that it’s become very valuable for its users. This makes acquiring new users easier because the value proposition has gone up. It also makes being displaced by a competitor less likely, since growing numbers of users are already locked-in.

Creating an online network has traditionally been difficult. The Catch-22 that online networks face on the path to achieving a network effect is that they don’t have much value to offer when they’re small. Until recently, the proven method for growing an online network had was to have a viral product, good marketing, lots of cash, and a bit of luck. But now this has changed. This chicken-and-egg problem has been overcome, or at least mitigated, by network stacking.

In Viral Loop, Adam Pennenberg coined the term “stackability” to describe a condition where “a viral network can be laid over the top of another, each fostering the other’s growth.”[1] Penenberg cites PayPal as the web’s first stackable network, which makes it a logical case study if we’re to fully understand the benefits that stacking conveys to today’s online network.

PayPal, the Web’s First Stackable Network

As I detailed in The PayPal Wars, when PayPal launched in late-1999 the company was targeting personal payments such as splitting up dinner bills and reimbursing roommates. When I joined the company in December of that year, it already had a $10 “refer-a-friend” bonus program to encourage users to send money to their personal acquaintances. But the entry of several well-heeled competitors, including Yahoo and Bank One, prompted PayPal’s executive team to look elsewhere for growth.

eBay ended up being the answer. Over the course of 1999, the auction site grew by 350% to a total of 10 million users, and it averaged 3 million live listings every day. As I wrote in The PayPal Wars:

[CEO] Peter [Thiel] and the management team believed that we needed to find the fastest way possible to scale up PayPal’s customer base. If we could increase our number of accounts to reach critical mass before our competitors, the resulting network effect would freeze out any opponents… [We would] concentrate our efforts on this popular auction site where there was an obvious need for faster and easier person-to-person payments. We would build PayPal’s payments network on top of eBay’s marketplace.[2]

When the decision was made in early-2000 to focus on servicing eBay’s users, all of the company’s product and marketing efforts shifted to focus on this. The engineering team built a request money feature, so auction sellers could send their buyers an invoice. We developed special PayPal logos for sellers to insert into their auctions, and we later automated the process for them. We created “Pay Now” buttons that would dynamically appear in closed auctions to prompt buyers to use PayPal to send money to the seller.

The new focus worked. In spite of eBay subsequently rolling out its own in-house credit card processing service called Billpoint, PayPal surged from 100,000 accounts in January 2000 to over 1 million users just two months later. By the end of March 2000, the company was averaging 18,000 new signups and over $1 million in payment volume per day based largely on traffic from eBay.

The next two-and-a-half years witnessed a competitive back-and-forth battle between PayPal and our reluctant host. eBay had no desire to see an uninvited payment service sprout up on our its auction site, but as our fledgling network grew so too did our network effect. In spite of numerous promotional campaigns and restrictions aimed at PayPal’s appearance on the site, eBay was unable to oust PayPal and ultimately acquired the company in October 2002. By the time of the acquisition, PayPal had grown to 20 million users and facilitated over $16 million in payments per day.

The Benefits and Risks of Network Stacking

A close look at PayPal’s effort to build its network on top of eBay offers valuable insight into the benefits of network stacking. The strategy conveyed a number of benefits that were critical to PayPal’s growth. Specifically, network stacking:

1. Provides a centralized source of customer acquisition. By focusing on eBay, PayPal’s team was able to find users in one distribution channel rather than having to chase a fragmented market through mass marketing. In the days before targeted online ads, the proverbial Super Bowl commercial was a common tool for dot-coms trying to reach a broad audience. PayPal avoided that by knowing where to look for its customers.
2. Delivers users with similar interests who are highly viral. People with common interests tend to communicate with one another. eBay’s buyers and sellers had a common desire—getting transactions done—and this turned eBay users into highly viral PayPal evangelists. They were eager to talk about the service and promote it to others, accelerating the growth of the payments network.
3. Overcomes the Catch-22 of network effects. Stacking PayPal’s network on top of eBay’s enabled us to acquire new users before PayPal ever manifested its own network effect. As Metcalfe’s Law predicts, networks can grow quickly as their value increases, but there’s no guarantee that a young network will ever get to the tipping point where that starts to happen. Stacking helped mitigate this risk for PayPal and enabled the service to grow to a point where eBay could not displace it.

The case study of PayPal makes it clear that networking stacking can deliver many benefits to a young online network. But network stacking is not without risk. As PayPal’s example demonstrates, this strategy:

1. Promotes a large degree of dependency. As PayPal scaled it also became highly dependent on eBay. With over three-fifths of PayPal’s business related to eBay, diversification became a key priority for the company in the months before the acquisition. Had eBay ousted PayPal before it achieved a network effect, it might have been disastrous to the company.
2. Depends on a functionality gap of the host network. Before PayPal, eBay buyers had to mail paper checks or money orders to sellers, slowing down the purchasing process and lowering inventory turnover. PayPal flourished on eBay because it succeeded in filling a functionality gap. Had eBay managed to solve this problem on its own before PayPal achieved a network effect, it might have turned our auctions-focus into a failure.
3. Strengthens the host’s network. Building our payments network on top of eBay’s auction network solved a shortcoming on their site. But unlike most symbiotic relationships, eBay wanted to displace PayPal because our presence impeded them from launching their own payment solution. Given this adversarial relationship, it could be said that strengthening the host network exposed us to more risk.

None of these risks overshadowed PayPal’s need to scale to create value and achieve a network effect, prompting the company to continue to use eBay as a means of growth. But as time went on, it became increasingly obvious that network stacking had exposed PayPal to these dangers.

Stacking on Top of the Social Graph

PayPal created the web’s first stackable network, but it would by no means be the last. Flash forward to the present, and you can see that the stackable network strategy that PayPal pioneered is now being employed across the web, powered by the social graph.

In the past half decade, Facebook, Twitter, and LinkedIn have themselves achieved massive scale while also opening up their application programming interfaces (APIs), allowing third parties to import over user information. Online networks of all types now have the opportunity to acquire users by leveraging the social networking giants. Fledgling networks can draw on the “Big 3” social sites—as well as their Google, Yahoo, and Microsoft accounts—to post information and send invitations to the friends and connections of their users.

It seems that most of the major players realize the value in reciprocal value in making social graph information accessible to third parties. The commitment has been codified through efforts such as the “Open Social” common API (supported by Google, LinkedIn, and Yahoo) and Facebook’s “Open Graph” data-sharing system.

The scale of the social graph is far larger than the opportunity PayPal saw on eBay’s network. The Big 3 likely account for a combined 800 million people worldwide, with Facebook alone claiming a quarter billion active users.[3] If we assume that the average social graph member has 150 total connections (which could be conservative since Facebook claims its average member has 130 friends), we’re talking about 120 billion nodes in the social graph that currently exist and are available via the APIs of the Big 3. And that’s not counting the billions of contacts stored in Gmail, Yahoo Mail, and Hotmail.

Clearly, the social graph of today provides an immense growth opportunity. We’ve seen this over the past couple of years as networks have begun to scale at incredible rates. For example:

• Zynga: The popular social gaming network, which has used Facebook as a distribution platform, was started in 2007 and reached 100 million users just two years later.[4]
• FourSquare: By allowing users to post to Twitter and Facebook when they check-in at physical locations, FourSquare grew from 1 million in May 2010 to 7.5 million users in March 2011.[5]
• BranchOut: The professional networking service allows users to send invites to their Facebook friends; it grew by nearly 1 million users in just a week this June after making changes to its product.[6]

And these are only three examples of privately held companies whose networks have soared at rates unimaginable back in the early days of PayPal. Yelp, Quora, Hunch, SeekingAlpha, Yammer and dozens of other innovative online networks are growing at impressive rates thanks to the social graph.

We’re seeing this same phenomenon here at CapLinked. Our network for private investing has seen 3,500% user growth YTD, pushing us to 75,000 total accounts. Viral invitations have been the primary driver of our growth as users send invites to contacts in their email address books and LinkedIn connections. Our vision for CapLinked is to “map the investor” graph, and in less than a year we’ve made serious progress in this objective. This would not have been possible even 5 years ago.

The Implications of the Stackable Web

What are the implications of this? I believe we’re entering an era where targeted social networks, able to scale by network stacking, will thrive and fill in the functionality gaps left by the Big 3.

We’ve seen this in the past couple of years in several specific categories, including question & answer (Quora, StackOverflow), location (Yelp, FourSquare), and job searching (BranchOut, TopProspect). These “functional networks” aren’t as broad as the Big 3, but because their focus is more targeted they can create value in ways that the Big 3 can’t.

Look for many more functional networks to sprout over the next several years, fueled by the decline in startup costs and the allure of being acquired by the Big 3 or another major site. And while many of these startups won’t survive, a decent number of them will grow to the point of achieving a network effect thanks to network stackability.