REPRINT R1610D PUBLISHED IN HBR OCTOBER 2016

ARTICLE SPOTLIGHT ON BUILDING THE WORKFORCE OF THE FUTURE The Performance Management Revolution The focus is shifting from accountability to learning. by Peter Cappelli and Anna Tavis

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

SPOTLIGHT

W hen Brian Jensen told his audience of HR executives that Colorcon wasn’t bother-ing with annual reviews anymore, they were appalled. This was in 2002, dur-ing his tenure as the drugmaker’s head of global human resources. In his presentation at the Wharton School, Jensen explained that Colorcon had found a more effec- tive way of reinforcing desired behaviors and managing perfor- mance: Supervisors were giving people instant feedback, tying it to individuals’ own goals, and handing out small weekly bonuses to employees they saw doing good things.

The focus is shifting from accountability to learning. BY PETER CAPPELLI AND ANNA TAVIS

The Performance Management Revolution

2  Harvard Business Review October 2016

SPOTLIGHT ON BUILDING THE WORKFORCE OF THE FUTURE

COPYRIGHT © 2016 HARVARD BUSINESS SCHOOL PUBLISHING CORPORATION. ALL RIGHTS RESERVED. This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact

customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

ARTWORK Ben Zank Going Nowhere, Untitled 8 Giclée on paper, 2015

October 2016 Harvard Business Review 3

FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR VISIT HBR.ORG

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Back then the idea of abandoning the tradi- tional appraisal process—and all that followed from it—seemed heretical. But now, by some estimates, more than one-third of U.S. companies are doing just that. From Silicon Valley to New York, and in offices across the world, firms are replacing annual reviews with frequent, informal check-ins between managers and employees.

As you might expect, technology companies such as Adobe, Juniper Systems, Dell, Microsoft, and IBM have led the way. Yet they’ve been joined by a number of professional services firms (Deloitte, Accenture, PwC), early adopters in other industries (Gap, Lear, OppenheimerFunds), and even General Electric, the longtime role model for traditional appraisals.

Without question, rethinking performance management is at the top of many executive teams’ agendas, but what drove the change in this direc- tion? Many factors. In a recent article for People + Strategy, a Deloitte manager referred to the review process as “an investment of 1.8 million hours across the firm that didn’t fit our business needs anymore.” One Washington Post business writer called it a “rite of corporate kabuki” that restricts creativity, gener- ates mountains of paperwork, and serves no real purpose. Others have described annual reviews as a last-century practice and blamed them for a lack of collaboration and innovation. Employers are also fi- nally acknowledging that both supervisors and sub- ordinates despise the appraisal process—a perennial problem that feels more urgent now that the labor market is picking up and concerns about retention have returned.

But the biggest limitation of annual reviews—and, we have observed, the main reason more and more companies are dropping them—is this: With their heavy emphasis on financial rewards and punish- ments and their end-of-year structure, they hold people accountable for past behavior at the expense of improving current performance and grooming tal- ent for the future, both of which are critical for or- ganizations’ long-term survival. In contrast, regular conversations about performance and development change the focus to building the workforce your or- ganization needs to be competitive both today and years from now. Business researcher Josh Bersin es- timates that about 70% of multinational companies are moving toward this model, even if they haven’t arrived quite yet.

The tension between the traditional and newer approaches stems from a long-running dispute about managing people: Do you “get what you get” when you hire your employees? Should you focus mainly on motivating the strong ones with money and getting rid of the weak ones? Or are employees malleable? Can you change the way they perform through effective coaching and management and in- trinsic rewards such as personal growth and a sense of progress on the job?

With traditional appraisals, the pendulum had swung too far toward the former, more transactional view of performance, which became hard to support in an era of low inflation and tiny merit-pay budgets. Those who still hold that view are railing against the recent emphasis on improvement and growth over accountability. But the new perspective is unlikely to be a flash in the pan because, as we will discuss, it is being driven by business needs, not imposed by HR.

First, though, let’s consider how we got to this point—and how companies are faring with new approaches.

How We Got Here Historical and economic context has played a large role in the evolution of performance management over the decades. When human capital was plenti- ful, the focus was on which people to let go, which to keep, and which to reward—and for those purposes, traditional appraisals (with their emphasis on indi- vidual accountability) worked pretty well. But when talent was in shorter supply, as it is now, developing people became a greater concern—and organizations had to find new ways of meeting that need.

From accountability to development. Appraisals can be traced back to the U.S. military’s

“merit rating” system, created during World War I to identify poor performers for discharge or transfer. After World War II, about 60% of U.S. companies were using them (by the 1960s, it was closer to 90%). Though seniority rules determined pay increases and promotions for unionized workers, strong merit scores meant good advancement prospects for man- agers. At least initially, improving performance was an afterthought.

And then a severe shortage of managerial tal- ent caused a shift in organizational priorities: Companies began using appraisals to develop em- ployees into supervisors, and especially managers into executives. In a famous 1957 HBR article, social

SPOTLIGHT ON BUILDING THE WORKFORCE OF THE FUTURE

4  Harvard Business Review October 2016 This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact

customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

psychologist Douglas McGregor argued that subor- dinates should, with feedback from the boss, help set their performance goals and assess themselves— a process that would build on their strengths and potential. This “Theory Y” approach to manage- ment—he coined the term later on—assumed that employees wanted to perform well and would do so if supported properly. (“Theory X” assumed you had to motivate people with material rewards and punishments.) McGregor noted one drawback to the approach he advocated: Doing it right would take managers several days per subordinate each year.

By the early 1960s, organizations had become so focused on developing future talent that many observers thought that tracking past performance had fallen by the wayside. Part of the problem was that supervisors were reluctant to distinguish good performers from bad. One study, for example, found that 98% of federal government employees received

“satisfactory” ratings, while only 2% got either of the other two outcomes: “unsatisfactory” or “outstand- ing.” After running a well-publicized experiment in 1964, General Electric concluded it was best to split the appraisal process into separate discussions about accountability and development, given the conflicts between them. Other companies followed suit.

Back to accountability. In the 1970s, however, a shift began. Inflation rates shot up, and merit- based pay took center stage in the appraisal process. During that period, annual wage increases really mattered. Supervisors often had discretion to give raises of 20% or more to strong performers, to dis- tinguish them from the sea of employees receiving basic cost-of-living raises, and getting no increase represented a substantial pay cut. With the stakes so high—and with antidiscrimination laws so re- cently on the books—the pressure was on to award pay more objectively. As a result, accountability

became a higher priority than development for many organizations.

Three other changes in the zeitgeist reinforced that shift:

First, Jack Welch became CEO of General Electric in 1981. To deal with the long-standing concern that supervisors failed to label real differences in per- formance, Welch championed the forced-ranking system—another military creation. Though the U.S. Army had devised it, just before entering World War II, to quickly identify a large number of officer candi- dates for the country’s imminent military expansion, GE used it to shed people at the bottom. Equating performance with individuals’ inherent capabilities (and largely ignoring their potential to grow), Welch divided his workforce into “A” players, who must be rewarded; “B” players, who should be accommo- dated; and “C” players, who should be dismissed. In that system, development was reserved for the “A” players—the high-potentials chosen to advance into senior positions.

Second, 1993 legislation limited the tax deduct- ibility of executive salaries to $1 million but ex- empted performance-based pay. That led to a rise in outcome-based bonuses for corporate leaders—a change that trickled down to frontline managers and even hourly employees—and organizations relied even more on the appraisal process to assess merit.

Third, McKinsey’s War for Talent research project in the late 1990s suggested that some employees were fundamentally more talented than others (you knew them when you saw them, the thinking went). Because such individuals were, by definition, in short supply, organizations felt they needed to take great care in tracking and rewarding them. Nothing in the McKinsey studies showed that fixed personal- ity traits actually made certain people perform better, but that was the assumption.

Idea in Brief THE PROBLEM By emphasizing individual accountability for past results, traditional appraisals give short shrift to improving current performance and developing talent for the future. That can hinder long-term competitiveness.

THE SOLUTION To better support employee development, many organizations are dropping or radically changing their annual review systems in favor of giving people less formal, more frequent feedback that follows the natural cycle of work.

THE OUTLOOK This shift isn’t just a fad—real business needs are driving it. Support at the top is critical, though. Some firms that have struggled to go entirely without ratings are trying a “third way”: assigning multiple ratings several times a year to encourage employees’ growth.

October 2016 Harvard Business Review 5

FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR VISIT HBR.ORG

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

told they were “average” than given a 3 on a 5-point scale. They especially detested forced ranking. As Wharton’s Iwan Barankay demonstrated in a field setting, performance actually declined when people were rated relative to others. Nor did the ratings seem accurate. As the accumulating research on appraisal scores showed, they had as much to do with who the rater was (people gave higher ratings to those who were like them) as they did with performance.

And managers hated doing reviews, as survey af- ter survey made clear. Willis Towers Watson found that 45% did not see value in the systems they used. Deloitte reported that 58% of HR executives consid- ered reviews an ineffective use of supervisors’ time. In a study by the advisory service CEB, the average manager reported spending about 210 hours—close to five weeks—doing appraisals each year.

As dissatisfaction with the traditional process mounted, high-tech firms ushered in a new way of thinking about performance. The “Agile Manifesto,” created by software developers in 2001, outlined several key values—favoring, for instance, “respond- ing to change over following a plan.” It emphasized principles such as collaboration, self- organization, self-direction, and regular reflection on how to work more effectively, with the aim of prototyping more quickly and responding in real time to customer feedback and changes in requirements. Although not directed at performance per se, these principles changed the definition of effectiveness on the job— and they were at odds with the usual practice of cas- cading goals from the top down and assessing people against them once a year.

So it makes sense that the first significant depar- ture from traditional reviews happened at Adobe, in 2011. The company was already using the agile method, breaking down projects into “sprints” that were immediately followed by debriefing sessions. Adobe explicitly brought this notion of constant as- sessment and feedback into performance manage- ment, with frequent check-ins replacing annual ap- praisals. Juniper Systems, Dell, and Microsoft were prominent followers.

CEB estimated in 2014 that 12% of U.S. compa- nies had dropped annual reviews altogether. Willis Towers Watson put the figure at 8% but added that 29% were considering eliminating them or planning to do so. Deloitte reported in 2015 that only 12% of the U.S. companies it surveyed were not planning to rethink their performance management systems.

So, by the early 2000s, organizations were us- ing performance appraisals mainly to hold employ- ees accountable and to allocate rewards. By some estimates, as many as one-third of U.S. corpora- tions—and 60% of the Fortune 500—had adopted a forced-ranking system. At the same time, other changes in corporate life made it harder for the ap- praisal process to advance the time-consuming goals of improving individual performance and develop- ing skills for future roles. Organizations got much flatter, which dramatically increased the number of subordinates that supervisors had to manage. The new norm was 15 to 25 direct reports (up from six be- fore the 1960s). While overseeing more employees, supervisors were also expected to be individual con- tributors. So taking days to manage the performance issues of each employee, as Douglas McGregor had advocated, was impossible. Meanwhile, greater in- terest in lateral hiring reduced the need for internal development. Up to two-thirds of corporate jobs were filled from outside, compared with about 10% a generation earlier.

Back to development…again. Another major turning point came in 2005: A few years after Jack Welch left GE, the company quietly backed away from forced ranking because it fostered internal competition and undermined collaboration. Welch still defends the practice, but what he really sup- ports is the general principle of letting people know how they are doing: “As a manager, you owe candor to your people,” he wrote in the Wall Street Journal in 2013. “They must not be guessing about what the or- ganization thinks of them.” It’s hard to argue against candor, of course. But more and more firms began questioning how useful it was to compare people with one another or even to rate them on a scale.

So the emphasis on accountability for past per- formance started to fade. That continued as jobs be- came more complex and rapidly changed shape—in that climate, it was difficult to set annual goals that would still be meaningful 12 months later. Plus, the move toward team-based work often conflicted with individual appraisals and rewards. And low in- flation and small budgets for wage increases made appraisal-driven merit pay seem futile. What was the point of trying to draw performance distinctions when rewards were so trivial?

The whole appraisal process was loathed by em- ployees anyway. Social science research showed that they hated numerical scores—they would rather be

6  Harvard Business Review October 2016

SPOTLIGHT ON BUILDING THE WORKFORCE OF THE FUTURE

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

better job of coaching and allows subordinates to process and apply the advice more effectively.

Kelly Services was the first big professional ser- vices firm to drop appraisals, in 2011. PwC tried it with a pilot group in 2013 and then discontinued annual reviews for all 200,000-plus employees. Deloitte followed in 2015, and Accenture and KPMG made similar announcements shortly thereafter. Given the sheer size of these firms, and the fact that they offer management advice to thousands of organizations, their choices are having an enor- mous impact on other companies. Firms that scrap appraisals are also rethinking employee manage- ment much more broadly. Accenture CEO Pierre Nanterme estimates that his firm is changing about 90% of its talent practices.

The need for agility. When rapid innovation is a source of competitive advantage, as it is now in many companies and industries, that means future needs are continually changing. Because organiza- tions won’t necessarily want employees to keep do- ing the same things, it doesn’t make sense to hang on to a system that’s built mainly to assess and hold people accountable for past or current practices. As Susan Peters, GE’s head of human resources, has pointed out, businesses no longer have clear annual cycles. Projects are short-term and tend to change along the way, so employees’ goals and tasks can’t be plotted out a year in advance with much accuracy.

At GE a new business strategy based on innova- tion was the biggest reason the company recently began eliminating individual ratings and annual

This trend seems to be extending beyond the United States as well. PwC reports that two-thirds of large companies in the UK, for example, are in the process of changing their systems.

Three Business Reasons to Drop Appraisals In light of that history, we see three clear business imperatives that are leading companies to abandon performance appraisals:

The return of people development. Companies are under competitive pressure to upgrade their tal- ent management efforts. This is especially true at consulting and other professional services firms, where knowledge work is the offering—and where inexperienced college grads are turned into skilled advisers through structured training. Such firms are doubling down on development, often by putting their employees (who are deeply motivated by the potential for learning and advancement) in charge of their own growth. This approach requires rich feed- back from supervisors—a need that’s better met by frequent, informal check-ins than by annual reviews.

Now that the labor market has tightened and keeping good people is once again critical, such companies have been trying to eliminate “dissatis- fiers” that drive employees away. Naturally, annual reviews are on that list, since the process is so widely reviled and the focus on numerical ratings interferes with the learning that people want and need to do. Replacing this system with feedback that’s delivered right after client engagements helps managers do a

CAN YOU TAKE COGNITIVE BIAS OUT OF ASSESSMENTS? A classic study by Edward Jones and Victor Harris in the 1960s demonstrated that people tend to attribute others’ behavior to character rather than circumstances. When a car goes streaking past us, for instance, we think that the driver is a jerk and ignore the possibility that there might be an emergency. A good workplace example of this cognitive bias—known as the “fundamental attribution error”—is to assume that the lowest performers in any year will always be the worst performers and to fire them as a result. Such an assumption overlooks the impact of good or poor management, not to mention

business conditions that are beyond employees’ control.

Of course, this model is highly flattering to people who have advanced into executive roles—“A” players whose success is, by definition, credited to their superior abilities, not to good fortune. That may be partly why the model has persisted so long in the face of considerable evidence against it.

Even when “A” players seem to perform well in many contexts (and that’s rarely measured), they may be coasting on the

“halo effect”—another type of bias, akin to self-fulfilling prophecy. If these folks have already been successful, they receive more opportunities than others, and they’re pushed harder, so naturally they do better.

Biases color individual performance ratings as well. Decision makers may

give past behavior too much weight, for instance, or fall prey to stereotypes when they assign their ratings.

But when you get rid of forced ranking and appraisal scores, you don’t eradicate bias. Discrimination and faulty assumptions still creep into qualitative assessments. In some ways the older, more cumbersome performance systems actually made it harder for managers to keep their blinders on. Formal feedback from various stakeholders provided some balance when supervisors were otherwise inclined to see only the good things their stars did and failed to recognize others’ contributions.

Anytime you exercise judgment, whether or not you translate that to numerical ratings, intuition plays a part, and bias can rear its head.

October 2016 Harvard Business Review 7

FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR VISIT HBR.ORG

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

of store-level goals. Nonetheless, Rob Ollander- Krane, Gap’s senior director of organization per- formance effectiveness, says the company needs further improvement in setting stretch goals and focusing on team performance.

Implications. All three reasons for dropping an- nual appraisals argue for a system that more closely follows the natural cycle of work. Ideally, conversa- tions between managers and employees occur when projects finish, milestones are reached, challenges pop up, and so forth—allowing people to solve prob- lems in current performance while also developing skills for the future. At most companies, managers take the lead in setting near-term goals, and employ- ees drive career conversations throughout the year. In the words of one Deloitte manager: “The conver- sations are more holistic. They’re about goals and strengths, not just about past performance.”

Perhaps most important, companies are over- hauling performance management because their businesses require the change. That’s true whether they’re professional services firms that must de- velop people in order to compete, companies that need to deliver ongoing performance feedback to support rapid innovation, or retailers that need bet- ter coordination between the sales floor and the back office to serve their customers.

Of course, many HR managers worry: If we can’t get supervisors to have good conversations with sub- ordinates once a year, how can we expect them to do so more frequently, without the support of the usual appraisal process? It’s a valid question—but we see reasons to be optimistic.

reviews. Its new approach to performance manage- ment is aligned with its FastWorks platform for cre- ating products and bringing them to market, which borrows a lot from agile techniques. Supervisors still have an end-of-year summary discussion with sub- ordinates, but the goal is to push frequent conversa- tions with employees (GE calls them “touchpoints”) and keep revisiting two basic questions: What am I doing that I should keep doing? And what am I doing that I should change? Annual goals have been replaced with shorter-term “priorities.” As with many of the companies we see, GE first launched a pilot, with about 87,000 employees in 2015, before adopting the changes across the company.

The centrality of teamwork. Moving away from forced ranking and from appraisals’ focus on individual accountability makes it easier to fos- ter teamwork. This has become especially clear at retail companies like Sears and Gap—perhaps the most surprising early innovators in appraisals. Sophisticated customer service now requires front- line and back-office employees to work together to keep shelves stocked and manage customer flow, and traditional systems don’t enhance performance at the team level or help track collaboration.

Gap supervisors still give workers end-of-year as- sessments, but only to summarize performance dis- cussions that happen throughout the year and to set pay increases accordingly. Employees still have goals, but as at other companies, the goals are short-term (in this case, quarterly). Now two years into its new system, Gap reports far more satisfaction with its performance process and the best-ever completion

A TALENT MANAGEMENT TIMELINE The tug-of-war between accountability and development over the decades

■ Accountability focus ■ Development focus ■ A hybrid “third way”

WWI The U.S. military created merit- rating system to flag and dismiss poor performers.

1940s About 60% of U.S. companies were using appraisals to document workers’ performance and allocate rewards.

1970s Inflation rates shot up, and organizations felt pressure to award merit pay more objectively, so accountability again became the priority in the appraisal process.

WWII The Army devised forced ranking to identify enlisted soldiers with the potential to become officers.

1950s Social psychologist Douglas McGregor argued for engaging employees in assessments and goal setting.

1960s Led by General Electric, companies began splitting appraisals into separate discussions about accountability and growth, to give development its due.

8  Harvard Business Review October 2016

SPOTLIGHT ON BUILDING THE WORKFORCE OF THE FUTURE

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

says the company has been tracking the effects on business results, and they’ve been positive so far.

Challenges That Persist The greatest resistance to abandoning apprais- als, which is something of a revolution in human resources, comes from HR itself. The reason is simple: Many of the processes and systems that HR has built over the years revolve around those per- formance ratings. Experts in employment law had advised organizations to standardize practices, de- velop objective criteria to justify every employment decision, and document all relevant facts. Taking away appraisals flies in the face of that advice—and it doesn’t necessarily solve every problem that they failed to address.

Here are some of the challenges that organiza- tions still grapple with when they replace the old performance model with new approaches:

Aligning individual and company goals. In the traditional model, business objectives and strat- egies cascaded down the organization. All the units, and then all the individual employees, were sup- posed to establish their goals to reflect and reinforce the direction set at the top. But this approach works only when business goals are easy to articulate and held constant over the course of a year. As we’ve discussed, that’s often not the case these days, and employee goals may be pegged to specific projects. So as projects unfold and tasks change, how do you coordinate individual priorities with the goals for the whole enterprise, especially when the business objectives are short-term and must rapidly adapt to

As GE found in 1964 and as research has docu- mented since, it is extraordinarily difficult to have a serious, open discussion about problems while also dishing out consequences such as low merit pay. The end-of-year review was also an excuse for delaying feedback until then, at which point both the supervi- sor and the employee were likely to have forgotten what had happened months earlier. Both of those constraints disappear when you take away the an- nual review. Additionally, almost all companies that have dropped traditional appraisals have invested in training supervisors to talk more about development with their employees—and they are checking with subordinates to make sure that’s happening.

Moving to an informal system requires a culture that will keep the continuous feedback going. As Megan Taylor, Adobe’s director of business part- nering, pointed out at a recent conference, it’s dif- ficult to sustain that if it’s not happening organically. Adobe, which has gone totally numberless but still gives merit increases based on informal assessments, reports that regular conversations between manag- ers and their employees are now occurring without HR’s prompting. Deloitte, too, has found that its new model of frequent, informal check-ins has led to more meaningful discussions, deeper insights, and greater employee satisfaction. (For more details, see

“Reinventing Performance Management,” HBR, April 2015.) The firm started to go numberless like Adobe but then switched to assigning employees several numbers four times a year, to give them rolling feed- back on different dimensions. Jeffrey Orlando, who heads up development and performance at Deloitte,

1980s Jack Welch championed forced ranking at GE to reward top performers, accommodate those in the middle, and get rid of those at the bottom.

2000 Organizations got flatter, which dramatically increased the number of direct reports each manager had, making it harder to invest time in developing them.

2016 Deloitte, PwC, and others that tried going numberless are reinstating performance ratings but using more than one number and keeping the new emphasis on developmental feedback.

1990s McKinsey’s War for Talent study pointed to a shortage of capable executives and reinforced the emphasis on assessing and rewarding performance.

2011 Kelly Services was the first big professional services firm to drop appraisals, and other major firms followed suit, emphasizing frequent, informal feedback.

Adobe ended annual performance reviews, in keeping with the famous “Agile Manifesto” and the notion that annual targets were irrelevant to the way its business operated.

October 2016 Harvard Business Review 9

FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR VISIT HBR.ORG

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

Still, given how reluctant most managers are to single out failing employees, we can’t assume that getting rid of appraisals will make those tough calls any easier. And all the companies we’ve observed still have “performance improvement plans” for employees identified as needing support. Such plans remain universally problematic, too, partly because many issues that cause poor performance can’t be solved by management intervention.

Avoiding legal troubles. Employee relations managers within HR often worry that discrimina- tion charges will spike if their companies stop bas- ing pay increases and promotions on numerical rat- ings, which seem objective. But appraisals haven’t prevented discriminatory practices. Though they force managers to systematically review people’s contributions each year, a great deal of discretion (always subject to bias) is built into the process, and considerable evidence shows that supervisors dis- criminate against some employees by giving them undeservedly low ratings.

Leaders at Gap report that their new practices were driven partly by complaints and research showing that the appraisal process was often biased and ineffective. Frontline workers in retail (dispro- portionately women and minorities) are especially vulnerable to unfair treatment. Indeed, formal rat- ings may do more to reveal bias than to curb it. If a company has clear appraisal scores and merit-pay indexes, it is easy to see if women and minorities with the same scores as white men are getting fewer or lower pay increases.

All that said, it’s not clear that new approaches to performance management will do much to miti- gate discrimination either. (See the sidebar “Can You Take Cognitive Bias Out of Assessments?”) Gap has found that getting rid of performance scores in- creased fairness in pay and other decisions, but judg- ments still have to be made—and there’s the possibil- ity of bias in every piece of qualitative information that decision makers consider.

Managing the feedback firehose. In recent years most HR information systems were built to move annual appraisals online and connect them to pay increases, succession planning, and so forth. They weren’t designed to accommodate continuous feedback, which is one reason many employee check- ins consist of oral comments, with no documentation.

The tech world has responded with apps that enable supervisors to give feedback anytime and to

market shifts? It’s a new kind of problem to solve, and the jury is still out on how to respond.

Rewarding performance. Appraisals gave managers a clear-cut way of tying rewards to indi- vidual contributions. Companies changing their sys- tems are trying to figure out how their new practices will affect the pay-for-performance model, which none of them have explicitly abandoned.

They still differentiate rewards, usually relying on managers’ qualitative judgments rather than numerical ratings. In pilot programs at Juniper Systems and Cargill, supervisors had no difficulty allocating merit-based pay without appraisal scores. In fact, both line managers and HR staff felt that paying closer attention to employee performance throughout the year was likely to make their merit- pay decisions more valid.

But it will be interesting to see whether most supervisors end up reviewing the feedback they’ve given each employee over the year before determin- ing merit increases. (Deloitte’s managers already do this.) If so, might they produce something like an annual appraisal score—even though it’s more carefully considered? And could that subtly under- mine development by shifting managers’ focus back to accountability?

Identifying poor performers. Though manag- ers may assume they need appraisals to determine which employees aren’t doing their jobs well, the traditional process doesn’t really help much with that. For starters, individuals’ ratings jump around over time. Research shows that last year’s perfor- mance score predicts only one-third of the variance in this year’s score—so it’s hard to say that someone simply isn’t up to scratch. Plus, HR departments consistently complain that line managers don’t use the appraisal process to document poor performers. Even when they do, waiting until the end of the year to flag struggling employees allows failure to go on for too long without intervention.

We’ve observed that companies that have dropped appraisals are requiring supervisors to immediately identify problem employees. Juniper Systems also formally asks supervisors each quarter to confirm that their subordinates are performing up to company standards. Only 3%, on average, are not, and HR is brought in to address them. Adobe reports that its new system has reduced dismissals, because struggling employees are monitored and coached much more closely.

10  Harvard Business Review October 2016

SPOTLIGHT ON BUILDING THE WORKFORCE OF THE FUTURE

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.

formal feedback is essential to accountability. And Medtronic, which gave up ratings several years ago, is resurrecting them now that it has acquired Ireland- based Covidien, which has a more traditional view of performance management.

Other firms aren’t completely reverting to old approaches but instead seem to be seeking middle ground. As we’ve mentioned, Deloitte has backped- aled from giving no ratings at all to having project leads and managers assign them in four categories on a quarterly basis, to provide detailed “perfor- mance snapshots.” PwC recently made a similar move in its client-services practices: Employees still don’t receive a single rating each year, but they now get scores on five competencies, along with other development feedback. In PwC’s case, the push- back against going numberless actually came from employees, especially those on a partner track, who wanted to know how they were doing.

At one insurance company, after formal ratings had been eliminated, merit-pay increases were be- ing shared internally and then interpreted as per- formance scores. These became known as “shadow ratings,” and because they started to affect other talent management decisions, the company eventu- ally went back to formal appraisals. But it kept other changes it had made to its performance management system, such as quarterly conversations between managers and employees, to maintain its new com- mitment to development.

It will be interesting to see how well these “third way” approaches work. They, too, could fail if they aren’t supported by senior leadership and rein- forced by organizational culture. Still, in most cases, sticking with old systems seems like a bad option. Companies that don’t think an overhaul makes sense for them should at least carefully consider whether their process is giving them what they need to solve current performance problems and develop future talent. Performance appraisals wouldn’t be the least popular practice in business, as they’re widely believed to be, if something weren’t fundamentally wrong with them. HBR Reprint R1610D

record it if desired. At General Electric, the PD@GE app (“PD” stands for “performance development”) allows managers to call up notes and materials from prior conversations and summarize that informa- tion. Employees can use the app to ask for direction when they need it. IBM has a similar app that adds another feature: It enables employees to give feed- back to peers and choose whether the recipient’s boss gets a copy. Amazon’s Anytime Feedback tool does much the same thing. The great advantage of these apps is that supervisors can easily review all the discussion text when it is time to take actions such as award merit pay or consider promotions and job reassignments.

Of course, being on the receiving end of all that continual coaching could get overwhelming—it never lets up. And as for peer feedback, it isn’t al- ways useful, even if apps make it easier to deliver in real time. Typically, it’s less objective than supervi- sor feedback, as anyone familiar with 360s knows. It can be also “gamed” by employees to help or hurt colleagues. (At Amazon, the cutthroat culture en- courages employees to be critical of one another’s performance, and forced ranking creates an incen- tive to push others to the bottom of the heap.) The more consequential the peer feedback, the more likely the problems.

NOT ALL EMPLOYERS face the same business pres- sures to change their performance processes. In some fields and industries (think sales and financial services), it still makes sense to emphasize account- ability and financial rewards for individual perform- ers. Organizations with a strong public mission may also be well served by traditional appraisals. But even government organizations like NASA and the FBI are rethinking their approach, having concluded that accountability should be collective and that supervisors need to do a better job of coaching and developing their subordinates.

Ideology at the top matters. Consider what hap- pened at Intel. In a two-year pilot, employees got feedback but no formal appraisal scores. Though supervisors did not have difficulty differentiating performance or distributing performance-based pay without the ratings, company executives returned to using them, believing they created healthy com- petition and clear outcomes. At Sun Communities, a manufactured-home company, senior leaders also oppose eliminating appraisals because they think

Peter Cappelli is a professor of management at the Wharton School and the author of several

books, including Will College Pay Off? A Guide to the Most Important Financial Decision You’ll Ever Make (PublicAffairs, 2015). Anna Tavis is the academic director of Columbia’s program in human capital management and the Perspectives editor at People + Strategy, a journal for HR executives.

October 2016 Harvard Business Review 11

FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR VISIT HBR.ORG

This document is authorized for use only by Amy Hauenstein (hauenstein@northwestern.edu). Copying or posting is an infringement of copyright. Please contact customerservice@harvardbusiness.org or 800-988-0886 for additional copies.