[beat]

This is relevant to my interests. As I develop my own enterprise SaaS product, my first-pass solution was to develop first as SaaS (to reduce my time to market), and later make an "enterprise" version that they could install themselves, on their own hardware, using a more traditional pricing model. I was dreading this because it sounds like a lot of work and support, but then my target market is businesses that have pretty severely broken internal processes, so they're probably backwards about pricing as well.

I brought this up with a friend who works in enterprise sales and he cautioned me against it. His company did the same thing - for a while. They found it so problematic to support the "enterprise" version that it wasn't worth it, and now they're a pure SaaS play. They find it easier to solve the problem on the sales side (convince customers to go SaaS, and skip the ones that won't), than to "give the customer what they want".

I suppose this is a case of choosing your customers.