As We Now Think

In books like Here Comes Everybody and What Would Google Do?, social media are presumed to empower and engage all consumers equally. Because these new technologies lower (or virtually remove) the cost of creating and sharing content, it has been assumed that anyone and everyone is involved in the process.

Clay Shirky describes an episode wherein irate travellers, stuck on a plane for an excruciatingly long delay, organized a mass protest over the internet. Soon, the world knew of their ordeal, and the airline looked horrible as a result.

Jeff Jarvis tells the story about PC manufacturer Dell ignoring customers and losing market share as disappointed customers began sharing horror stories of faulty laptops via blogs and comments. Eventually, the company realized what was happening, and began to engage disgruntled ex- and current customers, going so far as to have top executives write their own blogs and respond directly to questions and complaints.

The observation that both authors (and many others) make is that social media will force institutions to provide better service, because the information infrastructure acts as a mechanism for constant feedback and monitoring. While I do not disagree with this idea, the reality of the situation has some nuance that is only now becoming apparent.

What the authors are missing is that the information gathered and shared by social media–the sum total of the communication–will always suffer from a bias similar to that of self-selection in social science.

Very little research is conducted wherein the pool of respondents themselves choose to take part in a study.  As any researcher knows, self-selected groups are not random samples of a larger society, and findings from such groups are not generalizable to larger groups.

A good example of why this is problematic would be the non-scientific polls that TV stations run all the time. A news anchor might introduce ‘tonight’s question’, and invite viewers to respond online or via text message. At the end of the show, the station will reveal the results.

These results are absolutely worthless for anyone trying to learn anything about the world. If FOX News and msnbc were to pose the exact same question they would find wildly divergent results. That much is obvious, but perhaps less obvious is the fact that neither could validly state that their results were representative of even their own viewership. This is because even among their biased pool of viewers, only a select (non-random) sample will be motivated to make the extra effort to respond to the poll.

What does this all have to do with social media? While Facebook, Twitter, et al. may not be attempting to undertake valid academic studies, the same issues come into play when observers try to claim that they will create anything approaching consumer democracy or force companies to behave better.

From the perspective of institutions, what is important is to isolate those groups and individuals most active in content creation and sharing and catering to their tastes. Of course, this is a very, very complicated task. There are two main complications in doing so: (1) the inherent difficulty of predicting (or tracking) who most likely to be active and influential in this sphere, and (2) the fact that different people will be motivated to create and share under different circumstances.

The former is difficult, although not impossible, to overcome. The best place to start looking for the answers would be the academic work of Eszter Hargittai and Richard Florida. Hargittai, through surveys conducted at the University of Illinois-Chicago, has demonstrated that while most college students are consumers of social media, only a non-random sample are producers. Demographically, we can tell who these people are: the better-educated, privileged members of society. Of course, we need to understand that only a small sample of this group actually produces content, so we can look to Florida’s work on the ‘creative class’ to figure out what motivates certain people to create and share content and ideas more than others.

The point is, we can figure out who is more likely to create and share content online.

I opened this post with references from Shirky and Jarvis to illustrate the latter point. People who would normally be less likely to be active in discussing companies or experiences will be far more likely to do so when those experiences fall to an extreme end of the spectrum. If you want to see an example, check out reviews on Yelp, Amazon, or any similar site. On five-point scales, you will see an exorbitant amount of fives and ones, and relatively few threes or fours.

In part, this may be due to a tendency to exaggerate one’s opinion on the internet. But there is also certainly the dynamic that people just are not interested in taking the effort to share their opinions of relatively adequate of average experiences. And it’s hard to imagine that they ever will be.

Really, we should recognize that the story we’re seeing unfold isn’t exactly ‘Here Comes Everybody’, but rather ‘Here Come Some People All the Time, and Some Others Some of the Time, Based on the Quality of their Individual Experiences’. It isn’t catchy, and it isn’t simple. But neither is the world. And the lesson for companies and institutions isn’t ‘if you don’t provide good service, people will talk’, but ‘if you provide terrible service, people will talk. On the other hand, if you provide adequate service, you don’t really have to worry’.