Friday, March 12, 2010

Who's using social media most?

There are now more Facebook profiles than the entire population of the United States, Canada and the UK combined: it seems like pretty much everyone is busy blogging, tirelessly tweeting, or fiercely friending.

But who is most active with social media? Forrester Research has done a nice job of explaining the demographics of online participation, based on a taxonomy they created for the popular book Groundswell. Using their handy online tool you can see how online partication differs between men, women, age groups and geographies. What Forrester doesn't do is explain the types of people that get involved – their mental, emotional and social characteristics, as well as their motivations. What compels people to blather online? What traits do they have in common? Are the most active participants – what Forrester refers to as “Creators” - different from the lazy bums who do nothing with social media (what Forrester acidly calls the “Inactives”)? More to the point, what assumptions can marketing and communications professionals make about the people that actively engage with them through social media?

There are many competing and largely incompatible theories on this question, but lets look at two of the front runners:
  1. The Wisdom Of The Crowd thesis

    Advocates believe that the great mass of people that haunt social media and the Web are wise, well-intentioned, motivated and overall a pretty good facsimile of the population at large. They give us unfiltered access and insight into our customers, employees, voters, or other publics. True acolytes of this theory believe in an almost mystical ability of crowds to accurately guide opinion and decision making, an argument borrowed from the work of James Surowiecki who wrote the popular Crowdsourcing book back in 2004. Ask a crowd of kids to guess the number of jelly beans in a large jar, and the while individual answers will vary wildly, the average of all answers will be extremely close to the true number (I know, I tried this at a school fair).

  2. The Madness Of The Masses thesis

    There's a whole stack of theories penned by lofty thinkers like Freud, Tarde and Bernays that see crowds as irrational herds, spawning Orwellian DoubleThink or manipulated GroupThink or all sorts of other Thinks. But you don't need to wade through all these turgid academics tomes, just read the comments on an average blog. A great deal of the blogoshere and social media seems to attract rabid extremists and flat-out nutters. It just sounds mad.
Both theories can't be right. The Wisdom thesis seems to argue that the most active social media participants are reasoned and can fairly represent underlying populations, which is why crowdsourcing advocates argue they can be co-opted by marketing types to help develop and test products, ideas, messaging and so on. On the other hand, the Madness thesis argues that social media activists are just that – at best unrepresentative, at worst a herd led by extremists, zealots, and the marginal. If the Madness thesis is right, marketing types would interact with the social media Creators in a very different way. So, which thesis is correct?

Lets start by explaining crowdsourcing, which will require we look at some statistics and probability theory. Imagine you are faced with a question for which there are only two possible answers – say, who will win in an election, Democrats or Republicans. There are numerous factors that might determine the result, and any given individual will only have a partial understanding of those factors. Lets suppose that given all the factors people comprehend, that each individual has a 51% chance of being right in selecting the election winner. On this basis, if you were to ask one person for their pick, there's an about even chance they'd get it right – you might just as well flip a coin. However, ask 10,000 people for the answer, and the majority selection will be extremely accurate. The same statistical logic applies when you have complex problems with many possible answers, like the jelly beans in a jar problem, and you apply it to a group of smart 5th graders who can make a well-reasoned guess. Ask enough 5th graders, and the average answer you get will be weirdly close to the real jelly bean count. This is called the Condorcet Theorem, but here's the rub: if the individual group members are less than 50% likely to be correct in their selection, then as you increase the group size the probability that the answer will be correct starts to approach zero.

Condorcet has important implications for crowdsourcing. First, it explains the otherwise mystical ability of some crowds to make very good average guesses. Second, it suggests that we should never rely on the wisdom of a crowd if we know or suspect they are poorly informed or biased. In fact, these people will lead us away from the correct answer. At the very beginning of his book, Surowiecki gives a classic example of a crowd that is well informed and unbiased – attendees of a county fair guessing the weight of a cow. Can we say the same about Forrester's social media Creators?

I think there is some evidence to suggest that social media Creators are likely a self-selecting group of activists, passionate evangelists or committed detractors. The comments I get on my corporate blog or in response to an informal online survey are not going to be representative of the underlying population of all customers. Borrowing from Condorcet (and breaking with strict statistical logic), the opinions I see online are not going to approach some kind of truth – they are just as likely to head in the opposite direction. And the more people I interact with through social media, the more they may lead me astray.

Or at least that's the risk. Truth is, we don't know enough about social media mavens and we could do with some good research as a complement Forrester's Technographics.

We in marketing should be careful to understand who is active in social media – what is driving them, what motivates them, and to what extent they are outliers and unrepresentative of our target audiences.


6 comments:

Frank Days said...

I totally agree with your last point about participation bias. In some social channels (let's say Twitter) there are are real biases based on age, gender, occupation and income that can make it an unreliable predictor of the real world.

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...

Nice fill someone in on and this post helped me alot in my college assignement. Say thank you you as your information.

Anonymous said...

i seriously enjoy your own writing choice, very helpful,
don't give up and also keep posting due to the fact that it simply good worth to read it.
looking forward to see more and more of your current posts, thanks!

Anonymous said...

This is an interesting article. Thanks for sharing.