Most technology vendors, especially those that have been around the block a few times, reach a point when they must decide whether they're going to stick to their core products and business models or head in a new direction. Their dilemma: If they move too early, they stand to cannibalize their revenues and shrink profit margins; if they move too late, nimbler competitors will pick off their customers and eventually leave them behind.

Among the vendors that have come to this crossroads at least once in their storied lives are Apple, CA, DEC, Dell, EMC, IBM, Novell, Sybase, 3Com, and Wang. The extent to which they reinvented themselves as customers and industry trends dictated is reflected in their balance sheets and income statements, or in their epitaphs.

Software as a service is starting to put the major software industry players to just such a test. IBM, Oracle, Microsoft, SAP, and other entrenched purveyors of on-premises software suites and platforms have considerable installed bases to protect, whether those were organically grown or acquired. Snapping at those vendors' heels are scores of venture-backed SaaS startups, promising that their Web-delivered sales force automation, marketing automation, business intelligence, procurement, network management, payroll, expense management, HR, and other services are the future of business software. Enterprise-scale contracts landed by vendors such as Salesforce.com, SuccessFactors, and Workday show that SaaS is no longer just a small-customer phenomenon.

But let's not get swept up by the SaaS euphoria too quickly. While 30,000- to 100,000-seat deals for sales force automation and human capital management services are starting to emerge, those are relatively straightforward processes. Enterprise-class applications for managing supply chains, transactions, and regulatory compliance are far more complex and tend not to be the same from organization to organization. For all the advances in SaaS scalability and customization, there are many processes for which on-premises software will remain the rule. Component-based applications built around a service-oriented architecture may one day replace the software suite and platform approach, and SaaS will fit into that architecture, but that mass movement is still years away.

Meantime, if you think the software giants are just hanging on to a dying business, consider that four vendors with more than $5 billion in annual software revenue--IBM, Microsoft, Oracle, and SAP--generate an average operating margin on that revenue of about 30%, according to a recent article in Barron's. That compares with operating margins of about 16% for software vendors with $1 billion to $5 billion in revenue and 10% for those with $250 million to $1 billion in revenue. The conventional software business is still plenty profitable, especially for large-scale vendors.

So if the Big Four or Five software companies aren't overhauling their product lines and business models to emphasize SaaS, you can understand why. All of them, of course, are emphasizing the "c" word--choice--as they roll out software services here and there. For the bluntest talk on this subject, you can count on Oracle chief Larry Ellison. Subscription-based software is "very interesting," he told financial analysts last fall, noting Oracle's dabbling with Siebel CRM On Demand, "but so far no one has figured out how to make any money at it." Ellison should know: Besides owning a big chunk of Oracle, he holds a majority stake in SaaS vendor NetSuite and a small stake in Salesforce.

Some SaaS vendors want to engage in a debate about the cost and manageability merits of premises-based vs. single- and multitenant vs. hosted software architectures. In the end, customers will decide which conventional or SaaS approaches make the most business sense. If nothing else, though, it's exciting to the see the software industry on its toes.

Rob Preston,
VP and Editor in Chief
rpreston@techweb.com

To find out more about Rob Preston, please visit his page.