On the folly of rewarding A, while hoping for B

A new year means new goals and targets for many teams. While targets can help motivate and orient a team, coaches need to make sure their targets are pointing athletes in the right direction. A poorly chosen target might reward the wrong behavior, and history has given us countless examples of this happening.

Targets gone wrong: an NHS case study

In the 1990s, the British Government began setting targets for the nation’s National Health Service (NHS). These targets were—initially at least—logical and sensible: reduce waiting times in Accident and Emergency (A&E) departments, for example, and all non-emergency surgeries to take place within two years. The initial effectiveness of these targets led to the introduction of both more total targets, and more aggressive targets, by Tony Blair’s Labour Government. One of these targets was that 100% of patients attending an A&E should be seen, treated, and admitted or discharged within four hours. NHS Trusts that did not meet these targets could be fined.

On the surface, this seems sensible, until we consider that there are some patients that might need to spend more than four hours in A&E; perhaps they need further investigations to get to the source of the problem, or have such a non-urgent problem that they don’t represent a priority. By setting this target (later reduced to 98% in 2004, and 95% in 2010), the government started to misalign the outcome they wanted—efficient and effective healthcare—with what they ended up driving.

To achieve their targets, many A&E departments could just transfer the problem elsewhere, admitting patients to their required specialist wards within 4 hours, and letting those wards deal with the patients. This enabled them to see and admit the patient within four hours, but increased patient numbers in other wards. Furthermore, as NHS Trusts were fearful of being fined for missing targets, they began to employ more managers, to drive performance in line with meeting the ever-increasing number of targets they were given. Because healthcare in the UK is free at the point of service and paid for by taxpayers, the NHS doesn’t collect any revenues; it therefore has a reasonably static pot of money to use. By employing managers—who require a wage and an office—the NHS was less able to pay for doctors, and also had potentially less space to house patients. This, in turn, lead to worsening working conditions for doctors and nurses, including increased hours of work, making them less likely to hit their targets, meaning more managers would be employed, making the problem worse—essentially putting the NHS into a downward cycle.

Where rewards go wrong in sport

This is a typical example of what Steven Kerr refers to in his classic 1975 paper “On the folly of rewarding A, while hoping for B.” In the paper, Kerr outlines that all animals, including humans, seek information as to which activities are rewarded, and then do—or at least pretend to do—those activities, to the exclusion of activities that are not rewarded. This causes an issue when we try to create performance-based incentives, as we end up rewarding—or trying to reward—the end result, and not the process; which, as in the NHS example outlined previously, can cause further issues.

We can often see this in sport. As an example, let’s say that a Governing Body wants to employ the coaches of its elite athletes. On the surface, this is a very good idea, but one potential downside is that it might cause coaches to stockpile athletes, in the hope that one of them will become elite and they will then get rewarded. This might result in fewer elite athletes overall, because they are all vying for the same coach’s attention. Recognizing this problem of stockpiling, the same Governing Body may then wish to reward the development coaches of future elite athletes; perhaps they develop a program by which the person who coached an elite athlete when they are 18 is rewarded in the case of that athlete winning a medal, even if they are not coaching that athlete at the time they win the medal. Again, this is a good thing, but it runs the risk of just transferring the problem down the chain; perhaps instead of stockpiling elite seniors, coaches might stockpile good junior athletes, in the hope that they will be rewarded in future.

Similarly, for an individual athlete within a training program, we might reward—through praise and attention—good training performance and testing scores. This, in turn, may lead to a scenario in which the athlete does all they can to achieve great training and testing performances, at the expense of fatigue and future under-performance in competition. Here, the athlete is substituting future success—the outcome athlete and coach want—for immediate attention and gratification. In this case, this praise and attention, which the coach hopes is driving performance, actually sacrifices it due to misaligned incentives.

Focusing on behavior

In his paper, Kerr identifies some of the key causes of this issue. The first is a fascination with an “objective” criterion; we seek to provide a quantifiable standard against which to reward performance, without thinking of the unintended consequences. The second is that we overemphasize highly visible behaviors; we may reward a team-mate who exhibits visible and obvious positive teamwork behaviors but who has highly destructive, non-obvious behaviors that occur in the background. The third is hypocrisy; the manager or coach say they want to reward a certain behavior, but when the star player demonstrates damaging behaviors, they are still rewarded due to their on-the-field performances.

So what’s the solution? Kerr writes that coaches and managers should explore what behavior is actually being rewarded, versus what we are hoping to reward. They should consider what the unintended consequences of this rewarding behavior might be and consider whether the rewards are actually driving negative behaviors. It’s not that rewards are dangerous in and of themselves, it’s just that we need to spend time considering whether they are actually driving the outcome that we want—or, instead, and incentivizing negative behaviors.