The tumultuous launch of HealthCare.gov presents issues of complexity, failure and consequence unlike anything likely to be faced by state or local government officials. Yet just about every government has experienced some major information technology project that went awry.
I remember when I was deputy mayor of New York City and Mayor Michael Bloomberg had had enough of the delays and expenses associated with a payroll modernization implementation known as CityTime. Bloomberg asked me to remediate the project. The project's costs had ballooned from $63 million to nearly $700 million. In the end, the prime contractor had to reimburse $500 million to the city and the effort was brought to a successful close by moving the project management to a city agency.
Implementation snags almost inevitably come up on projects of such a large scale. Those of us who follow statewide benefit system modernization efforts have seen major issues from Florida to Texas to Indiana, where the state and IBM ended in court in a very messy and expensive exercise in finger-pointing over a $1 billion welfare IT system.
So what can we learn from these large, complicated and flawed implementations? I would suggest the following:
• Nothing will work in the absence of high- level executive leadership and a skilled government project manager. The more agencies a project touches, the stronger that leadership needs to be. Of course, a good leader manages collaboratively -- until a decision needs to be made. Negotiating and listening have their place, but decision-making delays not only slow a project down but also magnify problems.
• The project's "owner" not only needs to incorporate technology experts from the beginning but also should bring public employees who will be using the new system into the process from the start. Whether they sign up people for welfare benefits, administer payroll, help people choose the right health insurance product or apply collective bargaining rules, these workers know the endless variety involved in human behavior. The new system must reflect their knowledge; without it, the chances of failure increase immensely.
• Establish a specialized group to handle all details of a major project from contract negotiation through program management. Service level agreements need to be thorough, but at the same time there must be a change-order process with aligned goals since the contract will be outdated the day after it's signed. In New York City, Chief Information Officer Rahul Merchant was instrumental in the creation of the Technology Development Corporation, a city-affiliated nonprofit that works to manage complex IT projects in a timely and cost-effective manner and also supports best practices for all city technology projects.
• Get the networked governance right. Sometimes government takes the role of general contractor for a project. This puts a premium on a skill not widely held in government, that of the collaborative integrator of a complex system. In other cases, government allows a contracted integrator too much control, both over the project and its scope and over subcontractors. The CityTime project piled subcontractors on top of subcontractors, not only driving up costs but also adding unnecessarily to the project's complexity.
• Elevate substance over procedure in procurement. Procurement generally and IT procurement in particular too often revolve around issues of compliance, not value. A technically compliant procurement that results in a totally insufficient result should not be the highest goal of government. More program involvement that elevates best value is needed.
• Never underestimate the added complexity of a government system compared to a commercial one. Government systems that allow access to a program or benefit come with almost unimaginable layers of rules and procedures. Indiana's once-troubled benefits eligibility system simply did not have enough power to handle the volume and complexity of the rules that had to be applied. Modularized systems implemented in chunks help guard against disaster. Big-bang systems rarely work well.
• When in doubt, don't do it. Today so many options exist that building a massive government system should be the last option. Officials should ask what the private sector already does that can be leveraged. We don't often look to the Internal Revenue Service for lessons in efficiency, but when it began moving to electronic filing of tax returns it pushed some of the most complex issues off onto the H&R Blocks of the world, which wrote the front-end software that posted returns to the IRS.
HealthCare.gov, as it was originally launched, seems to stand alone in the breadth and depth of its technical failure. But it serves to remind us that putting up a big technology project is not merely about the mechanics of implementation. The underlying lessons should help state and local officials and managers dramatically increase the chances of success when they undertake their own complex projects.