The Data from Your Social Network Connections SUCKS!

On the April 2nd episode of their terrific podcast Marketing Over Coffee (SUBSCRIBE! NOW! DO IT!) John Wall and Christopher S. Penn very kindly addressed a comment I had posted with regard to the Forrester Technographics Tool. As a quick summary, this tool uses demographic data from various countries to determine the statistical likelihood that a particular audience will partake in a particular social network activity. (For a much more detailed explanation, check out Forrester’s Groundswell site.)

In answering my comment, Chris got to speaking about demographics. Now, Chris Penn is an exceptionally smart guy, but I have take issue with this statement from the podcast:

…if you have a Myspace page, you have all the demographic data you could possibly ever want — age, ethnicity, religion — and yeah of course some people don’t put accurate information in, you have to accept that…

The idea that data from a Myspace page could be compared to something with the statistical rigor that organizations like Forrester use struck me as a bit crazy, but I decided to run my own little experiment with a comparable set of data, namely my own Facebook connections.

(For the record, I do have a couple of Myspace accounts that I could have used, but for the purpose of this experiment it turned out to be a lot easier to use Facebook for a few reasons. I believe the premise that Myspace and Facebook data are roughly comparable isn’t too much of a stretch to make my point.)

Using the Facebook “Sort Friends By…” application, I was able to sort my friends by their self-declared gender, age & home town. I then went through and manually identified the people who hadn’t done so themselves as being male or female as case might be. Here were the results:

A few interesting observations:

• Gender was the most reported characteristic, with 72% of both males and females providing their gender in the their gender in their Facebook profile
• Home town was next most reported, but females and males did not report equally. 59% of men were willing to give their town, but only half of women did the same.
• Age (not too surprisingly) was the least reported by both genders, but the gap was different again. 42% of males gave their age, but only 35% of females provided theirs.

So not only does the willingness to report specific pieces of demographic data vary depending on what is being asked, it also varies to different degree in the two different genders, making it difficult to control for reports vs. non-reports.

The upshot — Facebook demographic data, at least for my group of friends, doesn’t appear to give a very accurate picture of who that group is at all. I suspect this would extend to other peoples connections on other social networks, including Myspace.
Sorry Chris!

Popularity: 58% [?]