Every spring and fall, skeins of geese fly over our house as they migrate. I always wonder how the lead bird gets to be the lead bird. Is it the one with the best sense of direction, the one that knows best how to use the prevailing winds, or simply the dominant male? It turns out that Twitter followers flock around specific people in the same way that birds follow their leaders. So what makes a lead Tweeter?
I have little doubt that by now most people in the blogosphere know of the research I’m referring to. On June 1, Bill Heil and Mikolaj Piskorski, graduate students at the Harvard Business School, published their analysis of the activity of 300,000 random Twitter users during the month of May, 2009. (Click here for their post on a Harvard Business blog.)
Some of their findings are startling, although, I am glad to say, they seem entirely consistent with my musings on the topic a couple of weeks ago. Here are the two key takeaways from the research as I see them:
First, the top 10 percent of prolific Twitter users account for over 90 percent of tweets. Not only does this prove that Pareto’s Law (aka the 80/20 rule) is alive and well on Twitter, it points to the fact that the vast majority of people are followers, not Tweeters. In fact, the median number of tweets per person is one. Count ‘em—one. As I suggested in my previous post, much of what takes place on Twitter is just good old-fashioned celebrity worship.
Second, men have 15 percent more followers than women. A man is almost twice as likely to follow another man as he is to follow a woman. A woman is 25 percent more likely to follow a man.
I don’t buy Bill and Mikolaj’s suggestion in a comment following their post that these differences are the result of what the two sexes write about. I think it is because of the difference in perceived status. It may be that the glass ceiling is, sadly, alive and well.
Commenting on the Harvard post, “Michelle” says the findings are interesting but suggests “It sounds like a fairly straightforward example of gender stereotypes.” I am not sure it is fair to suggest this behavior is the direct result of stereotyping, although I am certain it is an indirect result. People, male and female, want to associate with others who have power or expertise or who are outstanding in some way. Our society tilts the playing field in favor of men making it to the top in their chosen professions (as does biology, since men don’t get pregnant). Ergo, more men get followed on Twitter.
I have no idea whether geese pursue an equal opportunity policy when it comes to selecting a flight leader. Unfortunately, if they are like Tweeters, the lead goose is probably the dominant male after all. So what do you think of my theory? Equal opportunity posting, please.
(3 votes, average: 3 out of 5)
Loading ...
Email This Post
June 15th, 2009 at 3:54 am
Leader? I thought that geese (much like cyclists) take turns at leading, then drop back as they get tired. It’s all a matter of efficiency in ‘taking the wind’. Let’s face it, in a migratory species, an individual that lacks a sense of direction is hardly likely to contribute to the future gene pool. Is this relevant to your argument? Well I guess to answer that we need a longitudinal study. If it has any bearing at all, we’ll see that over time, there is a lot of churn at the top of the league tables, as good old fashioned celebrity worship finds new icons to follow.
June 15th, 2009 at 4:37 am
The flocking of birds is an interesting parallel to use because - as I understand it - it’s a classic example of how our observations suggest an order that (from the birds’ POV) isn’t there. We see birds flying in formation (or in a cloud) and assume it must be pre-planned somehow. In fact what’s happening is that each individual bird is aware of the movements of one or two other individual birds, and group behaviour emerges from that.
Twitter, it seems to me, works in a similar way. We see patterns emerge at a macro level but the extent to which they affect the actual user experience is very much open to question. From the perspective of an individual user, does the 90/10 thing matter? I’d argue not.
June 15th, 2009 at 6:14 am
At the risk of having ornithologists swooping down to correct me, I think geese rotate leadership. I believe another takes over when the lead bird tires. And the honking we hear is encouragement (Go leader!). So if I am correct, then birds have cracked the glass ceiling and humans follow.
June 15th, 2009 at 7:54 am
Well, I guess this proves I should be more careful in choosing my analogies in future! However, I am not sure that it matters. The data still suggest that people tend to follow far more than they lead. Does that not have important implications for how Twitter is used and whether it has a real marketing role?
June 15th, 2009 at 8:22 am
I’m not sure the data does suggest that. The median tweet of one seems to me simply to back up the Nielsen study from a month or two back saying Twitter has a high churn rate: most accounts turn up, post once (or not at all), and leave. They’re not leading OR following, they’re just inactive.
Are the 90% simply following the 10%? If that was the case, what a great finding for marketers - it makes finding your influentials drastically easier! But I suspect the 10% are mostly talking to other 10%-ers and the 90% aren’t doing much of anything.
And as I say, none of this makes any difference to the individual user experience. Twitter isn’t a community so there’s no reason participation inequality should make a difference to the UX: everyone on the service bar the 1000 in my immediate network is mostly invisible to me. Whether they’re tweeting regularly or not is of no consequence.
That said, this research might help brands on Twitter to focus on some useful questions:
- How many of my followers are active users?
- How many of my followers are interacting with me on any kind of regular basis?
These are questions they should have been asking anyway, of course
June 15th, 2009 at 8:46 am
Thing is, the methodology for that study is opaque to say the least; they seem to have left in all the inactive users, for example. There’s no explanation for how they arrived at the neat quadrants for male/female following. It might be true, for many of the reasons you state, but I just want to see how they got there. The general tenor of the HBR review also makes me think they don’t understand Twitter terribly well. A good study of Twitter might at least sort out the Twitter mega-celebs from the neighbourhood networkers.
June 15th, 2009 at 8:57 am
The number of people who will start a conversation is way smaller than the number who will join it. If it ends up failing on the weight of making money, I think Twitter will be remembered as the thing that made immediacy a relevant notion for people and brands. Twitter begat a change in Facebook that has made engaging with content simple and immediate.
Local brands would be wise to grab onto this immediacy angle since it’s something big brands will never be able to do well. Immediacy as a sort of intimate relationship doesn’t scale. A twitter-like communication channel offering relevant updates using local language cues is the local brand’s new advantage.
June 15th, 2009 at 12:59 pm
Good point Alison, the outliers may really hide some interesting patterns.
Matt, another good point about Twitter and local brands. A while back I was looking for examples of brands which had used Twitter successfully. The ones that did so tended to be location based, especially the mobile food vendors in big cities who used Twitter to update their location.
June 17th, 2009 at 7:29 pm
HBR is not a refereed journal so anything goes. (Ever see a fact in Quelch’s writings?) I have seen incredible crap published there–no facts, bad assumptions, blah blah. yesterday the reach of tweets about my blogging for http://www.fastcompany.com was about 50,000. I have no idea how my ideas could have spread that far in the media and marketing community without Twitter. BTW, I think Twitter is actually having an effect on what is going on in Iran. Twitter’s impact is simply amazing.
June 17th, 2009 at 8:07 pm
Hi Joel, thanks for weighing in.
I am sure your ideas have merit and the Twitter format forces simplicity which makes them easily digestable and shareable. They would have eventually reached 50,000 people in a non-Twitter world, just nowhere near as fast (and you would never have known about it anyway).