Roll Your Own Future

Now here's a real classic on the comeback trail: developing your own applications. Sounds so retro, doesn't it? The kind of thing start-ups do when the CEO doubles as the chief product engineer and surrounds himself with a cabal of MIT grads writing code. So what's going on when large pharmaceutical companies, insurers, hotel chains, health care providers and online powerhouses like travel firm Orbitz are found, in this day and age, productively rolling their own?

Computerworld reporter Gary H. Anthes answered that question last week in his cover story about the many sensible, cost-saving and even surprising reasons why companies build their own applications rather than buying into more packaged software ["Roll Your Own," QuickLink 47884]. What he uncovered flies in the face of conventional wisdom that buying is better than building -- a belief assiduously promoted by software vendors of all sizes.

And no wonder. The lifeblood of so many software companies increasingly flows directly from their maintenance and support fees, which have risen to nosebleed levels of 18% to 25% annually to offset the economic drag of lower sales in recent years.

Take Oracle as Exhibit A. When the database maker posted its financial results in mid-June, the single biggest factor cited as offsetting its slow-moving application sales was rapidly growing revenue from those fat fees for software maintenance. That revenue is increasing nearly twice as fast as new license revenue, CEO Larry Ellison said.

But it's not just the high cost of applications and their hefty annual fees that are driving development of homegrown applications. Ranking high as reasons for this approach are dissatisfaction with complacent vendors that don't respond quickly enough to user needs, and dismay over software suites overloaded with features and fiendish complexity. At Reinsurance Group of America, for example, a $35 million global enterprise administration system that was developed in-house not only fueled a competitive leap past the company's rivals but also was vastly preferable to the nightmare alternative of integrating more than a half-dozen commercial packages to provide similar capabilities.

Yet the greatest reason of all to roll your own is the ability to tailor IT to your business, to control the fate of applications too vital to trust to outside developers. It's about enabling (may Nicholas Carr forgive us here) a competitive edge that really does matter.

At Reliant Pharmaceuticals, for example, CIO Ron Calderone wisely heeded user resistance to complicated sales force automation tools and built a relatively simple system using speech recognition technology for the field agents. A packaged SFA system would have cost $4 million to $6 million, Calderone reckoned, but he delivered just what his business comrades needed for about 15% of that.

"Simple and inexpensive" are often the magic words associated with the best in-house application projects. We're hearing that mantra more often these days, particularly as open-source software carves inroads at the enterprise level. As Orbitz CTO Chris Hjelm put it in our story, "We are largely an open-source shop, so when we think about buying software, there's a general aversion to it."

The "buy vs. build" debate will no doubt go on forever. But the combination of open-source software with sophisticated development tools and standardized Web services is dramatically changing the face of that argument. When companies go looking for technical creativity, innovation and a competitive edge, they won't be buying that off anybody's shelf but their own.

Maryfran Johnson is editor in chief of Computerworld. You can contact her at

Copyright © 2004 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon