Where have all the minutes gone?

There's been a lot of coverage lately of the real state and size of Skype's user base, but assuming we won't ever get any audited figures until the IPO prospectus is filed [;-)], I think it might be more instructive to look at usage patterns. Sad though it may sound, I happen to have been tracking Skype minutes of use for nearly a year now (a task helped dramatically by the statistics RSS feeds created by Skype), taking observations somewhat sporadically at 8:00 AM GMT (or as near as I can to it), and typing them into a spreadsheet. My methodology (perhaps too lofty a word) has been to take the number of minutes recorded since the previous observation, and divide by the number of intervening days. Since the start of the year, here's what I have observed, for what it's worth (I don't claim that these are precise figures, just my observations of what Skype has reported, subject to caveats regarding unequal distribution of datapoints, etc.):

Average minutes of use per day

Jan - 28,954,133
Feb - 37,533,906
Mar - 41,745,885
Apr - 41,732,959
May - 39,451,552
Jun - 38,479,729
Jul - 35,754,556
Week of 12 - 19 Aug - 36,601,232

It's interesting to note that downloads of Skype are up 61% since the end of March, and the active user base reported by the company has nearly doubled, yet looking at the concurrent user stats, we seem to be stuck at a high point of just over 3m on any given day, and average minutes of use in the most recent week (again based on my calculations) appear to be down a little more than 12% from March levels. I'm at a loss to explain why such strong growth in the user base should not translate into heavier usage figures, unless the IM functionality of Skype is acting to dilute voice. One Platinum Circle reader pinged me yesterday with a timely observation:

"My kids spend hours a day using Skype IM, but never use the voice features."