Other agencies could benefit from NMCI's hard-learned lessons

The long travails of the Navy-Marine Corps Intranet program offers numerous lessons that can be applied by federal agencies designing very large contracts, and by integrators looking to bid on them.Professionals involved on both sides of the NMCI program agree on many of the basic lessons:DO THE DILIGENCE. The Navy's streamlined acquisition process for NMCI, its adherence to the traditional practice of accepting questions from prospective bidders and providing written answers, and its own ignorance of the condition of its infrastructure combined to hide the number of legacy applications and their potential impact on other aspects of the contract.For instance, the requirement to support legacy applications fundamentally clashed with information security requirements and dragged down the entire program for several years.'In hindsight, I think they might think they should have spent a little more time behind the contract,' said Chip Mather of Acquisition Solutions Inc., a consulting company that advises government agencies on their acquisitions and procurements. 'A little more due diligence would have turned up [issues] like the number of legacy applications.'Mather said his firm now recommends replacing the Q&A process with a due-diligence period, to help competitors get a reality-based understanding of possible constraints and the culture in which they will be working.An EDS executive familiar with NMCI agreed with Mather's suggestion.'With a deal the size of NMCI on the commercial side, the company would have had six months of due diligence, and the contract would not be signed until the due diligence is completed,' he said. 'That's a valid lesson learned.'PRICE WITH PURPOSE. Optimism has no place in pricing strategies on government contracts'too many things can happen.One former Navy official said it seemed that EDS had priced very aggressively, and built its plans on the assumption that everything would go right. In the real world, he said, you may not know what will go wrong'just that something will rise up to trip you.Mather put more of the responsibility for pricing problems on the government side of the equation. 'I think that what caused the EDS pricing model to go so awry was the government meddling, [the] insistence that they apply Defense Department weapons system testing to a [commercial] product and service solution. The two do not mix,' he said.The EDS executive said that if his company could go back and do it again, they would separate the question of dealing with legacy applications. NMCI could have been set up just as a network, which would have allowed a lower security profile. Then the Navy could have slowly brought legacy applications online and monitored for security issues.RESISTANCE IS INEVITABLE. The failure to adequately address the sweeping cultural changes the contract required of Navy personnel has haunted NMCI from the beginning.When there is this kind of resistance to change, 'the government reverts back to really bad behavior ... abysmal behavior, really destructive behavior,' said Mather. 'Under performance-based relationships you get paid for results'the contractor's success is your success, the contractor's failure is your failure. So we're focused very hard on shared goals and objectives, to the point that we're recommending [that] the success measures in program offices should be the same as for the contractors.'THE FEWER METRICS, THE BETTER. While service level agreements'the metrics upon which a contractor's performance will be measured'are critical, EDS was overwhelmed by far too many SLAs.'Best practice in industry in this area is fewer, more meaningful metrics,' Mather said. 'Our approach is that competitors should propose them. Government, quite honestly, is awful at it.'The combination of different metrics may drive prices, for instance, but the government has no sense of where price point breaks are, he said.'They [the Navy] learned this lesson, because they subsequently negotiated down from hundreds of metrics to a real handful,' he said.SHARE THE LOAD. As a seat management contract, NMCI may just plain have been too big for a single vendor.'I think there's a limit to how much you can outsource,' said Warren Suss of Suss Consulting Inc. 'One issue with NMCI might be whether the Navy went overboard.'He pointed out the difficulties in coming up with three standardized desktops that had to suit the needs of all 400,000 users.'You need to introduce more flexibility into these kinds of outsourcing deals without leaving the government vulnerable to the big switch'the get-well strategy that says we'll lowball them and then make it up on the changes,' he said.A former officer in the Program Executive Office for IT agreed. 'I think perhaps that a seat management contract, without having alternatives in place, was simply destined for failure,' he said. 'My gut reaction? Unbundling the services and [going] for a more regional approach with multiple vendors.'