toxicTom: On a more serious note I was thinking what would be really be possible to assume from information even coming from "minimal" profiles. It shouldn't be hard to write some crawler that visits all the profile pages (going through the forum, friend lists etc... I wouldn't be surprised if somebody already has that...
Now you can build a database of usernames and (unreliable) location (from the forum), and join date.
You also know the number of owned games in any case, achievements and play time for people who have used Galaxy before.
The interesting thing now, is not that one-time snapshot. You could watch all the known accounts over a longer period of time.
For instance you could assume the taste in games from when the number of games goes up by one after certain releases.
You could assume a "scrimper" if someone almost never buys outside large sales, but then usually a number of games.
For users having time tracking still on, who can monitor the increase of hours played. During the week, or only on weekends? If you know the location even what time of day. This allows to make assumptions about work status, family or single...
This is all very interesting data for marketing purposes. Of course individual cases will be wrong. Someone playing all day might be working night shift...
Since many people use same or similar usernames and avatars across several sites, you can cross-reference, and in some cases even pinpoint the individual. But even if not, find the same handle on a motor forum or a football site tells you more...