• by David Honig , February 2, 2009

Spats have more in common with Twitter than you thought

Something extraordinary is happening: we are witnessing the creation of an entirely new Internet model that few people could have foreseen even three years ago. According to some analysts, as much as 70 percent of consumer time online is now spent viewing content created not by professional editors, but by fellow consumers. Most eyeballs are trained on social media sites like Facebook, MySpace, Flickr, and YouTube.

Social scientists, among others, are fascinated by these developments. Many have noted that in social media we see a classic example of homophily - the powerful tendency of like-minded people to gather in closely-knit clusters around common interests, including brand affiliations.

Since the dawn of Geocities, figuring out how to monetize the social space has been a thorn in marketers' sides. The industry has not been able to find a way to fund social media through ad revenue alone. In fact in 2008, when evaluating their partnership with MySpace, Google executives recognized that "social networking inventory is not monetizing as well as expected." The industry's monetization problem may have occurred because marketers tried to apply models from other online contexts to the social space. For example, cpm and cpc models have worked with traditional mass media display and search, respectively, but have had limited effectiveness in social media. It is time to look at an old problem in a new way.

Incorporating social graph information into online advertising targeting methodologies will instantly add value to any behavioral, demographic, or contextual advertising plan. Most marketers would intuitively agree that the social groups of their existing customers (aka "network neighbors") represent a desirable audience. Here's why:

As summarized by the old proverb dating back to Plato's The Republic, "birds of a feather flock together," people tend to form social groups of similar, like-minded individuals both on- and off-line. This tendency to interact more and form stronger social ties with those similar to us is a classic sociological and psychological concept, but one that's very relevant.

Members of these groups tend to influence each other through word-of-mouth communication, and they tend to engage in collective purchasing behaviors.

Marketers who can successfully leverage the power of homophily - while respecting users' privacy - will find that they've harnessed an analytic tool that has the potential to be highly predictive on a massive scale.
A simple definition of homophily is that "contact between similar people occurs at a higher rate than among dissimilar people" and that "people's personal networks are homogenous with regard to many sociodemographic, behavioral, and intrapersonal characteristics" as defined by Miller McPherson, Lynn S. Lovin and James M. Cook in their essay "Birds of a Feather" about social networks. Similarity in factors such as "ethnicity ... age, religion, education, occupation, and gender" operating in shared geographic, family, and organizational environments play a huge role in determining how personal relationships are formed.

Why's that matter? Homophily is a robust phenomenon because it forms the structure in which almost every type of social relationship exists. You will find homophily in "marriage, friendship, work, advice, support, information transfer, exchange, comembership, and other types of relationship[s]," say McPherson, Lovin and Cook.

Whom does homophily affect? Everyone. Everywhere. It is global cultural phenomenon: Researchers in the early 1990s were able to find approximately the same levels of homophily in a Chinese city as in the United States. It is pretty amazing that two completely different cultures can have the same level of homophily. In our increasingly fragmented social space, it is essential for marketers who want to target specific groups to keep this in mind.

And homophily is highly predictive. In essence, the power of people's self-selection in personal relationships creates social groups that rival the best segments created by demographic, behavioral or psychographic targeting. For example, researchers have shown that marketing to people socially connected to an existing customer of a product has significantly higher response rates than methodologies uniformed by this social information, including demographic and behavioral targeting.

Homophily is a concept originally observed and formulated in the 1920s, but it has become increasingly relevant in the Internet and mobile age. New forms of technology allow people to interact with each other and form social ties and groups (through social networks, blogs, Internet chat, Web video, mobile phones, texting, etc.) much faster and more effectively than ever before.

In the 1920s, your social groups would have been comprised primarily of similar friends from your family, neighborhood, school, work and social clubs. For example, if you lived in Tampa, cultivating numerous friendships with people in San Francisco, Chicago and Dallas would not have been easy for the average American.

The advent of Internet and mobile technologies makes it extremely easy to interact and connect with people all over the world. Due to the popularity and high penetration of next-generation communications channels, geographical barriers have been drastically reduced, if not completely eliminated for many Americans. We can expect Internet users to use these new communications media to continue to seek out and bond with similar people and form even stronger relationships with their "real-world friends" online.

The advent of social media necessitates a radical change in the online marketing paradigm. The next generation of targeting strategies that incorporate a privacy-centric, group-oriented methodology will provide better audiences for brand advertisers. Moving forward, rather than asking your customers "Where do you live?" Ask: "Who do you know?"