Making the case for performance-based service contracting

The same day back in 1994 when President Clinton signed the Federal Acquisition Streamlining Act, senior agency officials gathered in a meeting room nearby to sign a pledge for a pilot program on performance-based service contracting (PBSC). Almost four years later, the results are in: PBSC is a winner for both the federal government and the taxpayers.

PBSC calls for the government to express its requirements in performance-oriented terms, such as: "The vendor will develop a system for processing veterans' claim benefits that will process the benefits with an average cycle time of x, an error rate of y and customer satisfaction survey results of z." It is an alternative to detailed design specs that tell the vendor exactly what technical solution to use or to vague statements of work such as "deliver a system to improve our benefits processing." The method is widely applied to hardware buys, but it has been slow to come to information technology services or system development buys.

In the pledge, the 27 participating departments and agencies agreed to convert about 60 contracts to PBSC. The pledge was carefully designed to come as close as possible to a controlled scientific experiment. Only previously awarded contracts up for recompetition were eligible. The scope and volume of work had to be comparable. The idea was to allow an apples-to-apples, before-and-after comparison where the only thing that changed was introduction of a performance-based work statement, a quality assurance plan, positive and negative incentives around the performance parameters and a fixed-price contract structure. Due to the strict criteria and agency cancellations of a number of the contracts in the original pledge, only 26 contracts from 15 agencies ended up participating in the pilot.

The Office of Federal Procurement Policy has just issued its report on how these contracts worked. The results are an overwhelming demonstration of the power of PBSC to produce better value for the taxpayers. On average, prices— compared with the costs of the previous buy of the same services— declined by 15 percent in nominal dollars. After taking inflation into account, savings were about 30 percent. The price declines were dramatic— 21 percent in nominal dollars— for contracts changed from a cost-reimbursement to a fixed-price environment. The move to a fixed-price environment also resulted in a 95 percent decline in the number of contract audits.

But it wasn't just that prices went down. Customer satisfaction with vendor performance went up by almost 20 percent— from a mean of 3.3 to a mean of 3.9 on a five-point scale— compared with the previous contract for the same services. This increase in customer satisfaction held even in cases where the contract was converted from cost-reimbursement to fixed-price, contrary to claims that moving to a fixed-price environment creates an adversarial relationship between government and vendor. In the pilot contracts, there was no change in the number of contracts awarded to small businesses.

There was one negative. Lead time went up dramatically— over 70 percent— for those pilot contracts described as being for "professional and technical services." That was due to the increased time it took to develop a performance-based request for proposals for the first time rather than just taking the old one and putting a new cover on it. Much of that extra time will disappear when these contracts are recompeted, although this will be no consolation to people working on PBSC for one-time requirements. The source-selection streamlining introduced since the pilot contacts were awarded should help as well. And frankly, the extra time is worth it to get the kind of improved results that the pilot showed.

Of the 26 contracts, only two were for IT services (one for telecommunications support and another for data entry), although a number of the other contracts probably had IT elements. PBSC is still in its infancy for more complex systems buys. However, as agencies develop performance measures for their IT investments for capital planning/Clinger-Cohen purposes, these measures should be plugged into contracts supporting those investments. It's time for the IT community to get moving on PBSC.

-- Kelman was the administrator of the Office of Federal Procurement Policy from 1993 to 1997. He is now Weatherhead Professor of Public Management at Harvard's Kennedy School of Government.