3 Ways to Use Social Network Analysis for Marketing
Social marketing is hot. What once was vaguely referred to as “word of mouth” is now a huge business. Facebook is worth nearly $100 billion and social listening and community assignments have become lucrative sources of ad agency fees.
Strangely enough, social network analysis has largely been ignored. In an industry obsessed with metrics, that seems like quite an oversight.
Since the late ‘90’s the science of social networks has been an intense field of study. It has changed the way we combat terrorism, study ecologies, fight disease and even evaluate organizations, but has yet to carry over to ad agencies and marketing organizations. Here are three areas where insights from the scientific community can be directly applied.
1. Early Warning Social Listening
Epidemics are difficult to control because by the time you know you have one, it’s usually too late. The infection has taken hold in a sizable proportion of the population and is spreading at geometric rates. The time for crises prevention has passed and you move straight into full-fledged crises management.
Wouldn’t be great if we could find a way get a head start? That’s a question that was tackled by Christakis and Fowler, leaders in the field of social network science. In a study of 744 Harvard undergraduates, they we able to isolate a network central group that predicted the outbreak of the H1N1 influenza virus on campus in 2009.
They found that they could get gain a full two weeks and lower the inoculation threshold from 90% to 30%. If we could adapt the same algorithmic approach to social listening, we could get a similar head start on outbreaks of sentiment. Some important PR crises could be averted and important opportunities could be uncovered and acted upon earlier.
Moreover, information overload is one of the biggest obstacles to effective social listening and focusing on network central individuals would help us cut down on the number of conversations we have to follow. We would have more signal and less noise.
2. Evaluating and Tracking Network Health
A few hundred people scattered in a park probably won’t communicate very much. Take those same people and put them at a backyard barbecue or a wedding reception and it will be hard to hear yourself think above the din. The number of people in a social network doesn’t matter nearly as much as how that network is structured.
Researchers at Northwestern University used a similar insight in their study of Broadway musicals. They found that the most determinant factor in the success of a play was not the marketing budget or even the track record of the director, but the relationships amongst the crew and cast.
They evaluated connectivity using a fairly standard metric known in the academic community called Small World Q.
Information transfer among consumers is just as important as it is between Broadway cast members. Once we learn how to evaluate “Q” for varying consumer and product segments we will be able to target viral campaigns to communities most likely to be able to carry them. It would also help us build more effective communities.
We currently have no way of evaluating community health. So, for instance, if we add a lot of followers we automatically assume that we’ve made our network stronger when, in actuality, we could very possibly have weakened it. Without a valid connectivity metric, we’re operating blindly.
3. Integrating Marketing Campaigns
Over the years, we’ve become extremely good at measuring specific media and marketing channels, but we are still in the dark with regard to how they work together. Pioneering network theorist Duncan Watts has worked on exactly that problem. The result is what he calls Big Seed Marketing.
Ironically, his thinking flies in the face of much of what you hear in the marketing industry these days. While many marketing gurus wax poetically about targeting “influentials,” Watts thinks that it’s more trouble than it’s worth. Rather, he suggests finding people who are susceptible to your message and reaching a lot of them.
If that seems a whole lot like traditional marketing, it is, but with a twist. He points out that new technologies give us a greater ability to create sharing mechanisms and that makes all the difference.
Further, his insights suggest that we can improve our passive model of paid, owned and earned media into a more actionable one of seeding, sharing and converting:
A point that Watts keeps coming back to is that the old two-step theory of communication where opinion leaders influence everybody else has been long discredited (although for some reason marketers still cling to it). In actuality, influence flows throughout a network and everybody influences everybody else.
Therefore, it makes a whole lot more sense to “seed” the network with a message targeted to people likely to receptive to it and then encourage and enable them to share it. There is very little evidence that social media can replace more traditional efforts, but a wealth of data that suggests that it can play an important role in amplifying it.
We already know a lot about seeding a network through cheap reach and quite a bit about how to convert consumers on owned resources. However, we know relatively little about how to get them to share messages. We need to undertake a concerted effort to learn more.
A Future Marketing Revolution
The new science of social networks is one of the most exciting areas of research today. At places like the Sante Fe Institute and others, it is allowing us to ask new questions and often find viable solutions to old problems.
Amazingly, despite the millions we invest in research every year, network insights have largely been absent in the field of marketing. Part of the reason for that is that the data sets we work with in marketing are often too large to make some forms of social network analysis feasible. Yet, that doesn’t explain the wholesale neglect that occurs today.
There is a wealth of data out there that we can put to use. After all, a successful marketing campaign is not unlike a flu epidemic nor is a consumer network so dissimilar to the cast of a Broadway play. The general knowledge exists and, in some cases, viable algorithmic approaches are there for the taking.
The real reason that more rigorous use of network science hasn’t been deployed by marketers is that most of it is still buried in hard-to-digest academic papers. Nevertheless, it is there, waiting for us to take notice and, when we do, the impact will be tremendous.