Looking at my iTunes library, the playing times of the first five songs of Help! are:
Now, let's look at the first five songs of Coldplay's X&Y:
So if an equal number of people decided to listen to Coldplay as those that decided to listen to The Beatles (I'm not claiming this is true), the latter would come out on top because more songs would be played during a given period of time. Thus, Last.FM would show The Beatles as being more popular.
There's obviously a lot more going on here. Not everyone is like me and listens to full albums while they work. It would also be interesting to see what the psychological "repeat-value" of the Beatles is. My hypothesis is: because Beatles ditties are so short, they'll be played more often anyway.
I would also hypothesize that if your band wants better Web 2.0 popularity, make shorter songs.
Last.FM should also add a feature to rank artists by total time listened to. That information is obviously collected since it's the basis of scrobbling.