Kill your performance ratings
Neuroscience shows why numbers-based HR management is obsolete.
s+b celebrating 25 years: Articles make their mark when they effectively challenge assumptions and practices that are widely accepted. In 2014, neuroscientist David Rock and colleagues Josh Davis and Beth Jones convincingly argued that the most powerful tool organizations use for human capital management was broken. They urged leaders to evaluate people based on the unique role they play in moving the organization forward.
Evidence is mounting that conventional approaches to strategic human capital management are broken. This is particularly true for performance management (PM) systems—the appraisal approaches in which employees (working with their managers) set goals for the year; managers interview others who have worked with them and write up an appraisal; employees are rated and ranked numerically; and salary, bonus, and promotion opportunities are awarded accordingly. A 2013 survey by the Society for Human Resource Management asked HR professionals about the quality of their own PM systems; only 23 percent said their company was above average in the way it conducted them. Other studies uncovered even more disdain. According to the Corporate Executive Board (CEB), a management research group, surveys have found that 95 percent of managers are dissatisfied with their PM systems, and 90 percent of HR heads believe they do not yield accurate information.
The performance management systems in many companies are misleading, cumbersome, and complex, requiring some HR departments to put aside an entire quarter to manage them. More important, they can be counterproductive. In the context of neuroscience research, most PM practices turn out to damage the performance they are intended to improve. That’s because they are based on a fundamental misunderstanding of human responses, as revealed in recurring patterns of mental activity.
There are at least two basic problems with performance management. First, labeling people with any form of numerical rating or ranking automatically generates an overwhelming “fight or flight” response that impairs good judgment. This neural response is the same type of “brain hijack” that occurs when there is an imminent physical threat like a confrontation with a wild animal. It primes people for rapid reaction and aggressive movement. But it is ill-suited for the kind of thoughtful, reflective conversation that allows people to learn from a performance review.
For example, a supervisor might say, with the best of intentions, “You were ranked number 2 this year, and here are some development actions for the future.” In this company, which scores its appraisals on a 1–3 scale, a 2 ranking is supposed to represent high praise. But a typical employee immediately disengages. Knowing that others were ranked still higher is enough to provoke a brain hijack. The employee may not say anything overtly, but he or she feels disregarded and undermined—and thus intensely inclined to ignore feedback, push back against stretch goals, and reject the example of positive role models.
The second problem with PM is that it fosters an incorrect but prevalent view of human growth and learning. As Carol Dweck, the Lewis and Virginia Eaton Professor of Psychology at Stanford University, has discovered, most people hold one of two implicit theories about human growth and learning. The “fixed mind-set,” as she calls it, holds that intelligence and talent are basically established at birth and remain static throughout life. People are born smart or not, and there’s not much anyone can do about it. The “growth mind-set,” by contrast, holds that people learn, grow, and improve all their lives. This is accurate; most people do learn throughout their years. But they could learn far more effectively, and bring more of a high-performance attitude to everything they did, if they weren’t held back by the mental paralysis associated with the fixed mind-set.
Few people are thoroughly inclined toward either the fixed or the growth mind-set. Some people, for instance, might go to work with a fixed mind-set about their ability to be creative, believing that they can never become any better at innovating new products than they are today. But they might have a growth mind-set when playing classical piano, associating the rigors of daily practice with their ability to improve.
If you want a high-performance organization, you have to reverse the destructive effects of conventional PM.
Unfortunately, the fixed mind-set is prevalent in many organizations—and reinforced by PM systems. This induces many employees to avoid the kind of effort that leads to learning and professional growth. Think back to your days in school. Chances are, no one liked to appear as if he or she had to work really hard to get good grades. They feared that others might think they were not naturally talented. Similarly, in a work environment where the fixed mind-set holds sway, people will typically strive to avoid difficult challenges. Any stretch goal or strategic imperative, no matter how worthy, will be seen as an invitation to fail. Unsurprisingly, this mind-set also makes people more likely to cheat. In one study, people primed just before a test with fixed mind-set suggestions (for example, statements that they could not change or grow) were 300 percent more likely to cheat than their growth mind-set counterparts.
The effects on organizations are devastating. Conventional performance management has been linked to high levels of attrition, low productivity, and significant problems with collaboration. PwC’s 17th Annual Global CEO Survey, conducted in 2013, found that 93 percent of the CEOs surveyed recognized the need to change their talent practices—something their companies were doing wasn’t working. Meanwhile, according to Korn Ferry Institute leadership development researcher Robert Eichinger, the ability to “grow talent” is ranked 67th out of 67 competencies for managers, despite decades of investment in PM systems. In other words, on average, managers are worse at developing their employees than at anything else they do.
If you want a high-performance organization, you have to reverse the destructive effects of conventional performance management. You need to find ways to evaluate people that recognize the unique role each person has played in moving the organization forward. These evaluations must be based on a growth mind-set: They must recognize that with the right context and conditions, anyone’s abilities can be improved, especially given the expansive, flexible nature of the human brain.
The starting point may be educating your company’s leaders about your current performance system, especially if it is based on numerical rankings. Because these rankings are so deeply ingrained in most organizations, and so closely tied to both the fight-or-flight response and the fixed mind-set, you may need to build awareness to get people ready to consider change.
Ranking and Its Discontents
In theory, a standardized, objectives-based PM system should be a straightforward exercise. It seems logical to set goals, provide feedback about progress during a six- or 12-month cycle, agree at the end of the cycle on whether these goals were achieved, and link an individual’s compensation to the results. It is also clearly easier to manage this approach with a common rating system for all employees. Thus, by the 1970s, most large organizations adopted a rating system to define human performance. Typically, these ranked people on a 1–3 or 1–5 scale. The lowest number denoted an outstanding performer, and anyone with the highest number was a problem employee.
Soon, many companies discovered that managers tended to rate everyone in the middle. As one executive from a large food manufacturer (with a 1–5 rating system) explained, “Anything other than a 3 requires extra work for a manager. You have to justify it if you give an employee a 1 or 2, and you have to put the employee in a performance improvement plan if you rank them 4 or 5.”
Starting with General Electric in the late 1980s, under Jack Welch’s direction, companies tried to solve this problem by requiring managers to differentiate their people and spread out the ratings on a curve. In this system, known as “forced ranking” (or “rank and yank”), a specific percentage of people had to be ranked at the top and bottom ranges, with their promotions and bonuses affected accordingly. A multibillion-dollar human capital management software industry emerged to support this type of system, which became commonplace in large companies, where it was seen as a consistent way to make rewards match performance. As of 2012, according to the Wall Street Journal, more than 60 percent of Fortune 500 firms still used this approach.
Yet both managers and employees find these systems dispiriting and exhausting. Kansas State University management professor Satoris Culbertson, who studied the response to more than 200 performance reviews, argues that the mere act of receiving a numerical rating can be perceived as negative feedback, and even people with a growth mind-set don’t react well to negative feedback. “You would think that having so many smart people try thousands of variations of a [ranking] scale over decades, someone would have found the right way to do this,” says Brian Kropp, who leads the CEB human resources practice. “But no one has.”
The SCARF Hypothesis
Why do performance rankings trigger a fight-or-flight response in employees’ brains? A neuroscience-based framework called the SCARF model, codified by David Rock, posits that five organizational factors have an immense, but often unnoticed, effect on negative human reactions. These factors are status (the perception of being considered better or worse than others); certainty (the predictability of future events); autonomy (the level of control people feel over their lives); relatedness (the experience of sharing goals with others); and fairness (the sense of being respected and treated equitably, especially compared with others). When an organization’s perceived level of any of the SCARF factors is low, people feel threatened and perturbed. Even if they don’t express it, the feeling is there, and it often impairs their productivity and willingness to show commitment. Performance rankings trigger that perturbed feeling in all five ways:
1. Status. People in lower-status positions have been labeled with numbers throughout history, as a way of dehumanizing and demoralizing them. In performance rankings in most organizations today, any number except 1 automatically signifies a lower-status position, with pay levels and promotion prospects to match. People carry that number, and the insult implicit in it, mentally around with them for a year, until their next performance review. As Korn Ferry’s Eichinger says, “Imagine getting a family together, lining all the kids up in a row, and telling them how they rank compared to each other.” Parents wouldn’t do it, because it would make most of the family members uncomfortable—including whoever was ranked at the top.
2. Certainty. The process of determining how people are rated is usually set in the human resources department, and is often opaque to everyone else. People may work as hard and as cleverly as they can, but they still don’t know if this will get them a higher rating. That’s because the rankings reflect not just their individual performance, but their perceived contribution compared to a cohort of their peers. If everyone is performing optimally but the manager can reward only the top 10 percent with the highest ranking, the employee’s sense of control is thrown out the window.
3. Autonomy. When a student gets a poor grade at school, there is generally a clear path to improvement—a path that may involve studying harder or seeking extra help. Whatever it is, the individual has some control. In an organization, a clear path is not always evident. Improvement may depend on factors (such as customer response to a product or the willingness of others to collaborate) that employees feel they cannot control or even influence. Though workers may actually have more influence than they think, the ratings trigger a sense of lack of autonomy. They reinforce the perception that the employee is neither trusted nor empowered. Moreover, in most PM systems, the focus is on past performance, not on future potential. This sends an unconscious message to employees: Their capacities are fixed and may never change. If those workers have not had much autonomy, or experienced it in the past, they are unlikely to see that change in the future.
4. Relatedness. If only one person can be given the top rating and get the best bonus, suddenly employees have good reason to undermine one another’s ratings, rather than to collaborate. “We spend performance season getting battle ready,” explains one manager. “Two weeks before my review, I begin to prepare my attack.” In one celebrated case, Microsoft’s ranking system (since discontinued) was blamed for the company’s decline in performance. Although Microsoft did not apply ranking quotas to individual small teams, a Vanity Fair article quoted a programming staff member who clearly believed the company did, and who said, “It leads to employees focusing on competing with each other rather than competing with other companies.” A similar unintended consequence occurs up the hierarchy. If employees feel that their bosses are comparing them against their peers, they will not openly share information that might compromise their ranking.
5. Fairness. A CEB study showed that at least two-thirds of the people paid as top performers are not actually seen by their peers as contributing the most to the enterprise. Perhaps that’s why another survey, this one from the research firm i4cp, found that 75 percent of the respondents believed performance systems were not fair. This represented a threefold increase since 2008, when it was only 25 percent. Unfairness is perhaps the biggest problem with forced ranking, because the system is set up in a way that makes the decisions seem more arbitrary every year. In most companies, after several years of a forced ranking system, poor performers have already left. Everyone who remains performs above expectations, almost by definition. But managers are still forced to rank their subordinates on a 20/70/10 scale, and the bottom 10 percent are required to leave. Thus, every year, at least one highly motivated, highly capable employee is ranked as a bottom performer. This ends with the employee exiting the company, and the manager deeply frustrated.
Of course, all these factors reinforce one another in ways that worsen the effect. For example, when people feel that others were ranked more highly without merit, and have no recourse to complain, the combined lack of fairness and autonomy can generate a much stronger emotional reaction than either would alone.
The result of all this? People feel unappreciated. They become more conservative. They set their goals low to ensure that they are seen as succeeding. They retreat from candid conversations about development, because the whole issue of progress and feedback is so emotionally charged. The experience becomes one of “ticking a box.” There is little of the type of conversation that actually promotes personal growth.
The rating system is particularly harsh on those who conduct the appraisals. Supervisors feel pressure to continue to show improvement, raising some people’s rating over time. They also feel pressure to differentiate, leading them to scapegoat some of their subordinates as poor performers. Steven Rice, executive vice president of human resources at the technology company Juniper Networks, explains why it deliberately shut down its forced ranking system: “The critical practice of letting someone know where their performance authentically stood became hijacked by artificially categorizing individuals into forced ratings in order to meet a fixed compensation budget. The process lost its integrity. In the majority of situations, it rendered the performance feedback incongruent with compensation and the rating. Ironically, an HR process designed to drive fairness resulted in mistrust.... Managers blamed it for tying their hands and wrote the whole process off as unhelpful.”
Only one person typically feels neurologically rewarded by the PM exercise. It’s not the high performer, but the senior executive who oversees the ranking system. The feelings of status, certainty, and autonomy that occur when one is presiding over a forced ranking system are intrinsically rewarding. Even the act of categorizing information into groups, according to one study, activates the reward center of the brain. That’s why part of the education effort must include senior executives, who may not see the problem because the PM process reinforces their own cognitive reward while diminishing rewards for everyone else.
Rethinking Evaluation
An increasing number of organizations have been listening to their employees’ complaints and taking a more sophisticated approach to performance management. They are replacing year-end appraisals and ratings with in-depth conversations, often drawing on the myriad data points now available about employee and company performance, such as sales information, organizational climate survey results, and employee engagement data. A few firms have begun to experimentally shift away from the conventional PM approach. The companies that have joined this trend, either in pilots or full rollout, include Adobe, Cargill, ConAgra, Gap, Intel, Juniper Networks, Medtronic, and Sears. One noteworthy example is Microsoft, which revamped its entire approach in 2013. It now focuses evaluation on results that people deliver together, leveraging and contributing to one another, emphasizing continual learning and growth. The company completely retired traditional PM tactics, including ratings, distributions, and annual reviews.
Only one person feels neurologically rewarded by the PM exercise: the senior executive who oversees it.
How do these organizations transform their destructive performance ranking practices into a system that can develop talent consistently and pay people fairly? They do it by not throwing out the old approach entirely. In the CEB study, goal setting was shown to increase performance by 36 percent. Even feedback, which is often destructive to learning, can be designed in a productive way (see “The Problem with Feedback” ). The structure of bonuses and the calculation of salary can also be improved.
As firms make the courageous transition to a no-rating system, they tend to choose one of two options. The first is highly structured types of conversations regarding employee performance. The HR department might lay out five or six topics to discuss in an annual review, such as career growth, contribution, collaboration, or innovation. The companies also provide guidance on how to talk about each topic.
The second option is a guided conversation. Instead of topics, the company provides a general framework (and often some training for managers). The conversation focuses on the goals people set for themselves and how they are progressing toward those goals, along with their contribution, past and present, to the company.
In either a structured or a guided conversation, one key element is to prime people—both the employee and the boss—to induce a growth mind-set. This improves how people listen to feedback, encourages them to set stretch goals, makes it easier for them to put in extra effort toward a worthy project, and helps them learn from positive role models.
You can hear the difference between the fixed mind-set and the growth mind-set in the subtle nuances of conversation, and managers can make a difference in a company by deliberately choosing one kind of phrasing over another. For example, in giving feedback to employees, the phrase “You did well; you must be talented” activates a fixed mind-set. Talent is perceived to be innate and changeless. If the manager says instead, “You did well; you must have worked hard on this” or “I see you put everything you had into this,” a growth mind-set is activated in the employee. The effort and creativity that people bring to bear makes a difference.
The employee can also prime the boss. For example, saying “I want to be the top performer” primes the fixed mind-set. It implies there can be only one. By contrast, saying “I want to take on challenges where I can learn new things” primes the growth mind-set.
It’s worth the trouble to prime people for the growth mind-set. In one study, priming a group of managers that way consistently made them more confident in their abilities and more likely to follow the example of a positive role model. That could be because the growth mind-set allows people to follow others with no perceived loss in status. Conversely, the fixed mind-set automatically implies a zero-sum competition: If someone rises in status, everyone else must fall. People holding that mind-set are more likely to attack one another’s success instead of focusing on their own development.
Another intriguing trend in productive performance review conversations surprised us at first. Some firms that got rid of most numerical ratings have left one type of rating in place: the determination of whether someone is essentially “in or out” as a fit with the company’s culture. At Juniper, this is defined as being a “J-Player” or a “Non-J Player.” A J-Player is someone who generally behaves according to Juniper’s values and delivers reasonably good performance. Juniper clearly and consistently explains which types of behavior result in Non-J Player status and helps those employees fit in if they choose to stay. More than 80 percent of the people rated as Non-J Players have opted to leave the company; they understood why they would never succeed there.
The success of this “in or out” rating system seemed disturbing until we recognized why it was necessary. Executives were reluctant to remove rankings, but not for the reasons we thought they would be. They didn’t care much about identifying problem performers. The new system addressed that issue. They wanted to weed out people who did not fit with their culture, and who were thus holding back their departments and colleagues. We also saw another virtue: This simple system decoupled the question “Should you stay here?” from the question “How can you grow here?” This meant that among those designated as J-Players, the conversation could focus on the individual’s professional and personal growth, without the anxiety of wondering whether he or she would be ranked poorly.
A third trend we are seeing is the effort, at many companies, to reframe the entire PM process as something else. “An engineer in our Bangalore Excellence Center,” says Steven Rice of Juniper, “pointed out that our performance management process was a violation of our values, because the forced ratings didn’t enable leaders to authentically provide feedback or truly trust their judgment to administer rewards.” That led Rice to realize that they could not fix the system piecemeal; Juniper had to imagine a whole new kind of practice, one that “delivers the benefits without the unintended negative consequences.”
Since 2011, Juniper has not given ratings to employees or kept documents of ratings. It also eliminated forced rankings. The new method focuses heavily on regular quality conversations between managers and employees, using the structured conversation model. Overall, Juniper has seen participation and satisfaction skyrocket among employees and managers. On the semiannual day of evaluation (known as “Conversation Day”), more than 88 percent of the participants reported that their conversations were “helpful” or “very helpful.” “We have increased differentiation and alignment of rewards against relative contribution,” Rice says. “Eighty-seven percent of the employees reported that they were willing to give extra discretionary effort, and 79 percent believe they can do their best work at Juniper.”
At Gap, managers conduct monthly conversations in a new performance management process called “Grow, Perform, Succeed.” (Its abbreviation, GPS, is also the company’s stock symbol.) By redesigning its PM system and giving it a new name, the company repositioned the process as less of a threat. That in itself is an important step to better conversations.
Sharon Arad, a Cargill executive, describes how the company made this kind of shift after reviewing its PM system a few years ago: “We found the system failed to generate quality conversations, leaving employees with a [ranking] that many viewed as a deficiency statement. In the end, the ratings given were not a trustworthy indicator of the actual status of performance or engagement.”
Many Cargill leaders wondered whether removing the ratings would bring about more desirable results and better conversations. So they set up a no-rating pilot of several thousand employees for three years. Every year Arad’s team compared the pilot group’s feedback to that of a random sample of rated employees. “Overall, 90 percent of the no-rating pilot participants reported, year after year, that their experience was positive,” Arad says. This was in stark contrast to the feedback that people normally gave about their performance management experience. Finally, this year, Cargill adopted the no-rating approach for its entire organization.
The results of no-rating systems are dramatically better than their rating and ranking counterparts—in satisfaction, retention, and engagement scores, which have been shown to correlate to organizational performance. We believe the number of companies operating this way will increase dramatically, to become the majority. It would be nice to think that eventually the ideal of labeling people numerically will be seen as a blip in the history of HR. Giving people a rating might be a useful tool in a company with a lot of “fat,” where it makes sense to shake people up and create competition, but in many of today’s lean businesses that demand a great deal from their employees, we need a better model. We need to improve the quality of the conversations we hold with workers. It’s time to kill your performance ratings.
The Problem with Feedback
Human resources professionals have long assumed that most employees appreciate feedback: direct commentary about their performance. It is taken for granted that feedback motivates everyone to perform, and that top performers find it particularly useful. None of this is true. Although those with a growth mind-set react more positively to it than everyone else, typical feedback is not motivating, rewarding, or pleasant. According to one study, everyone hates it.
Psychologists at Kansas State, Eastern Kentucky, and Texas A&M universities asked 234 staffers at a large university how they felt about performance reviews they had received. The respondents were split up into three subgroups: those who said they were considered competent, those who avoided difficult tasks (to avoid negative feedback), and those who focused on developing and learning skills. The researchers expected that the individuals motivated to learn and develop skills were most likely to take negative feedback in stride. But it turned out that all three groups found negative feedback demoralizing, and all to the same extent.
Managers who conduct performance reviews typically do not take into account the way people feel about feedback. Even innocuous remarks that carry praise can easily be misconstrued as criticism. There’s a biological reason for this. Until they’re about 12 years old, human beings obtain most of the resources for survival from other people. As a result, humans have well-developed neural circuits for reading the social environment. Having this type of brain makes us highly attuned to social threats and rewards.
Most feedback primes people with a fixed mind-set—for example, cuing them with statements like “This is how you are.” That in itself makes people less open-minded. In another study, Jennifer Mangels, now director of the Baruch College Dynamic Learning Lab, and her colleagues monitored brain activity in two groups of people while the participants were asked some common-knowledge questions about history, literature, geography, and other academic topics. The first group were people with a fixed mind-set; they had said that they believed intelligence could not change. The second group, with a growth mind-set, had said they believed it could. In each case, after the questions, the groups received feedback—commentary about their performance on the quiz—and then they were given the correct answers.
Compared to the growth mind-set group, those with a fixed mind-set had more activity in a frontal part of the brain that is associated with responses to negative sensations. When they received the correct answers, they also had a shorter duration of the brain activity associated with long-term learning. Their attention was more likely to move elsewhere—perhaps to reacting to the feedback. When they were given a retest a short while later, the fixed mind-set group also showed less performance improvement, even though they started out the study performing equally well.
In short, there is reason to think that most performance-related feedback from other people, even when well intentioned, diminishes employees’ ability to learn. To compensate, it must be carefully designed to foster a growth mind-set.
Reprint No. 00275
Author profiles:
- David Rock is cofounder and director of the NeuroLeadership Institute, a global initiative bringing neuroscientists and leadership experts together. He is the author of Your Brain at Work: Strategies for Overcoming Distraction, Regaining Focus, and Working Smarter All Day Long (Harper Business, 2009). He is also the CEO of the NeuroLeadership Group, a global consulting firm.
- Josh Davis is an assistant professor of psychology at Barnard College of Columbia University, and director of research and lead professor for the NeuroLeadership Institute. His research deals with adaptive emotional state control and mind–body connections.
- Beth Jones is a senior researcher based in Ridgway, Colo., and a consultant with the NeuroLeadership Group.