|
The past 50 years of
computing can be seen not only as a fantastic explosion of new
capabilities, but also as a progressive accumulation of costly
software junk. The entire history of computing can be characterized as
an evolutionary -- and increasingly expensive -- sequence of
build-and-destroy cycles. That has to change: We're on the verge of an
intolerably expensive new cycle. But thanks to the options being
offered by outsourcers and software innovators, there's a way to break
the cycle.
With each new computing cycle, much of what had been invested had to be junked for progress to take place. Consider the following:
By 1990, the chaotic proliferation of powerful personal computers had created a control crisis that it was hoped would be cured by client/server architectures. We have now come to the end of that cycle, and the pundits tell us that the Internet ande-commerce offer the magic solution. Be ready to write down about $3 trillion worth of hard-to-maintain and insecure programs. That $3 trillion is a significant share of total anticipated software spending in the next 10 years. With the hardware component of IT budgets dropping because of falling hardware prices, most of the world's spending for IT in the next technology cycle will be for software and software-related costs. With rising IT budgets, that suggests software spending in the next 10 years of well over $10 trillion. Unfortunately, migration, conversion, integration and maintenance will absorb most of all software costs if the current, wasteful practices are allowed to continue. Executives looking for competitive advantage won't be satisfied with only a fraction of their total software budget being deployed for profitable value creation. U.S. executives need rapid innovation, not expensive preservation of what is becoming an increasingly costly relic. Executives have three options. First, they can continue on the current course of escalating computer budgets and painfully extracting value-creating applications from whatever funding is left over from maintaining the systems now in place. Second, they can find outsourcers that will act as custodians of the junk pile. Third, they can find technically capable innovators to come up with technologically advanced solutions at radically lower prices than a firm could on its own. The third option is in reach. New software technologies -- such as Java running in a networking environment, and relying on processor-independent code -- have life-cycle costs of less than $1,500 per function point. That's far lower than the comparable life-cycle cost of more than $4,500 of the current installed base. (The estimates come from software metrics expert Capers Jones, a credible source of such data, and my own estimates of code obsolescence.) Clearly, there's money to be found in managing softwarelibraries for modernization if you can set up a clever way to manage that transition. There will be profit opportunities for the junkyard custodians as well as for the innovators who can lead us out of the next cycle of trillion-dollar waste. I'm betting that most companies will wager on the innovators. There's much more value available from delivering new profit opportunities than from managing a dying well of assets.
Strassmann (paul@strassmann.com) has managed his way through every computer cycle because money was always available to pay for replacements. Unfortunately for current IT executives, there isn't enough money left to do that again in the next cycle. |
Go back up to the Strassmann, Inc. home page.