A Pair of Pay-for-Performance CasesThere are a couple of short articles by Kevin Gray on BNET.com today that describe the contrasting experiences two organizations had with pay-for-performance compensation systems.
The not-so-hot case of pay-for performance was tried at Hewlett Packard in the early 1990s. As Gray explains, managers at thirteen HP worksites got approval
to adopt a pay-for-performance model, hoping to boost productivity and encourage a focus on team rather than individual performance. They designed a plan that tied 10 to 20 percent of their workers' pay to their team's performance.Unfortunately, the results were not good, even after managers made a series of adjustments, trying to tune the system so it would work as intended. The basic problem was that the existence of contingent pay meant workers' motivation was skewed and their morale was depressed. Employees:
- were resentful when circumstances beyond their control, such as slow delivery of parts, slowed them down.
- "refused to allow workers they saw as less experienced join them. Less movement between teams meant that less knowledge was shared or transferred among employees."
- found, if they missed their numbers, that they had trouble servicing mortgages and car loans whose size was based on the their income gross of bonus.
The moral of the story, Gray concludes, is that "Success is never merely about numbers, so don't turn your reward system into a numbers game." Instead, think about such performance issues as whether your compensation system is compatible with solid inter-group collaboration and steady talent development.
Oddly enough, the success story Gray writes about was largely a numbers game, but one that was set up in a context where employees and colleagues were satisfied that the numbers captured fully meaningful information.
The setting was North Shore-Long Island Jewish Health System (NSLIJ), which includes fourteen hospitals with, among them, almost 5,000 beds. The workforce numbers 38,000.
In 2003, NSLIJ agreed to participate in a pay-for-performance study organized by the Centers for Medicare & Medicaid Services (CMS). The goal was to see whether the pay-for-performance arrangement would improve the quality of patient care. Gray explains:
The rules were strict. The staff was given 30 measures to assess the treatment of thousands of patients. Heart attack victims, for example, had to receive aspirin within two hours of arrival, beta-blockers at discharge and smoking-cessation counseling. Pneumonia patients required flu screening and an assessment of the amount of oxygen reaching their blood. And surgery patients required antibiotics one hour before the first incision. The hospitals were graded on each criterion and given bonuses based on their performance.A big part of the reason NSLIJ succeeded with the program was that they invested the considerable time and effort required to get it up and running properly. They trained their staff, they created the necessary documentation, and they ensured that results were monitored by people with authority to make any needed adjustments.
Although Gray focuses on the experience of NSLIJ, he reports that the entire cohort of 275 participating hospitals registered encouraging results. For example, the program was credited with saving the lives of 2,500 heart attack patients in its first three years.
So, now, what's the moral of this contrasting pair of pay-for-performance stories? Most obviously, defining performance appropriately is essential for a program to pay off. What's appropriate in one industry and organization is going to be different from what's appropriate in a different setting.
But it also matters whether bonuses are awarded at the individual level, the team level (the HP case), or the organizational level (the NSLIJ case), which is not something Gray addresses.
I would argue that what is significant in this regard is the degree and nature of the linkages among individual, team, and organization performance. The tighter any linkage, the more important that a pay-for-performance program be directed at the highest level involved. For example, if work is largely accomplished through effective team performance, then well-formulated team bonuses are appropriate. Conversely, where linkages are loose, e.g., where a number of individual contributors are doing the work (as in many sales situations), individual bonuses are appropriate.