Expect Less

Get a grain of salt for all that Web-based hoopla and handheld whiz-bang and all that talk of solutions for the enterprise. We're now several decades into the information technology era, but the safest bet for a large-scale IT project -- in government or in the private sector -- is that it almost surely will fail to keep its initial promises. And it will cost at least double the early estimates for whatever portion of those promises can be salvaged. And it will take at least twice as long as was first forecast -- for the full set of original promises -- to complete just that salvaged portion. It is the rule, not the exception: IT projects will be overhyped, overbudget and way overdue.

The Census Bureau is taking a political beating right now for the failure of the decennial census modernization project, which had promised to shift the centuries-old paper-based program into the Digital Age. Bureau managers pointed their fingers at the lead contractor on the project, while the contractor pointed a finger back at the bureau managers. The bureau said the contractor didn't deliver the expected technology. The contractor said the bureau didn't explain what it really wanted, coming back way late with more than 400 requirements that should have been identified from the get-go. For example, the handheld devices that the contractor prepared for census takers to carry worked with Microsoft e-mail, but the Census uses Lotus. Oops.

But the Census Bureau is hardly the first and most certainly won't be the last federal agency to fall on its digital derriere. The Internal Revenue Service, the FBI, the Thrift Savings Board, the Pentagon, the Federal Aviation Administration -- let's just say it would be a shorter list of agencies that haven't had a major techno-meltdown than a list of those that have. The public eye is naturally more concerned with taxpayer-funded bungles, but the private sector has an even longer list of IT project flameouts.

What makes these debacles all the more frustrating is that managers in both government and business have racked their brains trying to find ways to avoid such failures. They chop up projects into little pieces. They tie them to the overall enterprise. They get independent auditors to watch every step. They rewrite contracts to emphasize results rather than deliverables. They read all about other organizations' lessons learned.

Yet the failures go on. IT mismanagement is like a constantly mutating computer virus; it's immune to the patches and filters that are designed to prevent it from spreading.

And so maybe the best way to deal with IT mismanagement is not to search for the magic cure that will solve the disease, but instead to handle it the way we grapple with many disappointing facts of life, such as the common cold and e-mail spam: Expect it will happen. The next time your agency embarks on a big technology upgrade that promises to solve myriad problems, confront your optimism. Halve your expectations, double your budget and prepare extra-strength patience. You're going to need it.

Brian Friel covered management and human resources at Government Executive for six years and is now a National Journal staff correspondent.