In the late 1960s, the sociologist Stanley Milgram investigated what came to be known as the “small-world phenomenon.” He conducted an experiment to find connections linking people in the United States who did not know each other. He gave a letter to someone in Nebraska, with instructions that the letter had to reach a particular person in Massachusetts. The first person was told only basic information about the “target,” such as his address and occupation, and was told to give or send the letter to someone she knew on a first-name basis, and that person was given the same direction, to deliver the letter to the target as efficiently as possible. Anyone who received the letter would follow the same instructions until the target was reached. Through repeated trials, Milgram found that it took five or six “steps,” on average, to get a letter from Nebraska to Massachusetts. This formed the basis for the well-known concept of “six degrees of separation” (shown in Figure 3-2).

Figure 3-2. Degrees of separation

In his book Six Degrees (W. W. Norton & Company), Duncan Watts points out that certain paths were considerably shorter because the Milgram package reached a “hub” or “broker” node that shortcut the average six steps. Figure 3-2 illustrates this by putting an additional circle on the hubs or broker nodes and showing the different paths on the right and left side.

In the online world, using LinkedIn or a similar network, members can easily browse lists and profiles of past business contacts who now live in Boston or who are stockbrokers in Massachusetts, and see on a map how many steps it would take to get an email introduction forwarded to the stockbroker online through first-degree contacts (see Figure 3-2).For example, LinkedIn will tell you how many steps there are between you and your destination mailbox (as shown in Figure 3-2).

Online social and professional networks use the connections found in the offline world and magnify them proportionally. In the LinkedIn system, the intermediary roles of the connectors, mavens, and salesmen remain basically the same but are enhanced by orders of magnitude. Easy and speedy digital connectivity shifts the focus in online social networking from creating, mapping, and expanding your network toward finding new ways to exploit, capture value in, and monetize your own network (see social network mappings in Figure 3-3, Figure 3-4, and Figure 3-5).

Figure 3-3. Social network hubs are keys to connection

Figure 3-4. Ties connecting separate groups may be weaker

Figure 3-5. How brokers connect separate social units

According to an MIT Tech Review article,^[20] Marcus Colombano, a media and technology marketing consultant in San Francisco, read about a company he thought should be his client. He popped its name into LinkedIn and found he was connected to four people with contacts at that company. He wrote up a proposal and sent it to a friend who had a contact who knew the CEO. Four hours later, he got an email from the CEO requesting a meeting (see Figure 3-6).

Figure 3-6. Connections across a large network can have tremendous value and speed

Do these speedy introductions lead to revenue for the startups that create and run these online networking platforms? The surprising answer is yes. Perhaps even more surprising, the case of LinkedIn, examined more closely later in this chapter, gives us a pretty good idea of what online connectors, mavens, and salesmen are willing to pay for access to these visibly more efficient search and linkage networks, as well as an idea of when the network’s best move is to offer some services for free. Of course, this presents new choices for the online social and professional networking companies that provide these services and platforms. Should advertisers or sponsors pay the freight or “freemium” users who are willing to pay for a “leapfrog link”?

The Web and related technologies help online connectors, salesmen, and mavens reach an exponentially larger audience, more frequently, more easily, and almost instantaneously. In comparison:

Social networks moved online steadily, but businesses took different approaches and found themselves with different groups of people doing different kinds of things. Business networkers needed one set of features, whereas social networkers needed another. Varying preferences led people to cluster on different sites, to move from site to site, and even use multiple sites at the same time.

In the study of high-tech marketing and management of innovation, the Rogers Adoption Curve is often used to explain the rate of adoption of a new technology or product.

In 1962, Everett M. Rogers, in his book Diffusion of Innovations, gauged how populations adopt new products and technologies. He later suggested that five factors explain why some new products succeed with high growth rates and rapid customer adoption and others fail:

The power of Web 2.0 technologies coupled with a freemium n-sided business model shows up in all five of these factors for Facebook, compounding its advantages and triggering its hypergrowth rate of adoption.

The Rogers Adoption Curve provides a useful framework to evaluate where on the adoption curve your product or service is getting roadblocked. For example, Geoff Moore’s book Crossing the Chasm (HarperBusiness) identifies a marketing “chasm” between the early innovators and adopters and the mainstream. As mentioned in Chapter 1, Web 2.0 companies like Flickr have attempted to close that chasm by using a freemium strategy and social network effects.

However, there are no variables or factors in Rogers framework that directly take “social influence” and the peer-to-peer pressure of online social networks into account. It’s difficult to quantify the impact of viral marketing, costless distribution, or viral distribution on adoption or growth rate. Nor are there quantifiable and standardized growth and diffusion patterns so that different kinds of products, services, channels, and marketing strategies can be measured and compared. The Rogers Adoption Curve and model rely on some fundamental assumptions about new technology adopters and markets:

The Bass Diffusion Curve is an important update to the Rogers model because it provides a radically different framework of customer/user adoption rates based on social influence. Social influence is arguably the key factor completely missing from the Rogers Adoption Curve. It also uses an exponential form, mathematically speaking, which makes the modeling assumption that the process of diffusion is viral, as in an epidemic. The general shape of the Bass Curve reflects a type of propagation through a population with a modified power law distribution. (And again, for quantitative projections of how many customers will adopt a new product and how quickly, the Bass Curve has been regularly empirically tested and validated.)

Frank Bass introduced this model in 1969 in a famous paper titled “A New Product Growth Model for Consumer Durables,” published in the Management Science journal. The key concepts are intuitive and simple: some users start the process by adopting a new product independently of the actions of others; others are influenced to adopt the innovation when they see others using it, being socially influenced by those who have already adopted the product.

The simple version of the Bass model has only three major parameters—m, p, and q:

The level of social influence can be expressed as the ratio of q/p. If the ratio is low, fewer social “followers” or “imitators” are influenced by the early independent adopters. If the ratio is medium or high, a substantially larger proportion of the potential first-time adopters in the market are strongly influenced by the initial group of independent adopters.

A few consumer durable goods examples make the curve intuitively easy to understand. The Bass Curve has been used to gauge the adoption of washing machines, TVs, leaf blowers, and cell phones.

These kinds of examples provide analogies in making marketing predictions for newer technologies. So, for example, an analyst might estimate the diffusion pattern of satellite radio to be similar in shape to that for satellite TV but faster because of ease of replacement and lower cost.

But what about online viral marketing and distribution?

The size of the Oakland Hills, California, firestorm in 1991 that burned more than 2,000 homes had more to do with the hot dry season and the gusty Santa Ana winds that spread the fire than with the actual spark that started it. Structural and coincidental factors were at work. In a similar way, fads and social influence “cascades” are triggered by influencers or sparks, but they require the right conditions to spread with explosive speed.

Highly influential people—celebrities like Oprah Winfrey—can make a book an overnight bestseller. But thousands of computer simulations reveal that global cascades are more likely to be triggered by accidental and unintentional influencers within a receptive social network structure. A cascade needs just enough of a critical mass of easily influenced people to propagate a chain reaction, one degree of separation at a time.

Sometimes, personal characteristics make a difference. Rogers’ risk profiling and psycho-graphic profiling of early adopters, mainstream users, and laggards—combined with Moore’s explanation of “crossing the chasm”—does a good job of explaining many high-tech marketing successes and failures. However, online web social networks make it easy for a relatively small number of users—like you and me—to trigger network effects.

Do we know how to “seed” and shape these new product adoption cascades or epidemics?

Not yet, but current research findings seem to indicate that the raw number of early users and their first-degree “friend” or online “buddy” linkages, along with their social interactivity or frequency of contact, someday may be more critical in this “online cascade” viral distribution model than highly paid influentials such as celebrities or brokers.

LinkedIn and Facebook are great Web 2.0 examples of the newer models of applied online social networks. LinkedIn offers a fascinating online business application of small-world networks and social network analysis mapping brought online. In contrast, Facebook provides a natural example of large-scale diffusion and node-to-node cascading behavior in social networks.

In 2004, there were 12,000 requests for introductions per month; 85% of requests approved by all the intermediaries were accepted by the target recipient. By mid-2005, requests for introductions had reached 25,000, and acceptance rates had increased slightly to 87%. As the membership base grew, so did search activity. Members executed 2 million searches in April 2004. By mid-2005, that number had more than doubled to 5 million.

LinkedIn’s ability to track the electronic activities and linkages of its membership gives us new insights into online social networks. The Harvard Business School case study revealed an unexpected usage pattern:

“A friend of yours is a friend of mine.” The common saying echoes the feeling that most of us have: two degrees of separation can be trusted. After all, you know your friend directly, and he wouldn’t introduce you to someone he didn’t trust or think highly of online or offline. “A friend of a friend of a friend,” however, is three degrees of separation, and already we feel a little less comfortable because we are further from a direct link. On the other hand, LinkedIn realized that allowing the possibility of four degrees of separation between a requestor and a recipient exponentially increased the reach of the system. For example, mathematically, if each contact had 20 ties, a requestor could access 137,180 unique contacts with 4 degrees of separation:

So, although it might be awkward for the “friend of a friend of a friend” to pass along an introduction when he does not know the originator of the message or the recipient, anyone along the path could decline the request anonymously. In mid-2005, LinkedIn announced that monthly requests were 25,000 and acceptance rates were at 87%. Part of the reason for the higher acceptance rate was the added endorsement feature. Members could collect endorsements or testimonials from their contacts, describing their qualifications and performance specific to a particular position or project. In the Harvard Business School case, LinkedIn noted that members were 3 times more likely to select an endorsed member from a search list, and requests initiated by endorsed members were 25% more likely to be accepted by both intermediate connectors in the path and the final target destination.

By carefully mapping and separating LinkedIn users into three distinct groups—relationship managers, connectors, and focused searchers—we see that Guericke and Hoffman were able to create a virtual two-sided network system. LinkedIn offers free services to the 90% installed base of relationship managers, while networkers and contactors are willing to pay a per-contact fee, as well as a monthly subscription fees for highly desired premium contact and networking services.

We see a pricing strategy similar to the other freemium cases where one side of the market sponsors or pays for the other, for example, Visa credit cards, eBay, and Google.

Note the importance of the relationship managers. In a classic business model, they might be seen as troubling parasites—90% of customers are using valuable services for free? (They see ads, but nothing makes them click on them, of course.) In LinkedIn’s model, relationship managers may be getting something for free, but they’re also giving LinkedIn the key information it can sell to other people. The cost of supporting those free users, even when it’s 90% of the base, is still less than the income potential from the 10% of paying customers.

LinkedIn makes much of its money by selling members the ability to contact anyone in the network for a fee, through three programs of highly focused feature changes:

Figure 3-9 shows the features LinkedIn offered in July 2007, along with their price points.

Figure 3-9. LinkedIn’s many offerings, all built on the same basic set of personal data

LinkedIn epitomizes the Web 2.0 model of free access to users who help build critical mass, while charging those who seek to use that mass for their own larger purposes. LinkedIn has stayed slim, focusing tightly on a core mission of business connectivity and leaving richer but much more resource-intensive forms of social networking to other companies.

The hypergrowth of the Facebook user base continued as it developed, from 5 million online users in late October 2005 to more than 7.5 million in mid-April 2006.

Facebook enjoyed an exponential growth pattern much like the S-curve discussed in Chapter 1. It benefited from direct network usage effects: the more a user’s friends use Facebook, the more valuable it is to the user. Additionally, peer pressure and social influence fuel the “contagion” and word-of-mouth effect. If your friends are not using Facebook, why should you?

College students didn’t want to have to wait to see their friends show up and start using Facebook. So, speed of adoption is important to get to critical mass quickly.

Speed is also important in the race to reach a high enough market share to tip the market and lock out competitors. Facebook reached critical mass and tipped the market before other competitors entered the same space. The Korean social network site Cyworld seized a similar early-mover advantage in its national market. When first-movers in strong network-effects markets reach a dominant market share, they become hard to displace.

The usage patterns were impressive for the frequency as well as retention. Facebook’s Mark Zuckerberg explained:^[21]

Facebook had competition. MySpace had become the fifth most popular web site by number of page views. It had already grown as a music destination to include more than 350,000 bands and artists. Its growth rate was phenomenal as well; in October 2004, it had 3.4 million users, and one year later it had 24.3 million unique visitors—an increase of 600%. On July 18, 2005, the multinational media conglomerate News Corporation announced that it would acquire MySpace’s parent company Intermix Media for about $580 million in cash. As of August 2006, MySpace had broken through the 100 million account mark.

Although MySpace and Facebook are targeted toward social rather than professional networks, Facebook focuses on relational clusters and existing groups and circles of friends; MySpace is targeted at helping users find new friends—discovering social affinity groups related to music, bands, and artists. Facebook helps college students keep up with high-school buddies in other colleges, as well as peers from sports, music, extracurricular, and volunteer activities. MySpace supports much more customization of profiles than Facebook, which has stuck with a simpler look.

This is consistent with Zuckerberg’s creation of Facebook as a directory utility for people to keep tabs on the people in their lives and as a way to express themselves. One-third of Facebook users list their cell phone numbers, and a large number check their Facebook pages a couple of times a day. Of course, college students are concerned about how their online Facebook profile and photos appear to others. They’re also interested in identifying others on campus who have similar interests and in looking for potential dates.

By January 2006, Facebook had become so vital to the university lifestyle that eMarketer reported that incoming students were creating their Facebook profiles long before they even set foot on campus. By contrast, MySpace has a lower frequency of usage, and its users are doing different things on it—uploading music and playlists, and exploring new music and artists. College students might categorize MySpace with sites such as Last.fm, a popular site for finding new music.

Because the utilities and motivations for college students who join Facebook are different than for those who join MySpace, it’s not yet clear whether the two can coexist in equilibrium or will spark a competitive race—in the college market and overseas.

Mark Zuckerberg recalled that:

In early March 2004, Facebook launched at Yale, Columbia, and Stanford:

By June 2004, Facebook was serving 30 colleges and had about 150,000 registered users. Zuckerberg and his two roommates, Dustin Moskovitz and Chris Hughes, spent the rest of the 2004 school year trying to respond to the demand from colleges across the nation:

They were able to build their system with minimal customer acquisition costs. The popularity of Facebook made it something users signed into automatically:

In September 2005, when Facebook opened its platform to high-school students, college students were asked to send invites to their high-school friends, who could then invite others:

In 2004, Facebook was a free service and accessible to anyone with an “.edu” email account, limiting its use mostly to college students. Members of the network were encouraged to create a personal profile, including contact information, interests, and current course schedule.

Unlike MySpace or Friendster, Facebook made its members’ schools their primary networks and offered only limited access beyond that, making it a network of peers, rather than a random assortment of people. In the past few years, Facebook has opened itself to anyone who wants to join.

Facebook was fairly open by default: everyone at a member’s school could see his profile, but members could control their settings so that only their friends could view their profiles on the web site.

In May 2007, Facebook launched a set of application programming interfaces and services that allow outside developers to add their own ideas to the Facebook user experience. Marc Andreessen, a founder of Netscape, posted a detailed and enthusiastic analysis of the Facebook applications platform:^[22]

Facebook’s API offers its users a wide choice of new services, as shown in Figure 3-13. Users don’t have to wait for Facebook to provide new functionality for them—anyone can develop applications, and good applications will get shared across user networks quickly.

Figure 3-13. Facebook’s application editor

Perhaps more startlingly, the Facebook API also gives developers a place to see their ideas zoom across a huge number of people. The amazing effectiveness of this viral distribution is best told in the words of the iLike blog, shortly after its launch as one of Facebook’s first third-party applications:^[23]

Facebook’s API approach is somewhat different from that of, for example, Google Maps. Rather than support new functionality outside of Facebook, it adds new functionality inside users’ familiar Facebook environment. Working this way, Facebook can share its success with a much broader ecosystem of potential partners who can leverage the huge social network inside of Facebook. Facebook itself, rather than the broader Web, acts as the platform here.

Beacon is Facebook’s targeted social advertising system. From the press announcement and launch on November 6, 2007, 44 high-traffic web sites—from eBay to Fandango to Travelocity—were enthusiastic supporters of a new way to socially distribute information and connect businesses with relevant users and their trusted network of friends on Facebook.

For example, the chief marketing officer at Travelocity used the following example:

The Beacon announcement was taken as a major step toward justifying the $15 billion valuation by Microsoft’s investment and proof of a business model for “monetizing” the reported 50 million Facebook users.

It didn’t work out, though, because users objected, and loudly. How did a platform hyped as the “most important marketing breakthrough in the past 100 years” become a major controversy about privacy and, in the eyes of several observers, “a disaster” that could bring Facebook down? After all, Facebook’s News Feed feature, introduced in 2006, was arguably even a bigger “privacy trainwreck,” and yet it’s a feature that many users now enjoy and consider positive and useful.

The most succinct explanation I heard online came from Andy Carvin, National Public Radio’s senior product manager for online communities and former coordinator of the Digital Divide Network, an online community:

In other words, Charlene Li, a prominent Forrester Internet analyst, could purchase a cocktail table at Overstock.com and discover that all her “friends” had received a message via news feeds about her purchase. Or some number of the 55,000 protesters who signed the petition from MoveOn.org to demand that Facebook change the Beacon system had their holiday gift purchases broadcast to all of their friends, including the intended recipients. Coca-Cola, one of the early advertisers, stated that it had chosen to withdraw its participation in Beacon, stating that it had had the impression that Beacon would be an explicit opt-in program.

Facebook’s CEO Mark Zuckerberg apologized publicly on December 5 and made Beacon an “opt-in” system rather than a default. But trust was further eroded by independent tests that seemed to indicate that Facebook was continuing to collect information on click and purchase activities for data mining, whether or not its users opted-in. Legal experts and advocacy groups continue to look closely at the advertising program, and complaints have been filed with the Federal Trade Commission.

But the larger question had been raised in too many minds and blogs: does social advertising and marketing mean the end of privacy?

Social networks have always been a key part of sales and marketing. When you’re trying to sell a product, you want to find key influencers. The acceleration that social networks provide in a Web 2.0 market has two major effects in this regard:

LinkedIn provides clear paths for its customers to invite other customers to join its system. Members who invite their friends receive clear and direct benefits, and joining the network costs nothing (except time) for the invitee. This makes it very easy for people to give or get a positive impression of LinkedIn, accelerating its growth tremendously. Facebook’s viral marketing benefited from similar effects, amplified by its appeal to contained groups of people. Figure 3-14 shows how growth curves may vary by the nature of the kinds of people who spearhead the use of the system and their social influence.

Similarly, both LinkedIn and Facebook grew more and more rapidly as the value of their networks climbed. “Are my friends using it?” is a critical question for possible customers, and it’s vastly easier to convince them to join as the number of their friends involved increases.

Figure 3-14. Social influence drives growth

The value of a social network lies in its members. Social network software and sites are tools for connecting and finding the members, but in a strong sense, the health of the business isn’t the software, it’s the customers (see Figure 3-15).

Social networks don’t have to be like Flickr or Wikipedia, where users spend lots of time creating material to share with others. Basic social network software focuses on capturing what users already have—acquaintances—and creating a forum that can be shared and expanded. Directories are still powerful tools, and enhancing those directories to support new kinds of contact between users draws in more people on the basis of the same relatively small amount of information.

Figure 3-15. Emerging and useful patterns in social networks

Some social networks give their users considerably more freedom to add information. LinkedIn sticks to the basics; Facebook lets users add more information in a fairly structured environment; MySpace lets users do all kinds of things to present themselves. The kind of information users are likely to add to an explicitly social network application may be different from what they post to Flickr or Wikipedia, but it can still draw new customers, often even more motivated customers.

Social networking is a hallmark of Web 2.0. Users are willing to share a certain amount of themselves, and that sharing creates new opportunities. It’s a situation where users who like the site will probably want to invite their friends to join, creating an army of potential marketers and evangelists. It’s a situation where anyone looking for a targeted group of people may be interested in paying for access to all of those people who just showed up, making it possible to sustain growth.

Social networking creates an optimal arena for positive network effects to appear, and it also offers opportunities that can be integrated into projects that aren’t necessarily “social networking” projects. Any time a site has registered users—or even identifiable users—there may be an opportunity to offer them a place to call their own and share it with others. Building a community of people who use the same product and who can interact with each other, for instance, is a lot more likely to generate user interest than sites where customers interact only with a company. It’s less controllable, of course!