I think the most interesting part of this is that it really lowers the cost of providing potentially great applications, since the variable cost in adding a new client in a custom way is very cheap. This should allow more smaller niches to grow in ways that before the advent of this delivery model were unable to do. Furthermore, I think it plays nicely into the hands of the microisv who is already going after the niche markets.

It is also interesting that different architecture thought is being pushed out onto the table for discussion to support this delivery mechanism like Multi- Tenant Data Architecture. The development of this idea will help more than just SaaS architected solutions.

It's exciting to see and I highly recommend checking out the two articles linked to in this post -- it might change the way you think of your business model as well as your IT architecture.