This is the sixth in a series of excerpts from the report A Matter of Perspective: A systems approach to communication and complexity. A copy of the entire report is available here
Much of what drives successful communication lies outside of our controlSean R. Nicholson
For as long as anyone can probably remember the statistical standard for defining societies has been demographics. These generic attributes, ranging from age, gender and race to education, employment and even home ownership, have been used to characterize what social scientists refer to as “representative agents:” factitious persons or groups who typify the behavior of broad swaths of the population. For marketers they may be women 18 to 49; for economists, the one percent; and for politicians, “the American people.”
The problem is, as populations atomize into ever smaller, self-defined segments whose needs and concerns overlap, demography is losing its appeal. In its place some practitioners have turned to psychographics to interpret consumers’ beliefs, personalities and lifestyles. Others though are captivated by the aura of affinity groups that coalesce around shared interests or objectives; and for which they have anointed a new version of the representative agent – the influencer.
In brief, the lore of the influencer goes something like this: certain individuals who are especially authoritative or passionate about a subject garner substantial numbers of friends, followers or connections who value their opinions. Organizations then promote and market themselves through these virtual persuaders to induce desired behaviors on the part of target audiences. Not surprisingly, however, the reality is more complex.
Among the first to challenge this conventional wisdom was Columbia University psychology professor Duncan Watts, who is also a principal research scientist at Yahoo. As early as 2001, he began questioning the notion that one person or a small group can drive collective behaviors online. Subsequent findings by social platforms Buzzfeed and StumbleUpon assert that when influential people do reach a wide audience their impact is short-lived. Moreover, in every instance the analyses concluded that content is more likely to spread when large numbers of ordinary people share it with small groups of other ordinary people, instead of when it comes from someone “special.”
But Watts has taken it a step further. He maintains that regardless who the sender is, the flood of ideas will only flow if the receivers comprise a critical mass of easily influenced people, who then pass the information on to other easy-to-influence people. Without them, he notes in his book Everything is Obvious, “not even the most influential individual could trigger any more than a small cascade.” By that reckoning the task before communicators is to identify who is truly impressionable.
There is certainly no shortage of material to sort through. To the contrary, website traffic, online searches, banner advertising, social media and smartphone use leave behind vast trails of personal information. It is estimated, for example, that 34,000 tweets are sent every minute. That comes to about one billion tweets per month; still far less than the 30 billion pieces of content posted on Facebook. Add to this the endless data churned out by the so-called “Internet of Things” – millions of objects embedded with readable sensors – and it has engendered a new lexicon to calculate it all with terms like gigabyte, petabyte, zettabyte, and most recently yottabyte, which is designated by the number 1 followed by 24 zeroes.
But data is just data without the right tools to analyze it; ergo the current allure of big data analytics. By examining massive amounts of digital information simultaneously across hundreds or thousands of parallel servers, organizations can discover once-shrouded paradigms among consumer behaviors; then try to respond to, and predict, outcomes in real time. As part of this process, companies are also attempting to translate much of this data in ways that will allow them to appreciate consumers in more subjective terms. Beyond simply counting likes, follows or retweets, they hope to uncover and exploit genuine attitudes, emotions and intent.
This sentiment analysis represents a major step forward for social metrics. Nonetheless, it has its limits. For one thing, it has a hard time handling sarcasm or cynicism. For another, persons of various ages, ethnicities, genders and geographies can use the same words differently, which further flusters machines unable to pick up on nuances. Most importantly, people filter their judgments and beliefs through a host of perceptions and cognitive biases that computers alone cannot infiltrate.
Enter new methods like neuroscience which melds the study of the brain with fields as varied as computer science, engineering, math, chemistry, physics, psychology and philosophy. But the impact of neuroscience in terms of inferring consumer attitudes and behaviors remains open to debate. In the meantime, the ability to understand how humans process information is still mainly the domain of other humans.
Cartoon: Sean R. Nicholson