Seems like a helluva lot of extra users to convert. 1 million by June 2004--19 months from Live's launch (US at least, a bit less in Europe), but 9 million more in the next 30 months? At their current rate of sales they would indeed be at ~25 million at the 5 year mark (~16 million installed base) so that would mean increasing Live's adoption from 6.25% to 40% in that extra time. (Will 40% of people even have BROADBAND by that point? Last I heard we were only in the high teens for broadband adopters.)
I'm not sure about their current numbers, but their main page still flashes "more than 500k" (which they announced hitting back in April, as I remember), so unless they're holding back for some reason I guess they have somewhere between 500-600k active subscribers right now, or 5-6% adoption. That's some pretty low moving in the next year to bust things wide open afterward. We also right now, of course, have a bit of confusion since we know two months of Live come free with every Xbox sold right now, so there's likely lots of people coming and going. It would be most interesting to see an actual adoption rate of people who tried and continue to subscribe, and those who tried and left (and what kinds of time periods they were on for.)
Certainly as more and more games filter in the popularity will grow, and I have no doubt Halo 2 will get plenty of interested parties on board (and back on board), but I don't think the end of the run will anywhere near where they were hoping. (Especially since they plan to get Xbox2 out before that point, which will almost assuredly confuse the sales of the current gen.) ALSO especially since they planned on 10 million at $10, and since their current monthly subscription is $5, matching the dollar amount would require them to have 20 million active subscribers, or raising their price. (Which they might be able to do once enough people are "hooked" but might also cost them both active subscribers and potential joiners in the process.)
Also keep in mind that $2 billion over the five year mark--even going by their HOPED target number of 10 million--amounts to $200 per person. A year of Live now costs $40 ($60 if you perpetually pay month-to-month)--minus a lot of profit for the first year from the kit--and only started off at $70 (and the rates went down to $50 or so before the first renewal date I believe). The five year mark lost a bit from the time of press announcements to its launch, but even assuming they had their hoped-for numbers AT that mark paying FOR the five full years, the numbers would barely break out of the red.
The numbers are simplistic, but even going by their own projections and what I must call "very optimal targets," the service doesn't stand a chance of making a profit unless the situation drastically changes.
