This slideshow came from the recent Forrester Consumer Conference (see here and here)
The demographics summary:
- The numbers: 44 million active users
- Demographic drill down (34% work as professionals)
- Facebook users are aging
- How to apply it
I also found an indepth analysis of user behavior on Facebook and some specific thinking about metrics.
- Comparison of old/new metrics
- Nature of conversation on Facebook
- Segments by app usage
These figures are really fascinating.
Posted by: Clay Newton | October 13, 2007 at 02:25 PM
Hi Beth,
I personally think the numbers on Facebook, et al, are pretty bogus. Typically, when asked to fill in their age or date of birth during an online sign-up process, people simply lie. I'm not sure of the reason -- perhaps it's an almost unconscious thing; rage against the machine. As well, sometimes there is an incentive (like if there are enforced restrictions by age, etc.)
Regardless of why, people nevertheless do. I find that I do -- if only to frustrate the incessant attempts to pin down my particular marketing demographic. Moreover, when folks lie about their ages, they strangely tend to UP their age. I personally seem to gravitate towards 1916 as the year of my bogus birthday.
(You can see these patterns when matching survey responses against known demographic trends -- there are always more folks in their 80's + than there "should" be... )
Hence, all those 24 – 35 year olds that Forrester is counting are probably more like 14. Then again, a number of those 14 year olds are probably in their 40’s.
The same is true for responses abotu profession, etc.
The short story here: the numbers are bull, and nobody really knows who’s on first, and what’s on second.
Gavin
Posted by: Gavin | October 14, 2007 at 12:31 PM
Gavin,
Do you have any empirical data around your suppositions? Yes, people do lie on sites at times when they are entering their birth date. But to my knowledge, that is not an overwhelming majority ... if it was, advertisers wouldn't pay money for those numbers, which means that the only sites that would collect them would be the sites that have a regulatory reason for doing so. I think you are making a hasty generalization based on your own behaviour. http://en.wikipedia.org/wiki/Hasty_generalization
Posted by: Clay Newton | October 15, 2007 at 08:12 PM
Clay,
Ah.. I'm not sure I need a definition of "hasty generalization," but thanks all same.
As to empirical data, absolutely, there is lots of it. Fact is, there is so much I haven't a clue where to start. Surveys are notorious for generating false answers. It's called survey bias – it’s pretty much the subject of 9 billion dissertations. (I’m exaggerating, it’s only 2 billion.) [Nope, I lied, it’s only 200,000.]
Survey’s—asking people for information about themselves or their opinions –have always been problematic. One must be very careful to not write questions that assume answers (like most political pollsters do), or one must be very cognizant of the hidden incentives (or disincentives) to give false answers that are built in to how the very questions are structured. Survey designers go to extremes to cross-check questions; asking the same question in multiple ways to try to eliminate the bias.
[I’d imagine if these social sites asked for both age and date of birth they’d get rather interesting inconsistencies, for example.]
Survey bias can be a function of the questions, of what the respondent "thinks" you want them to say, or what the respondent "feels" they should say. Television ratings are one prime example: back when you actually had to log your watching behavior, people would log "educational" programs and watch sit-coms. Even though the Neilson surveys were “anonymous,” folks felt a social pressure, or even an “public obligation,” to list public television shows while secretly watching "Three's Company."
Admittedly, asking one’s birthday or age, at first blush, does not seem to have any hidden incentives or disincentives. But, in reality, depending on the site (and purpose), there are many. The studies done on online dating sites are an easy and obvious case-in-point. People lie about their age (apparently men more than women with dating sites. Humm.. that’s interesting given the social bias. I digress.)
Anyway, for a good time, I invite you to simply Google “Do people lie about their age online” – lots of fodder, lots of studies, and lots of examples.
If you're really really interested, I'd suggest the "International Journal of Public Opinion Research" published by Oxford Journals, available here: http://ijpor.oxfordjournals.org/
The bottom-line with survey data (and providing your age or birthday is a form of survey data collection), is that the data are only valid when seen and compared over some period of time. The adage in social research is: “people consistently lie, but they tend to lie consistently.” Hence, you can ascertain change in attitudes by looking at the same data collected over time. I suppose this might apply to age, but given other factors (like how you actually have to enter the numbers) I would imagine that there are other biases present, such as the easiest numbers to pick from the list, or some such.
With sites like Facebook, or other social sites, there are incentives to lie about age. Younger people want to avoid the restrictions, older people want to not be thought of as old, and people like me like to torment the marketers, still others simply rage against the machine. My favorite quote is here is from Jerry Pournelle (a curmudgeonly science fiction writer) who on a tech podcast proudly announced he was (or is) a 13 year old girl on Myspace.
People lie about their age, their sex, their weight, and their marital status. They lie about the books they read, their education, their military service, the salary they make, and the movies they see. If we were to believe what people tell us about their lives and loves, we’d have to believe that we all *really* watch “Masterpiece Theater” and not “Dancing with a Fifth Grader” or “Are You Smarter than a Movie Star” or whatever it is; and that we're all much better looking, healthier, and much more fun, than we really are. I know I am.
Finally, you ask, that if this is true, why do folks use such data? Well, it’s what they have. They’re well aware of the survey bias — and depending on their goals, they might not care. Political pollsters, for example, might want to make a point. Marketers – in this case the people selling Facebook –actually feel that an older demographic is better – they have more purchasing power, etc. Hence there is an incentive for them to stress data that may or may not be that accurate or useful.
There are three kinds of lies, or so says Mark Twain, Lies, damn lies, and statistics. These are clearly statistics, and as we all know, 47.8 percent of all statistics are made-up on the spot.
Gavin
(Age 12)
Posted by: Gavin Clabaugh | October 16, 2007 at 10:10 AM
Great discussion, here's my personal opinion (although I'm a new Forrester Analyst)
Didn't we see some reports that up to a third of all personal data on social networks are bogus? That's quite possible. Also, some of the numbers that I've seen come from Facebook themselves, it's their data.
While surveys may have their problems, there are no 'perfect' forms of data collection.
What should we do? The key here is not to paralyzed by the data we may or may not have, but to take in account multiple sources of information, look at trends and make decisions.
At the end of the day, this is an age old internet debate, and can even go further back to any type of media, and one that will continue to plague us for nearly all time (unless we give up personal privacy)
Posted by: Jeremiah Owyang | October 16, 2007 at 11:36 PM
Here's a humorous datapoint!
http://xrl.us/7mkg
Posted by: Clay Newton | October 23, 2007 at 08:40 AM
Re: DOB on social sites, I wonder why they can't ask, but user the ability to select "hide DOB". We all know that it is useful for the advertisers that actually fund the existence of such site to have accurate DOB, but at the same time, there are people out there who simply do not feel comfortable doing so. Does anyone know if MySpace or Facebook has a "hide DOB" functionality that allows it to be hidden on the front end, but is captured on the back end so that it is on their back-end database?
Posted by: Jill | May 22, 2008 at 11:25 AM