I had a good conversation about this with Dave last week ... maybe he'll chime in! But I'm now convinced it all depends on where the early adopters are in their licensed-software depreciation cycles. So, if someone could tell us when the majority of the installed market will have fully depreciated their up-front investments, I think the huge attraction of monthly fees without up-front license fees (since they pay large maintenance fees with installed software anyway) will win a lot of converts as they rip-and-replace old systems with new, better, cheaper alternatives.

Having been involved in a pioneering SaaS business in the late '90's, TPN Register (it was called outsourcing then), and running a software company now (with a SaaS-type offering of our solutions), I look at SaaS from this perspective. Enterprise Fortune 500 -type customer's loved the "plug and play" deployment, having another organization host and administer the solution, and the avoidance of high, up front license fees. What they didn't love were the limitations with regard to customizion, potential for data security issues, and the fact that if the provider shut down operations, they had minimum recourse (unlike licensing, where they can get escrow accounts and other protection).

So for SaaS to go mainstream for large companies, we need to address the things customer's don't love listed above (I'm sure I missed a few)...and I think we're getting there, thanks to advances in technology and just changes in attitude.

Re: the mid-market, SaaS will be a comet once somebody makes the pricing work, which is difficult due to, among other things, the costs of supplier enablement (catalogs, EDI, etc.) and potential erosion of price points at the enterprise software licensing level (e.g. Ariba Buyer is Ariba Buyer whether it's for a $20B company or a $20M company, other than the price charged)!