A lot of vendors coming out of the conventional software world regard SaaS as merely a delivery option. I spotted this comment by Kana CEO Michael Fields the other day, in a story about the CRM vendor's plans to offer a hosted version of its conventional on-premises software:
"It's about offering flexibility of choice and customers being able to acquire software on the business model that works best for them and then be able to move from one model to another."
This is a typical comment. It seems to be the new mantra being chanted by conventional licensed software vendors everywhere. I'm sorry, guys, it just won't wash. If the hosted option is interchangeable with the same old on-premises software, then it's just a load of old SoSaaS. The whole point of the on-demand model is that it doesn't get installed on customer premises, and that's a big deal because it means the vendor (and the customer for that matter) can Forget about implementation, focus on results.
When vendors follow that through, they end up with applications that are totally different from anything architected for on-premises implementation. Patrick Grady, CEO of Rearden Commerce, succinctly expressed the difference in an InfoWorld article, which I quoted in March [disclosure: I am currently doing some paid consulting work for Rearden]:
"On-demand is not a hobby. If you don't have a single-instance, multitenant, on-demand, pure SOA-based platform, then it's all crap and the customer will not receive any of the advertised benefits."
All true on-demand vendors get this. Here's what Chris Cabrera. founder and CEO of on-demand vendor Xactly, wrote in an email I quoted in April:
"Enterprise software by any name, hosted or otherwise, is not a substitute for true on-demand ... On-demand vendors view the model as religion, not a catch-phrase or contract terms ... In the end, hosted enterprise software is just a new look for a very old and tired set of problems."
So can you go to SaaS and back again? I don't think so. Not unless you started from SaaS in the first place. If an application is architected from the ground up for the on-demand model, then it is possible to implement it in-house in your own data center (though why you should want to, no one quite understands ... on-demand vendors that offer this option find that few customers take it up). The reverse, however, doesn't apply. Unreconstituted on-premises applications have no future in the on-demand world. I know I've said it before, and no doubt I'll have to reiterate countless times again. If conventional on-premises application vendors want to embrace the on-demand model, they're going to have to do a whole lot more than simply start hosting their existing software in a data center. It's just not as simple and easy as that.