Opinion: IT's problem is too many dipsticks

Many years ago, a major truck manufacturer discovered that its costs were higher than those of its competitors because it used so many more parts to build its fleet. On further analysis, it learned that it actually used eight oil dipstick designs, while its competitors used just one. Though each dipstick added design and manufacturing costs, the eight designs contributed no meaningful marketplace value. Many other parts of the truck were similarly different, but not differentiating.

Everyone agrees about the problem in IT where, with very few exceptions, enterprise IT organizations are drowning in low-value work. Our research shows that roughly two-thirds of most IT staffers' time is spent on the installation, migration, patching, security and compliance work that, while clearly necessary, contributes little or no direct business value. In some organizations, the figure is closer to 90 percent.

Pretty much everyone also agrees on the cause of the problem: stovepiped information systems. They were built for different purposes at different times, span a mix of largely incompatible hardware and software. They weren't built for today's rapidly changing business climate, and they often can't enable the innovation and agility the marketplace now requires. In some cases, the instability of years of bridging and integration work even threaten the operations of the company itself.

Oddly, there is often no one to blame for this sad and very frustrating situation, especially on the IT side of the house. Most of the current mess evolved from making what at the time seemed to be one good business decision after another.

The solution to the problem is more complicated. From a technology perspective, the adoption of service-oriented architectures (SOA) is seen as the best path forward. By defining activities as high-level services, companies can mask many underlying system incompatibilities and achieve new levels of simplicity, reuse and agility. Or at least that's how the argument goes.

In the real world, things are a bit more murky. While SOA is becoming an important way of designing and thinking about information systems, it's no silver bullet. In the short run, SOA actually increases system complexity and requires conceptual and architectural skills that many IT departments sorely lack. Most enterprises are still experimenting, and even the leading vendors talk about a relatively slow transition, with the real momentum for SOA not building until 2007 or so.

But technology is at best half of the solution. Information system simplicity and agility will never be achieved unless businesses adopt the necessary discipline. After all, many of today's stovepiped systems were the natural offspring of stovepiped organizations and business processes. This is why so many companies have systems that include the same customer names, addresses, account numbers, dates and other data.

The task for corporate IT managers is to first identify where the equivalents of the eight dipsticks reside and eliminate them. Until needless complexity and redundancy are significantly reduced, companies won't realize the agility that IT should enable. In the end, the rigor and power of SOA will mean little unless it's matched by equally rigorous changes in business behavior.

David Moschella is global research director at the Leading Edge Forum, a Computer Sciences company

Join the newsletter!

Error: Please check your email address.
Show Comments