Nick Carr takes up where I left off with Salesforce.com's vision of moving beyond CRM and into the packages software territory dominated by Oracle, SAP, Microsoft and other packaged applications. He makes a good point about the SaaS purists and the old guard heading for a potential showdown:
How large a role will the SaaS model end up playing in enterprise computing? Will SaaS thrive on the periphery, limited to either smaller companies or a few well-defined applications like salesforce automation or call-center management, or will it move into the mainstream, displacing the old on-premises model entirely? For the moment, SAP and Microsoft seem to be betting on the former (while Oracle seems more agnostic, happy to buy its way to dominance no matter what the future brings).
SAP and Microsoft have both dipped their toes into the SaaS pool, but, to horribly mix metaphors, they appear to be taking the view that SaaS is basically an "on-ramp" to big, traditional software implementations. It's a well-contained, manageable supplement to their traditional enterprise products, not a big, hairy cannibal that's going to feed off those old products. Microsoft's Live services are aimed squarely at very small businesses - the ones that can't afford to buy their own servers or hire their own IT staffs. If those businesses stay small, they'll stick with the subscription-based or ad-supported Live services. If they grow, Microsoft will be able, it hopes, to lead them to more traditional products. SAP, for its part, is introducing a SaaS version of its own CRM system, but, as Business Week reported, "SAP says it believes most large corporations that buy CRM as a service will use it only temporarily and eventually move the applications on-premises."
He points out that RightNow CEO Greg Gianforte, among other firmly in the on demand camp, is seeing a shift to the majority of its customers choosing hosted solutions than on-premises installations, and questions how anyone could make a business case against the efficiency of a hosted, shared service model over time. What Nick calls an "old highway with high tolls" doesn't compute.
Pendulums do swing. Intuitively I agree with you but I have seen outsourced functions brought back in house. I have seen CIOs start to custom build more as the TCO of packaged software (SAP, Oracle - license plus maintenance plus systems integration partners plus perimum salaries for those skills) has grown...it depends on vendor performance and pricing and market evolution...right now SaaS vendors are eager to please but when they grow to Oracle or IBM sizes, who knows how they will behave?
In this case, I side with Nick. The value proposition of SaaS for many kinds of applications will grow increasingly attractive. The secret sauce won't be in some proprietary code for creating features, such as customizing a template or linking to a database--it will be in delivering software as a service, providing the underlying platform that provides functionality that can be leveraged across mulitiple departments, corporations, geographies with high performance, security and realiability as the lowest possible cost...