How Performance Reviews Pay Off
Sometimes, the least productive workers will bounce back the most.
Authors: Francesca Gino (Harvard Business School) and Bradley R. Staats (University of North Carolina at Chapel Hill)
This paper examines the impact of performance reviews on productivity, and finds that feedback delivered on a regular basis, whether positive or negative, tends to result in improved performance. On a short-term basis, though, the impact varies, sometimes in ways that are counter-intuitive: Positive reviews, for example, do little to boost productivity, and negative reviews that are somewhat vague and indirect cause performance to fall off, but reviews that are directly negative cause productivity to leap. The research offers guidance to managers concerning the pitfalls and potential benefits in framing their messages in reviews, and suggests there is a need to provide feedback on a frequent basis.
In an ideal scenario, employees would be evaluated through the use of objective standards, but as the researchers point out, in organizational settings this is rarely possible. Instead, the very nature of performance feedback promotes what they call social comparison processes, as employees are informed about their performance relative to that of their co-workers. In this study, employees were not told their exact ranking, or that of their co-workers, but were informed where they stood in relation to either the “bottom 10” or “top 10” in terms of productivity.
The researchers conducted their experiment at APLUS, the consumer finance subsidiary of Shinsei Bank, a medium-sized Japanese bank. The 70 employees in the study performed largely repetitive tasks: They entered information from customer applications into a central data system. Their salary was not linked to their performance, and they had no specific goals to meet, which enabled the researchers to weigh the effects of performance feedback in an incentive-free context.
At the beginning of the monthlong study, the workers were split into three groups. One received negative feedback on a daily basis, a second received positive feedback on a daily basis, and the third, acting as the control group, received no feedback. The groups were randomly chosen without regard to past performance — in fact, none of the workers had ever before received a performance review from the company. Over the course of the month, the researchers analyzed more than 480,000 data-entering transactions performed by the three groups. By tracking the completion times and accuracy of the employees’ efforts, the researchers were able to measure daily changes in productivity for each of the workers.
Employees in the “negative” group were told they fell into the bottom 10 if, in fact, that was the case that day (what the researchers called direct feedback) or that they were not in the bottom 10 (an indirect approach that implied poor performance). In other words, even the best-performing employees in the negative group would get indirect negative feedback and know only that they were not ranked near the bottom. Similarly, employees in the “positive” group were told that they ranked in the top 10 or that they were not in the top 10.
On a day-to-day basis, the researchers found that neither form of positive feedback had much effect on productivity. Bad reviews, however, carried far more significance. When workers first received direct negative feedback, their performance jumped 13.6 percent, on average, the next day. But when employees first received an indirect negative review, they faltered, dropping an average of 17 percent in productivity the following day. ip info . The difference, say the researchers, is that those in the bottom 10 were motivated to improve by the shame they felt, whereas those who were not in the bottom 10 simply felt relief.
In the long term, however, all forms of the performance feedback used in the study helped. The researchers found that both groups receiving feedback boosted their performance over the course of the month in comparison with the control group; the positive group’s productivity was up approximately 20 percent and the negative group’s about 30 percent. So although couching an employee’s performance review in positive terms may not make a difference the next day, it will over time. And whereas giving an employee indirect negative feedback can hurt performance in the short term, over the long term it’s still better than no criticism at all.
The researchers acknowledge that the study’s setting — a Japanese bank — may raise questions about whether the results are widely applicable. Because reputation and saving face are particularly important in Japan, employees there might react more strongly to criticism than workers in other societies. But the authors point out that the feedback was private and did not include specific rankings. In addition, because the company had no history of laying off employees for inadequate performance and offered no bonuses for working harder, the researchers could focus solely on the effects of the feedback itself. These two factors make the findings relevant beyond Japan, the researchers say. The researchers also believe the findings are applicable beyond job settings that are highly routine and quantifiable.
The results show that although regular feedback can improve worker performance over time, the pace of change can vary. Managers shouldn’t expect to see an immediate increase in productivity from their best workers. As for the rest, indirect praise isn’t likely to produce an immediate uptick, and indirect criticism may actually make things worse for a while. Telling underperforming employees that they are in the bottom segment — which, of course, could be defined more broadly than the “bottom 10” — offers the best chance of getting a quick and dramatic improvement.
Managers should consistently tell their employees where they stand: Whether presented in positive or negative terms, feedback tends to improve performance over time. In the short term, the biggest improvement may come from workers who are told they are in the bottom rankings.
Publisher: Harvard Business School Working Paper No. 11-078