Premium on Performance

eth Hall hated the idea of changing the contract for the Environmental Protection Agency telephone hot line she managed. As program manager, Hall's interest was getting fast, accurate answers quickly for people who called the safe drinking water hot line. She was happy with the cost-reimbursement contract that kept six operators staffing the phones full time.

Faced with pressure to move to a fixed-price contract with specific performance standards, Hall balked. She was convinced that changing the contract would reduce the management control and flexibility needed to ensure contract operators were well trained and that there were enough of them on duty to handle surges in the call workload. "We said, 'Thanks, but no thanks,' long and loud, and we lost," Hall says.

In her resistance, Hall was like many other program managers being dragged kicking and screaming into performance contracting. But once she was forced into experimenting with fixed pricing and paying for performance, Hall was sold. The converted contract was tougher to write than its predecessor, but, though the new pact was with the same firm, it reduced the number of hot line operators by two, cut costs by 20 percent to 25 percent and requires less management than the previous contract while still meeting all Hall's performance goals.

Today in government the premium is on performance. The goal is to get measurable results that citizens value, and the method is setting standards and rating performance against them. Agencies are keeping book and taking names on the performance of their programs, employees and contractors. The pressure to get results has spawned new attention to performance in every aspect of acquisition, from writing contracts and choosing contractors to overseeing and rewarding their work.

The federal budget may be in surplus, but no one is proposing to spend more money on government. Instead, the budget boon will go to cutting taxes or saving Social Security. Caps on agency spending are likely to persist. Agencies will continue to face pressure to produce more and better results faster with less money and fewer people. Already federal downsizing has trimmed more than 300,000 civil servants from the rolls since 1993. Growing missions and declining staff mean agencies increasingly are turning to contractors to either do work or provide information technology that will enable the fewer remaining civil servants to do more. In fiscal 1997, the government spent nearly as much on service contracts, $110 billion, as on its own payroll, $113 billion.

The acquisition reforms of the 1990s are changing all that by shifting more financial and performance risk to contractors in the hope of lowering costs and improving outcomes. The two key performance pressure points in acquisitions are performance-based service contracting and the use of past performance in source selection. In May, the Office of Federal Procurement Policy reported that a performance-based contracting pilot project brought 16 agencies cost savings averaging 15 percent, improved customer satisfaction 18 percent and cut the number of costly contract audits by 93 percent. In January 1997, OFPP reported that using past performance as a major criterion for selecting the winners of 60 contracts raised government customers' satisfaction by 21 percent.

Past Performance

Contractors traditionally have had to have satisfactory performance records to do business with the government. Contracting officers also had to determine that firms could meet a contract's performance schedule before allowing them to compete. Nonetheless, procurement officers often shied from disqualifying bidders for fear of running afoul of laws enacted to ensure fairness in competition. "In reality, only in the rarest cases, and then only with the most grievous performance records, were contractors determined 'non-responsive' and eliminated from the competition," according to a December 1997 "White Paper on Past Performance" published by the General Services Administration's strategic information technology analysis division.

Efforts to increase the importance and use of past performance information began in earnest at the end of the Bush administration in late 1992. Then-OFPP Administrator Allan Burman issued a policy letter requiring past performance evaluations on contracts over $100,000 and use of past performance information in selecting contractors. The 1994 Federal Acquisition Streamlining Act (FASA) codified the requirement to use past performance in distinguishing among contractors. A revised Federal Acquisition Regulation (FAR) Part 15, which took effect Jan. 1, mandates the use of past performance in awarding contracts worth $100,000 or more.

Burman, now a contractor with Jefferson Solutions, a Washington change management firm, is feeling the heat of his own policy. "I definitely don't want to have anything in the record that would make people say 'I don't want to do business with him because he failed to deliver,' '' he says. "You go the extra mile. If the past performance isn't there, you don't even get to play."

"The government always looked at past performance-we looked for responsive and responsible contractors and asked for [information about] their similar contracts," says Betty Bailey, EPA's acquisition and management director. But the current focus on past performance "puts more teeth into it," she says. "Before, you had to say, 'You violated the terms and conditions of your contract.' It was more adversarial. Now you say, 'You need to pay attention here because this is something I will be looking at for past performance,' '' says Bailey. "It lets me talk to the contractor [during] the current contract."

Performance-Based Contracting

These periodic check-ins with firms during the course of a contract become formalized under the performance-based contracting approach. In essence, performance-based contracts define work in terms of what needs doing as opposed to how it ought to be done, thus freeing contractors to exercise ingenuity and creativity in meeting agencies' requirements. This focus on results is intended to encourage contractors to come up with cheaper and better ways of accomplishing work.

Performance-based contracts define work in measurable, mission-related terms; contain performance standards; include quality assurance plans for measuring contractor performance against those standards; and-if the work is mission-critical or involves very large expenditures-provide financial incentives and penalties based on performance measurements.

Defense agencies have used the approach for some time and have reported it works. So much so that in 1994, OFPP decided to encourage the practice govern-mentwide by getting 27 agencies to pledge to apply it to contracts due to expire. Twenty-six contracts valued at between $100,000 and $325 million were recompeted as performance-based pacts under the pledge. They included such services as drug testing, air traffic control and weather observation, janitorial services, telephone hot lines, data entry, and base operations and maintenance.

History Counts

A good part of the significant cost savings and performance improvements reported by agencies participating in the pledge resulted from their own cleanup efforts. Performance-based contracts require more upfront work on the part of government organizations before they issue requests for proposals. The only way to write a performance-based work statement, the basis of a performance-based contract, is by first closely analyzing what services and outputs really are needed from contractors.

EPA signed OFPP's pledge and targeted Hall's contract for conversion to a performance-based pact. To prepare, Hall did the first extensive analysis of call volume since the safe drinking water hot line was created and contracted out in 1989. She discovered that call workload was much more predictable than she had thought and that even call surges came within a range. "We had thought it was very volatile. It wasn't," she says. That information proved invaluable in writing a contract that could be both fixed-price and performance-based. "It's critical to have a good history so you understand what you want to buy and contractors know what they are bidding on," she says.

Pledge agencies found that scrubbing predecessor contracts before recompeting them as performance-based vehicles led them to change work requirements, in some cases reducing the amount of work covered by the contracts. Closer scrutiny of in-house operations often revealed outdated requirements that could be dropped and opportunities to streamline and reengineer so contractor contributions were both more limited and more useful.

As Hall and her colleagues reconsidered the hot line contract, they realized they didn't need to prescribe the type and amount of training the contractor should provide to operators and the tests operators had to pass to demonstrate what they knew. The new contract prescribes only that contractor employees are trained and have answers for 90 percent of questions while the caller is on the line or that they call back with a correct answer within 15 minutes. "We don't care if they are using hypnosis tapes under their pillows as long as they can answer the questions correctly," says Hall.

Similarly, the Railroad Retirement Board reported to OFPP that the process of developing a performance-based work statement for a data entry contract allowed the agency to eliminate 14 percent of the work previously included. The performance-based contract cost 30 percent less than its predecessor as a result of the more accurate and realistic portrayal of the work.

Required vs. Desired

Another advantage of the performance-based approach is that when agencies are more certain about exactly what outcomes they seek, they are better able to write fixed-price contracts. In the past, cost-plus contracts have predominated in the federal services field. In general, these pacts reimburse contractors completely for allowable, allocable and reasonable costs of performing a service and include fixed fees or award fees, as well. Cost-plus-award-fee contracts provide extra fees based on periodic assessments of contractor performance.

In theory, cost-reimbursement contracts are to be used only for work of uncertain cost or technical requirements. In fact, they have long been used for predictable and routine work by program managers who want a tighter rein on contractors via subjective award fee determinations or who want to retain the right to add additional work not foreseen in the original contract. One of the goals of performance-based contracting is to force program staffers to determine what work actually is required and to stick to it as opposed to inking open-ended deals allowing them to add work they desire.

"In the performance-based service contracting pilot it was the fixed price that had everybody nervous," says Hall. "There was a sense that the hot line was the heart of the operation, and we have surges. What if the Today Show flashes the number across the screen?" This fear had led the program to write a cost-reimbursement contract that required six full-time operators to be on duty all day, even when they didn't get enough calls to keep them busy. "By our hefty analysis of call loads we were able to say, 'This is the way it really works,' " Hall adds. Knowing the true nature of the workload let Hall shift to the contractor the responsibility of having the capacity to handle surges as opposed to paying to keep extra operators on duty even when they weren't needed. "We now pay not for capacity, but for what actually gets done," she says.

Ideally, the vast majority of performance-based contracts will be fixed-price vehicles in which all costs and profit are agreed upon up front and cannot be adjusted in light of the actual cost of performing the work. Fixed-price contracts place more cost risk on contractors, but leave them free to use innovative practices to increase profit by holding down costs. The combination of fixed prices and performance-based contracts also pleases many vendors. "We like to do things on a fixed-price basis, and it's important in terms of knowing the cost [of a project] to know exactly what the client is looking for," Burman says. "It adds a lot of clarity to what we're doing and to deciding how many resources we'll need to do the job."

Fixed-price contracts clearly save the government money. Cost-reimbursement contracts must regularly be audited and end with a lengthy final audit to determine final fees to be paid to the contractor. Fixed-price contracts need far less auditing. Indeed, OFPP found audit costs fell 93 percent on contracts in the performance-based contracting pilot, largely because cost-reimbursement pacts were converted to fixed-price deals. Contract prices fell 21 percent, on average, on the seven contracts converted from cost-reimbursement to performance-based-fixed-price deals, while prices fell 13 percent when fixed-price contracts also became performance-based.

Almost everyone involved in performance-based contracting agrees the biggest challenge is crafting performance-based statements of work. Even the Defense Department, where performance-based contracting originated, still struggles to back away from precise descriptions of how contractors are to work, says Stan Soloway, deputy undersecretary of Defense for acquisition reform. "I think the degree to which we're really good at writing performance-based statements of work needs to be looked at," Soloway says.

Reduced Oversight

When Hall's staff took their first whack at a performance-based work statement, they applied numerical values to just about every aspect of the hot line effort. But they quickly realized such precision would hamstring the contractor. Instead, Hall's people homed in on what they really wanted to buy and created a flexible contract with accountability built in instead of imposed after the fact by federal managers. "Performance basing makes it more objective," Hall says. "It is asking them to report and measure themselves against statistics we used to use in oversight."

Reduced oversight is a chief source of savings. "In an era of limited resources, [a performance-based pact] takes less resources to manage," says EPA procurement chief Bailey. "It lets the contractor do what he does best." Hall says the performance-based, fixed-price deals require far less work for the contracting shop as well. Programs realize that savings indirectly, she adds, as their contracts move more quickly through the contracting office.

To help agencies struggling with work statements, OFPP has issued a number of sample statements (available at www.arnet.gov/References/References.html) for services most ripe for performance-based contracting, such as automated data processing, software development, call center operations and language training. The Naval Facilities Command has sample statements for 23 services, mostly related to facilities maintenance (see www.efdsouth.navfac.navy.mil/gpws/).

But no matter how much help is available, some work-research and development, for example-still is tough to render into performance-based contracts. Program and contracting people are wrestling with writing performance standards for the work of scientific discovery-how do you base such contracts on results that aren't yet known? For other types of work, setting an appropriate level of government surveillance is dicey. The thrust of performance-based contracting is to replace constant oversight of contractors with insight gained from firms' own quality assurance systems. But in some instances, closer surveillance still is necessary, says NASA procurement analyst Kenneth Sateriale. In fact, he says, on some programs involving great risk to human lives, such as space shuttle operations, a new and even deeper form of pervasive oversight may be needed.

"I am convinced insight is dangerous to do when you're coming from a circumstance where the government has traditionally had the know-how and the contractor has relied on it," Sateriale says. "If you pull it away, you may jeopardize human lives." The alternative is to force dedicated government employees into "well-intentioned illegal behavior" as they continue inserting themselves in contractor operations to ensure lives aren't lost.

Too Much, Too Soon

The Energy Department has learned that there's such a thing as too much emphasis on performance-based contracting. Energy has committed more than $13 billion to huge management and operations contracts at nuclear waste sites the agency must clean up. Until the agency issued a contract reform initiative in 1994, those contracts typically were cost-reimbursement deals with broad work statements and only a few subjective performance measures. Largely because many of these contracts have spiraled out of cost and schedule control and yet have been regularly renewed without competition, GAO put Energy contract management on its 1997 list of programs at high risk for waste, fraud and mismanagement.

In 1994, to shift cost risks to contractors and spur them to better performance, Energy created performance-based management contracts with objectives measures and related financial incentives. Performance-based incentives (PBIs) are tied to successful achievement of performance objectives and measures. They let the agency emphasize performance of specific critical work efforts by tying substantial extra fees to them based on a preset schedule in the contract.

Energy has had some significant success with the new contracts and incentives since they began in fiscal 1995. For example, at the Hanford Site in Washington state, a contractor developed a new deactivation approach that cut 16 months and $77 million from the previous schedule and cost estimate at one facility. Performance objectives and fees goaded the contractor at the Idaho National Engineering and Environmental Laboratory's (INEEL) advanced test reactor to achieve the lowest operating costs since 1991, the highest operating efficiency since 1969 and the lowest number of unplanned outages in eight years.

But PBI contracting also has brought problems to some of DOE's 130 waste sites. At the Richland, Wash., operations office, a PBI program set up in fiscal 1995 contained $14.2 million in fees covering 34 incentive opportunities with 86 objectives. The Energy inspector general reported on March 10, 1997, that Westinghouse Hanford Co. was paid incentives for work done before the fees were instituted and for tasks that never were completed. In addition, the contractor once received a $225,000 incentive for installing a waste tank ventilation fan, a project that cost just $24,766. On another occasion, the contractor received a $250,000 incentive for work that took just two days. What's worse, the IG unearthed evidence that the contractor created a "near miss" incident compromising safety and quality in order to earn $225,000 for installing an air compressor system 19 days ahead of schedule. Since the IG report, Energy has recovered $2.8 million in questioned fees.

An October 1997 DOE report suggests the problems arose in large part from hurried efforts to put the new contracts in place, inexperience among those writing them and a lack of guidance from headquarters. Field organizations lacked historical performance data on which to base objectives and fees, according to the report. At times they relied too much on contractors to develop work scope and performance objectives, or failed to analyze the value of work before setting fee amounts. The department issued a "lessons learned" document in March and has been rewriting its regulations covering the use of fees for management and operations contracts.

Fixed-Price Problems

Fixed-price contracts also have their dangers, DOE has found. The government's largest contractor, Lockheed Martin, has been forced to fire one of its own subsidiaries, and one federal agency is fining another in the wake of perhaps the most spectacular recent failure in Energy cleanup contracts.

Following its 1994 contract reforms, Energy launched a privatization initiative designed to move from cost-reimbursement maintenance and operations cleanup contracts to fixed-price deals under which contractors assume any construction costs needed to produce the final product-usually remediated waste-Energy is buying. The first test of the plan came at INEEL's Pit 9, a one-acre field filled with buried radioactive garbage and sludge trucked in during the late 1960s from the Rocky Flats, Colo., nuclear weapons lab. Energy sought to privatize the Pit 9 cleanup, offering a fixed price contract for removal of the wastes.

Contract winner Lockheed Martin Advanced Environmental Systems (LMAES) said it would complete the project by February 1999 for $200 million or would return all the payments. But on June 2, INEEL operations contractor Lockheed Martin Idaho Technologies Co. fired LMAES at Energy's behest after the firm had failed to remove even a single shovelful of waste from Pit 9 in almost four-and-a-half years and had sought to recover $257 million in costs through June 1997. In addition, the Environmental Protection Agency and state regulators slapped DOE with a $940,000 fine over Pit 9. The two firms' parent company, Lockheed Martin, is suing Energy, contending the department interfered with LMAES and prevented it from doing the work. Energy says the firm was badly managed and was using a flawed approach.

The General Accounting Office has pronounced the project a failure that bodes ill for DOE's plan to privatize another 11 sites. Yet privatization continues at Energy, albeit with closer and earlier attention to contractor performance problems, clauses making firms responsible for fines levied by EPA and other agencies, and renewed focus on past performance in selecting contractors.

Many observers say DOE never should have contemplated, and the vendor never should have signed, a fixed-price contract for such tricky, unpredictable work as nuclear cleanup.

"If the risk is unmanageable, the contractor can't put itself in the position of saying, 'I will do these things for a set amount,' " says Burman. "The agency needs to be fair because it doesn't want to set [the contractor] up to fail."

Energy now says it made a mistake by accepting LMAES' experience on smaller cleanups as evidence of the firm's ability to handle Pit 9. Former Energy Secretary Federico Pena told a House oversight panel in July 1997 that the department now doubly scrutinizes contractors' past performance on work "of similar nature, size, complexity and risk."

Past Performance Perils

Despite Pena's endorsement, the use of past performance information in selecting contractors has not been problem-free. Industry continues to complain that past performance information isn't always collected or used fairly. But in a 1996 ruling, GAO gave agencies wide latitude in using the information. GAO ruled it would review an agency's evaluation of past performance only against the contract's evaluation criteria. Further, GAO found an agency can hold past performance against a bidder even when the firm disputes the agency's interpretation of the facts.

Complaints about inconsistent information should diminish as agencies adopt standard reporting forms to automate collection of past performance data. The Energy and Defense departments and the National Institutes of Health currently have automated systems. "Without [an automated system], you have a hard time getting any data that makes sense. Everyone just takes it upon themselves and collects something different," says Paul Coombs, manager of Energy's past performance database. "Now we have standardized information and score information on the same criteria." Ninety percent of DOE's sites have information in the database, which is the core of a dial-in, client-server system. This summer, Energy will have a World Wide Web-based system to access the database in place.

NIH's Web-based Contractor Performance System (CPS) contains NIH data as well as that of 32 other agencies that subscribe to the service for fees ranging from $1,850 to $15,000 a year depending on the number of contracts for which information is collected. In addition, agencies pay monthly usage fees that run about $25. Ratings in the Energy and NIH systems cover contractor performance in terms of quality, cost, timeliness and business relations. The numeric ratings in the systems must include supporting narrative comments. NIH reports that the system already is easing users' workloads. One subscribing contract officer who received 50 requests for evaluations of a single contractor reported how easy it was to simply print out the information from CPS or to direct requesters to the system if their agencies subscribe.

Late last year, DoD standardized past performance rating elements by business sector and created a five-level rating system. Departmentwide collection of past performance information began in February. Previously, each military service and Defense agency-and even subdivisions of the services and agencies-had its own rating system. In the new system, rating elements differ among business sectors. For example, services and information technology contractors will be rated on overall quality of product or service, adherence to schedules, business relations, key personnel management and cost control (except for fixed-price contracts). Systems contractors, on the other hand, will be rated on product performance, systems engineering, logistical support, product assurance and other technical performance. The new standards apply to contracts over certain dollar thresholds: $5 million for systems and operations support and $1 million for services and information technology, for example.

DoD also is putting in place an automated approach to using past performance information. Since there are a number of automated collection systems in place throughout Defense, the department has focused on putting together an overarching data retrieval system so people can gather contractor data from any of the existing DoD or other government systems. This summer, DoD tested an off-the-shelf, Web-based retrieval system at five activities.

The collection and use of performance information is likely to bring vast changes in every aspect of acquisition, from what agencies buy, how they seek contractor bids and how contracts are managed, to what and how firms are paid and whether they will be able to win other contracts. "You can't look at past performance in a vacuum," Soloway says. "It starts with the issuance of the solicitation and making clear what the requirements are." Contracting in the performance environment leads to more productive communication between agencies and companies from the beginning to the end of contracts, and to better results.