Readers offer stuffing for IT turkey
Why do large projects fail? The debate goes on
- By Kevin McCaney
- Dec 01, 2009
Our Thanksgiving retrospective on IT turkeys – government information technology projects over the years that went wrong – prompted readers to suggest several projects that should have made it to the roast, while sparking a debate over why projects fail.
“I'm shocked that ALMRS (an ill-fated BLM project in the mid-'90s) didn't make the list,” one reader wrote. A fair point. The Bureau of Land Management project, called the Automated Land & Mineral Records System, fell short of its goals. As Red in Washington, D.C., pointed out, it was finally killed late in President Bill Clinton's second term, "after almost a billion was spent on it.”
“The one project that definitely should have made the list was [Federal Aviation Administration] modernization,” suggested another reader. “The big software contractor (the Alphabet predecessors of HAL) spent almost $1 billion and 10 years and could not produce a single workstation (the contractor was big in hardware, too). Each of them were always blaming the other party for 10 years and finally FAA pulled the plug and we may be still back with software written in Jovial 50 years ago (it still works though). So next time you get on board a plane, think about this.”
(The reader clearly meant "successors," not "predecessors" in referring to the malevolent computer of "2001: A Space Odyssey.")
Rachel of D.C. metro agreed: “Perhaps because it goes back so far, but the FAA's Advanced Automation System (AAS) disaster [was a] $7.2 billion program stopped in '94 after spending $1.2 billion and the ability of the FAA to convince folks on the Hill that the problem was the procurement regulations - which they were permitted to rewrite (and, hmmm, how different are they?? but never mind) -- and AAS was divided into many smaller programs that are disasters today.”
What other projects deserved "honorable" mention?
“I am amazed that the Defense Integrated Military Human Resources System (DIMHRS) didn't make the list,” wrote Fred from Washington, D.C. “Ten years an not a single transaction processed!! That's an ERP.”
“How about the $1B UFMS [Unified Financial Management System] project" at the Justice Department, wrote another D.C. reader, who was echoed by another reader in the nation’s capital: “That turkey has been on the watch list for years and component roll-out dates come and go year after year.”
Obviously, there are failed or troubled projects that didn’t make our list (which we didn’t claim to be comprehensive). But why do projects such as these – backed by millions of taxpayer dollars awarded after a competitive bidding process and overseen by auditors – fail in the first place?
“Almost any troubled contract or program suggests strongly that it takes two to tango,” wrote Michael Lent of Washington, D.C. “When Boeing was dogging the first task of SBInet [the Homeland Security Department’s Secure Border Initiative – Network], after a little passive-aggressive scolding by the customer, DHS and the company locked arms to defend the program. Witness the tortured defense of SBInet’s progress by [former DHS Secretary] Mike Chertoff. Look at the fumbling, ineffective stewardship of Deepwater by the Coast Guard, even after finding that the delivered boats from NG/LM were unacceptable. The companies are still there, and for a long time after the damaging disclosures continued to perform Lead Systems Integrator types of functions for the government. Which brings up Boeing, again, and SAIC; how much damage did they suffer for ‘cancellation’ of the Army’s [Future Combat Systems]? As LSIs, they were driving the bus as suppliers and in many (normally) government roles in the program. Nary a head has rolled in the suppliers or the Army; ‘things happen,’ eh? While the blame game is fun and fodder for the media (which very rarely digs up wrongdoing), and deadly serious for the customers and the taxpayer, it does not prove to be that consequential for the suppliers or the government, except in the most egregious cases, e.g., criminal prosecutions, which occur once in a blue moon.”
A Northern Virginia reader, however, put most of the blame on government managers. “Each one of these horror stories involves a different prime: SAIC, CSC, Northrop Grumman, AMS, Accenture, EDS, and Boeing. These are the giants with literally hundreds of successful IT projects to their credit, both commercial and government. What is the common thread then? Government mismanagement. Unclear and constantly changing requirements, understated workload estimates, and doubtlessly a decision and feedback process on the part of the government that would try the patience of Job. I have not been personally involved in any of these projects, but I have worked for two of the primes as well as having had a 25-year career in the government. In no case did I ever see a project lack for proper staffing and support by the contractor. What I did witness over and over again was poor management on the part of the government that bordered on malfeasance. That is where IT acquisition reform is needed.”
That comment prompted a back-and-forth over private enterprise. “You can't just cite government mismanagement for being at fault without at least noting another commonality, namely that primes are in it for profit,” wrote an anonymous reader. “Most of their strategies are influenced by the market and their bottom line. I've seen primes drive costs through the roof with contract mods (NMCI??), overconfident schedules, and positioning for more work. Government mismanagement may play a part, but these two are constantly playing two different ballgames.”
A commenter in Northern Virginia shot back: “Of course they are in it for a profit! Do you think that American corporations are philanthropic nonprofits that perform work for the government at cost only? (It occurs to me that you might actually be that naive. You wouldn't be the first government manager I've encountered who labored under that misconception.) However, what you are overlooking is that contractors are subject to audits and review of all of their proprietary cost information and can have their invoices reduced or denied. If primes are ‘driving costs through the roof with contract mods,’ then that is another example of government mismanagement. I'm glad you mentioned NMCI. The Navy represented in its solicitation that the number of local command-unique applications that the contractor would have to support and migrate to be approximately 10 percent of what EDS found in their initial site surveys. Whose mismanagement versus profit motivation caused that cost increase? If NMCI was such a pearl of government IT management then where is the acquisition strategy for NEXGEN?”
Another commenter suggested that perhaps the problem is a lack of accountability, writing: “Is this really surprising to anyone? How does the saying go: No one ever loses their job by hiring one of the biggest contractors? What happens to the purchasers that make the disastrous decisions? Nothing, they move to another agency and are never held accountable.”
And maybe, in some cases at least, the issue isn’t profit or accountability, but a sufficiently skilled workforce. “The IRS wants the cheapest possible people working on their projects,” writes an anonymous reader. “You can't do complex projects with people who have little or no experience, but that's all the IRS is willing to pay and that's what wins the contracts. If you always do what you've always done, you will always get what you always got. If the IRS wants a chance of fixing their system, they will prioritize getting older, more experienced (and therefore more expensive) people on their projects. No guarantees, but after 20 years it might help.”
Kevin McCaney is a former editor of Defense Systems and GCN.