Now That the DoJ Won’t Do It, Microsoft Should Break Itself Up

The federal government has dropped its antitrust efforts aimed at Microsoft, although the company has been officially acknowledged as a monopoly. XP is around the corner, and Microsoft is showing every indication that it plans to return with a vengeance to the practice of tying applications to the OS.

Now that Microsoft has essentially gotten its way, I would suggest its next organizational step (following the imminent XP release) be precisely the one it fought so hard against: it should break itself up.

Even without U.S. District Jusge Thomas Penfield Jackson's big mouth, Microsoft was never that likely to be forced into such an action, but a move in this direction could achieve some things the MS monolith could never accomplish. It would also be the ultimate "up yours" to the Department of Justice, something I know several Microsoft execs have been looking to deliver for many years.

Although the shareholders might squeal, a cleverly cleaved Microsoft could be an even more powerful set of entities. It could partially defuse the hostility felt by many in the software development community, it could enable Microsoft products to achieve a greater degree of interoperability with the rest of the computing world and it could enable the company to expand its reach into new markets without ultimately being tied to the desktop operating system. And, should its parts suddenly start having to behave like companies that value true partnerships, it could trounce the existing players at their own touchy-feely games.

Informix did something akin to this last winter when it split its applications business from its database business, based on customer fears that its applications would be tailored to work best with the Informix database. It knew that its customer base for these applications needed to span all databases in order to succeed, but the natural corporate tendency was to try to leverage them against the Informix database. With Informix facing strong pressure from bigger database competitors, such a practice meant shrinking market share for both lines of products.

Microsoft holds a dominant position in operating systems, so the need to untie applications is less pressing. But that will come, thanks to Linux and other open source efforts. Currently, Linux is in its infancy and applications have not yet generated the momentum to drive mass adoption on the desktop. However, the interface issues are being addressed and office-style applications already exist in fully functional forms, and a new generation of computer science graduates are entering the job market armed to the teeth with Linux programming skills. If the emphasis on total cost of ownership continues to strengthen, the economic arguments for Linux may offset the many benefits of Windows.

A split of Microsoft into independent units--perhaps focused on operating systems, enterprise and consumer applications and infrastructure--would also enable the company to avoid duplicating the efforts that others have already embarked upon. The .NET efforts, for example, are up against the Web services models suggested by IBM, Sun and other companies. The tendency for the Web services advocates to emphasize interoperability bodes well for the success of the concept. .NET, while it has advocates, is tied to other Microsoft products, and that in itself provides an excuse for decision makers to look at other companies' approaches. As a telecommunications executive once told me, "There's only one vendor for Windows, and once you're on Windows, you're a hostage."

A divided Microsoft would also be freed from the need to develop new products with the idea of interlocking them into other division's products in proprietary way. The more complex and varied Microsoft's applications become, the more difficult it will become over time to give them an edge with Windows and the slower product cycles will become. An example: the Windows operating system could be simplified and refined without the need to worry about support for an increasing number of features, many of which provide marginal benefits to most users.

Microsoft's only other realistic option is to acknowledge the increasing heterogeneity of the enterprise computing environment and make sure that its products are interoperable with those of its competitors. The much larger players in the hardware space have come to accept this reality in order to expand their markets and to appeal to choice-minded buyers. The Internet and the evolution of electronic business have now made software and data interoperability a critical consideration, and if a monolithic Microsoft can't accept this, perhaps a breakup would be the best thing it could do for its future.

And boy howdy, would it ever get judge Jackson's goat.