A couple of weeks ago Ofcom published its annual Communications Market Report which nicely distils lots of UK media research into one place (albeit a 334 page “place”).
The report noted that social networking is “growing more slowly than previously.” This isn’t perhaps all that surprising since at some point services like Facebook run out of new sign-ups.
But the piece of research that really caused some ructions was that summarised as follows:
Social networking is also maturing – literally. Use grew fastest among 35-54s – up by eight percentage points since Q1 2008 to 35%. Among 25-34 year olds use grew by six percentage points to 46% but it actually fell slightly among 15-24s – by five percentage points to 50%.
This couldn’t be true could it?
Fewer 15-24s are using social networking than previously?
Then there was that widely quoted “research note” written by a 15-year old Morgan Stanley intern (Seemingly, nobody at Morgan Stanley had previously bothered speaking to their teenage sons and daughter to find out how they were using social media). That report said that 15 year olds didn’t use Twitter. That might be true for that particular teenager, and it’s certainly true that Twitter appeals to older people, but taking one person’s experience at face value is always dangerous.
Well let’s have a look at some of that Ofcom report in a little more detail.

First of all, while the Ofcom report does indeed show a drop between Q1 2008 and Q1 2009, the 15-24 age group has the highest penetration of any age group.
Over the same period, the 25-34 age group has leapt from 40% to 46%, while amongst 35-44s the jump is even more marked going from 28% to 35%.
But the question remains: are the underlying figures accurate?
The source of the data is Ofcom’s Technology Tracker. Unfortunately, full results from this research don’t seem to be available, and although the source gives the sample size at around 6000 per sweep for Q1 2008 and Q1 2009 (it’s significantly lower at a mid-point sample taken in Q3 2008), we don’t know how many are 15-24s.
Supposing that the sample is split relatively evenly in line with population, there should be no doubt that the figures are accurate and the margin of error should be minimal.
The jumps around and changes in levels suggest something a little broader – perhaps people’s defninitions of a “social networking site” are changing, as well as their behaviour.
Interestingly, Comscore came out and pretty much refuted Ofcom’s findings in a press release that highlighted continued growth in the sector between June 2009 and June 2009 amongst 15-24 year olds. Facebook again shows excellent growth.
Now ComScore certainly employs a very different methodology to derive its figures to those of Ofcom. Some have poo-pooed Ofcom’s numbers because they’re based on information from as long ago as Q1 2009. That’s an eon in social networking terms! But I’m not inclined to agree. Yes more up to date data is always preferable, but this is only a few months old, and it’s not enough to have made a difference, except perhaps, to Twitter.
On a broader level, it seems clear that different age groups use social media very differently. Conventional wisdom doesn’t regard instant messaging as a “social media site” and very few 15-24s would probably regard it as so. Yet it’s vastly popular in this age group (TGI suggests that 45% of all instant messager users are in this demo, and I suspect that account for the vast majority of the messaging), and it’s certainly “social.”
The younger you are, the more your “social network” revolves around people you see very regularly: at school, college or socially. But even then, you still use these sites a lot.
Then you get older and lose touch with some people – you might find them again via Facebook, or stay in touch professionally via LinkedIn.
Much of the debate seems to be focused around how Twitter is slightly older (not that old if you’re a 62 year-old, former reality TV contestant and newspaper columnist). That’s because it’s something more than just that status update bit of Facebook. Yes – you stay in touch with friends via it, but you probably also have a wider selection of people you follow who are feeding you stuff that interests you.
It can take some time to learn this, which is probably why so many Tweets are babble. Of course one person’s “babble” is another person’s interesting bit of information.
And once you understand how Twitter works best, you don’t just subscribe to anybody’s Twitter stream. I’m only likely to subscribe to people I’m interested in, and those who have interesting things to say (in my eyes at least). So the signal to noise ratio is much better for me.
Indeed one clear element in the “babble” research by Pear Analytics is missing: we don’t know how many people followed the tweets that were sent. The research considered 2000 tweets pulled from the public timeline. How many followers those tweets had was not considered.
Somebody who just tells the world at large what they had for breakfast probably isn’t going to get a large number of followers outside their close circle of friends and family. Indeed, if you only participate to that level, you’re likely to find Twitter unappetising and become part of that other oft-quoted stat – the number of people who leave nearly as quickly as they join Twitter (or plenty of other social networking sites).
The tweets of more interesting opinion formers are seen by vastly more people. And they probably value those messages a great deal more. Put simply, one tweet by Joe Bloggs is not equal to one tweet by Stephen Fry. Not all tweets are alike*.
You can see the whole paper here, although as I say, I believe it’s flawed in its conception.
I’m not sure what my conclusion overall here is, except that you should always be a little distrustful of statistics unless you’re able to look at the full picture.
* In fact, Pear Analytics did understand this, and referred to a Gizmodo blog post and accompanying visualisation that noted that 75% of tweets come from 5% of the Twitter community. But they just didn’t actually make use of this information in their own research.