Software that actually works
The problem for all conventional software vendors is that their business model is built around products that don't work. For some reason, customers have bought into this. They hand over their money, take delivery of the product, and then start figuring out how to get it working. Mostly, they bring in specialist consultants to help them figure it out, at even more expense. Once they get it working, they then pay the software vendor even more in maintenance charges to ensure they'll get help fixing it when it breaks again.
On-demand vendors have a completely different orientation. They contract to deliver a working application, and customers start paying from the day they start using it — not before (except for setup costs, if required). The notion of charging some kind of maintenance fee on top to help keep the service running is almost ridiculous. In fact, most reputable providers operate under service level agreements that mean they get paid less, not more, if the product stops working.
This contrast in attitudes goes to the heart of why conventional enterprise software vendors have struggled — and will continue to struggle — with delivering on-demand applications. What they don't get is that the on-demand model isn't about delivering software per se. It's about delivering the results of successfully using the software. Greg Gianforte, CEO of on-demand CRM vendor RightNow Technologies, spelled out the difficulties this causes for conventional 'old guard' vendors in an email to me a couple of weeks back:
"Without a culture that truly values customer success, getting customers to renew on an annuity license plan is near impossible. The 'old guard' has never even been responsible for customer success, because they have relied on D&T, Accenture and others."
Conventional software vendors earn money when they ship their products to customers, irrespective of when the software is actually installed and working on the customer's premises. Customers can't even ask for their money back if they never get it working. Imagine if a company that had high project failure rates for implementations of its software were forced to return those license revenues. The effect on its balance sheet would be devastating.
This nightmare scenario (for vendors; the customers involved would probably be quite pleased) is unlikely to happen, of course — but ongoing maintenance revenues are another matter. It's reported that more than a third of Siebel's revenues last year came from maintenance, and about 45% of Oracle's. Competitors are already snapping away at these highly profitable revenue streams (which are reckoned to offer margins as high as 70%), but if Oracle is as serious as Larry Ellison claims about competing in the on-demand market, then the company is going to have start transitioning its business model away from maintenance charges and towards shipping applications that actually work, on demand.