Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Prints of Darkness ]   |  A Lose-Lose Situation  |   [ Conscientious Objector ] >>

Weekly Column

A Lose-Lose Situation: Sometimes IT Integration Just Isn't Worth the Trouble

Status: [CLOSED]
By Robert X. Cringely

Beauty is in the eye of the beholder. Back in 2000, the Department of the Navy, part of the U.S. Department of Defense, awarded the largest-ever single contract for U.S. government IT services, almost $7 billion. The project was called NMCI -- Navy Marine Corps Intranet -- and was going to provide all data, voice, and video services and desktop and laptop computers for the 400,000 plus Navy workers who weren't stationed aboard ships or submarines. For one monthly price of around $200 per seat, all the IT needs and worries of those 400,000 folks would be handled efficiently by EDS, the prime NMCI contractor. No more Navy PCs, these machines would be owned by EDS. No more military technicians, every move, change, or repair would be done by EDS or one of its subcontractors. Computing would be simplified, more efficient, and deterministic.

Too bad it hasn't worked out that way.

NMCI, as it is rolling out today, has few boosters in the U.S. military. The transition has been awkward and far more expensive than it was supposed to have been. While someone may argue that NMCI has improved military readiness or saved money, it is almost impossible to prove either claim. It isn't clear, in fact, that the U.S. Marines are still even a part of NMCI, though the Marines' own NMCI web site doesn't say that. What IS clear is that the whole NMCI experience is a prime example of how NOT to buy IT services.

Under NMCI, even moving a PC from one side of the room to another is supposed to require a call to EDS. But there are limits to NMCI. There can be no application development on NMCI machines, for example, so Navy software developers (they do exist) have to use the NMCI machine for e-mail and their old PC for writing code, with the two machines on completely different networks. So much for network simplification.

All of the NMCI machines run Windows 2000, which is supposed to increase network security. Yeah, right -- as long as Linux boot floppies can be kept off the base.

Once NMCI is fully implemented, the Navy will have a whole new computer network owned by someone else. The Navy won't be able to do anything it couldn't do before, and because of forced application streamlining, there will be some things the Navy used to do with its PCs that it will no longer be able to do at all.

It would be easy to write this column making EDS into the bad guy, but in many ways EDS is really the biggest victim -- stupid, but still a victim. NMCI is killing the company and its stock because, frankly, EDS under-bid the job. Their profit was to come in barrels of money associated with exceeding customer performance goals. If EDS did a fabulous job, they could earn up to an extra $50 per seat per month (that's $20 million per month) in pure profit. But of course customers AREN'T happy, so those bonuses aren't being paid and EDS is losing lots of money on the NMCI contract.

EDS reportedly bid the five year contract amount instead of the seven year amount required by the Request For Proposal (RFP). Originally EDS was obligated by their mistake to give the last two years of service for free. Now I hear the Navy has agreed to amend the contract, which is funny because NMCI has no specific funding from Congress. The idea was that it would pay for itself with savings, but these savings now appear to be fictional.

The RFP, itself, was flawed, failing to disclose hundreds of applications that would require support. Oops. The RFP basically called for someone to take over all the IT functions of the Navy and Marines. There was very little information provided on the existing systems, and almost nothing on the current applications. The RFP had requirements for a NEW environment and required service level agreements (SLA's).All 483,000 PCs had to be replaced in three years with a second replacement three years after that. There were timetables for a new e-mail system, new network, etc. There would be a regular network and some secret networks.

The Navy wanted a fixed price contract for all this. You were to bid on five years, and on a two-year extension to seven years.The RFP didn't allow contract language to mitigate financial risks. So if the Navy later revealed an extra 100,000 PCs or two previously undisclosed submarine bases, that was just tough. It was a fixed price contract entirely on the Navy's terms.

But wait, there's more! Prior to NMCI, the Navy had hundreds of firms providing parts of their IT service. This was a very disjointed effort with little central coordination. Many costs were hidden and unknown, so no wonder they wanted to improve it. But a key part of the Navy's "diversity" quotas were achieved by these IT contracts, which involved many small companies employing lots of women and racial minorities, all required by law. While there is nothing wrong with diversity, the NMCI RFP required the successful bidder to absorb these little companies as subcontractors, keeping the diversity intact.

The NMCI contract specifically required at least 40 percent of the work go to these small businesses. The life cycle of such a contract would normally involve layoffs in the second half of the term, made possible as productivity improved, but the NMCI contract didn't allow such layoffs if they messed with the Navy's diversity numbers. This led to a conundrum (note -- in 37 years of writing for publication this is the first time I have EVER used the world "conundrum"): the successful bidder could promise no minority layoffs and probably never show a profit as a result, or the successful bidder could promise no minority layoffs but not really mean it. Not good either way.

EDS won the bid, but little integration work was done to make sure the new network, the new servers, the new LDAP, the new workstations, and the old applications would all work together. They started building a new network and installing new workstations before they knew what they needed. When the new stuff arrived, they couldn't get rid of the old stuff.

Where was the oversight that should have caught these problems before they even became problems? There wasn't much oversight at all. Late in the RFP process, the General Accounting Office appeared and started asking questions, but in the end the Navy cut the deal without much, if any review outside of the Department of Defense. They got away with that, remember, because there was no specific appropriation for NMCI.

There is an important lesson in this NMCI saga for both providers and users of IT services. It is doubtful that there was any way EDS could make a profit on the project under the original contract terms. It is doubtful, in fact, that ANY company could have made NMCI work, though EDS is still doing a grand job of pretending that the project is a success.

The reality of IT service contracts is that in the end the price alone will win or lose the deal. Even if it is obvious the bidder has seriously missed something, most RFP teams and almost all goverment RFP teams will still go with the low bidder. If the low bidder is the wrong choice, it becomes the bidder's problem and they are to blame. It is never the RFP team's fault. So the Navy was sloppy, but it doesn't matter because EDS should have known better.

Here is why this disaster is happening and why it will likely happen again over and over. For an industry supposedly dedicated to increasing user productivity, the IT service industry, itself, is uninterested in efficiency and productivity. It is possible to be the low cost provider, but that requires being the most efficient provider. Being efficient is contrary to IT's billable hour mindset. Learning to become efficient requires time and painstaking effort. IT is generally interested only in the quick sale, the quick win. Disasters like NMCI are driven by commission-based sales teams that will do anything to win the bid, even if it means their company loses money overall. A 10 percent error on a $6.9 billion project like NMCI could destroy almost any IT provider, yet still the sales teams (from every company, not just EDS) bid away. They know their commission checks will be cashed long before it becomes clear that the project is impossible to do at a profit and sometimes impossible to do at all.

Comments from the Tribe

Status: [CLOSED] read all comments (0)