Intelligent Enterprise

The enterprise software market, and all of IT for that matter, has long struggled over two essential questions: What is the next killer app, and how can vendors get customers to pay for it? Leaving aside the first question for another column, one way to get customers to pay for new technology is to figure out a way to free up existing IT dollars for new development. And the easiest way to do that is to come up with some disruptive technology or business practice that lets companies throw out one piece of IT spending to free up capital for the next killer app.

The question of where new spending can come from is particularly important to the enterprise software market, which has been relatively successful of late in capturing sparse IT dollars and has no desire to see the trend go south again. Probably the most ambitious of all is Microsoft and its Business Solutions (MBS) group, the owners of the Axapta, Great Plains, and Navision ERP products. Microsoft's goal of growing MBS from its current $500 million per year to $10 billion in 2010 reveals an expected growth rate that has gotten competitors to sit up and take notice. There's only three ways to do this — organic growth, more acquisitions, and poaching existing competitors' customers — and not one spells anything but trouble for the competition.

The Linux Alternative

So with Microsoft positioned as the next big competitor and freeing up IT spending positioned as the worthiest of goals, it shouldn't be too surprising that enterprise software vendors are quietly pursuing Linux, open source, and other Microsoft alternatives, even as they continue to pledge fealty to the .Net deity. A number of these erstwhile competitors are realizing that getting customers to adopt non-Microsoft server, database, and, especially, desktop software could be the best way to free up the money customers need to buy their new killer apps. And getting customers dependent on non-Microsoft technology could also help competitors offset the advantages that Microsoft will undoubtedly have as its applications, server, desktop, and other software assets become increasingly intertwined and interdependent. In other words, the more ubiquitous anything-but-Microsoft technology can be in the enterprise, the more successful MBS's enterprise applications competitors will be.

The battleground for the anything-but-Microsoft option spans the entire technology stack — and doesn't necessarily require a Linux solution. SAP's announcement in November 2003 that it was cutting a deal with Sybase to make that company's database available for SAP's mid-market Business One application is emblematic of this fight. Absent an offering like Sybase, most Business One customers would probably pick up Microsoft's database offering. Instead, the new partnership helps keep the MBS technology advantage to the minimum.

Despite such deals at the database level, it's getting customers to shift to Linux, particularly in the desktop, that represents Microsoft's worst nightmare and the other vendors' secret hope. The desktop's role as the linchpin for Microsoft's next-generation enterprise software — the linkage of two major Microsoft technology initiatives, Project Green and Longhorn, will make the current ties between Windows and Internet Explorer look like child's play, which means that the most disruptive technology in the competition's arsenal may well be desktop Linux.

Getting even a small fraction of the desktops served by SAP, Oracle, and PeopleSoft to run desktop Linux could free some serious IT dollars and help shake off the all-for-Microsoft threat. Suppose you're a modest-sized SAP customer with 2,000 desktops using SAP R/3. At a conservative estimate of $3,000 in maintenance per desktop per year, you're paying $6 million to run your Windows desktops. One admittedly biased source, IBM, thinks that Linux on the desktop could knock about 50 percent off this cost. So let's play it safe and say IBM has overestimated by a factor of two. That still means using Linux could leave $1.5 million of play dough for your next project. And, in a not-so-amazing coincidence, $1.5 million just happens to be a nice average for the total acquisition cost of a new enterprisewide killer application.

How much capital could desktop Linux free up industrywide? Guessing the total number of ERP desktops is a highly ambiguous affair. SAP and PeopleSoft have bandied around an eyebrow-raising 50-million desktops each. But even if the two companies' combined total were one-tenth that size, $750 in savings per desktop would translate to $7.5 billion per year. Turn that savings into new license sales to SAP or PeopleSoft and you'll see smiles all the way from Pleasanton, Calif. to Waldorf, Germany.