At the Wall Street Journal last month, Wharton management professor Peter Cappelli challenged executives to think differently about their approach to employee performance and potential. The core of his argument is that these talent processes are less dynamic than the speed of the business. Business needs change, who you work with is different from project to project, expectations evolve, and personal circumstances or other factors can affect an employee’s potential contribution at any given time. Rarely, he argues, is an employee always an A, B, or C player:
Broad categories miss a lot of subtleties, so people’s true talents and strengths often get ignored. They also make it hard to recognize when people improve or stumble in their performance. The notion that good performers are always good also contributes to what psychologist Edward Thorndike called a halo effect, the mistaken belief that if you were an A player in one thing, you will be an A player in another. …
That means we assume A players perform better because they have more ability or talent than the others. But we don’t consider that, for instance, they might have gotten a string of easy projects where they could shine. Or maybe we think they’ve done a good job simply because we expected them to do a good job. Likewise, we assume that people performing poorly in their jobs are C players rather than people who are struggling with problems outside of work or have just been given an impossible assignment. Or again, maybe we judged them as performing worse than they actually did because we expected them to do poorly.
The cornerstone of his advice to solve this very real problem is to assess employees more frequently and objectively. He’s right about that, but this is just one of three things organizations need to do in order to improve their performance management strategies. They must also:
- Expand the number of people from whom managers receive feedback. Performance today is more collaborative, interconnected, matrixed and horizontal. In order to better determine the contributions that someone is making, managers need to ask more people.
- Make the feedback more forward looking. Rather than simply telling an employee how well they performed in the past, the best managers use past performance as a vehicle for talking about what they should be doing differently in the future. This approach lets managers use performance feedback as a development and evaluation tool.
Peter Cappelli suspects that it is. At HRE, the Wharton professor interprets the recent wave of organizations radically overhauling or even abolishing their performance review processes as a sign of a more fundamental transformation in the nature of this relationship:
By the 1990s, … [t]he growing interest in maximizing shareholder value and the increasing presence of financial thinking in the operation of big companies fueled the idea that employment is really an economic relationship: If you don’t give employees a financial incentive to work hard, they won’t do it. In that context, merit pay became the linchpin of the relationship with employees, and the purpose of performance appraisals was to determine merit pay. The goal was to measure past performance, document that measure and then hand out rewards based on it. …
Dropping performance appraisals — as so many leading companies seem to be doing now — cuts right at the heart of the idea that employment is an economic relationship and the notion that supervision is really just an accounting function: Total up the points and pass out the money. It’s taking us back to an earlier idea that employees actually need to be managed and that the supervisor should be helping them improve their performance and develop their skills. Merit pay still has its place, but that role is simply to encourage the main goal of improving job performance and skills.
Cappelli’s idea that the ongoing trend is not about performance appraisals per se, but rather the broader change in thinking that it represents, is an interesting way to look at the issue. To a certain extent, however, our research disagrees with the notion that “employees actually need to be managed and that the supervisor should be helping them improve their performance and develop their skills.” This describes a heavily top-down way of managing employees.
Instead, we would argue that professional relationships are indeed changing, but rather than reverting to a bygone era of performance management, the emerging management style is more collaborative, with performance and development becoming the shared responsibility of manager and employee. This change is also reflected in recent changes to performance management that focus on more frequent, informal, two-way conversations.
In the past year or two, many organizations have radically reformed their approach to performance reviews, in some cases ditching them altogether, and in many cases keeping the review but getting rid of performance scores or ratings, as GE became the latest major employer to do in July. On the other hand, our research at CEB has found that removing ratings tends to diminish the quality of the review process and doesn’t help performance; others have questioned this new trend as well. Looking at all the changes going on in this field (and drawing heavily on CEB research!), Knowledge@Wharton considers what’s working, what’s not, and what the future holds for performance management:
While the traditional annual performance review is surely dying, [Peter] Cappelli, who is also director of Wharton’s Center for Human Resources, has a wait-and-see attitude about whether employers will really create a different kind of relationship with employees, or end up doing less performance appraisal and replacing it with nothing instead.
“For a lot of companies that are thinking about this change, they are just copying what other companies are doing,” he says. “We will see a lot of false starts on this thing, and then they will discover their relationship with employees is worse off. The thing I would watch is to what extent this is an ideological battle. Is it all about the money, all about rewarding people — that [this is] how things get done, we have to punish the bad employee and fire them? Is it all about the economic incentive? Or is it much more about relationships and developing people and encouraging them to perform better? It’s an ideological divide that has to do with human nature. And to some extent that’s at the heart of this whole issue.” …
Good performance management looks backward and forward at the same time. Managers evaluate past performance to identify strengths and weaknesses, guide employees’ development, measure their individual contributions to the organization, and differentiate their rewards. Simultaneously, they must figure out how to drive performance to meet the challenges ahead of their team. At TD Magazine, Marcus Buckingham suggests breaking performance management into these two core components of assessment and acceleration and thinking about them as separate goals:
[I]t is incredibly inefficient to tackle assessment and acceleration together. Take the example of feedback. We think feedback is useful for accelerating performance. But even if it is necessary, what happens when that feedback seeps into the assessment realm? “Hey, everybody needs feedback. Let’s get even more of it. We’re going to aggregate that feedback from your bosses, and even your peers, over the course of the year, so that we can figure out what your performance was like and how to differentially invest in you.” Now we’ve taken feedback from the acceleration bucket and poured it into the assessment bucket. …
Rachel Emma Silverman broke the story at the Wall Street Journal this morning:
A new performance-management system asks employees and managers to exchange frequent feedback via a mobile app called PD@GE, in person or by phone. The messages are compiled into a performance summary at the end of the year. For GE, a longtime standard-bearer for corporate management, the shift reflects the realities of a new work climate in which employees expect more feedback from bosses and peers—companies, in turn, expect employees to act quickly on that feedback. …
Roughly 30,000 GE employees have tried rating-free reviews in the last couple of years. An internal study found that bosses could dole out pay and promotions effectively, and employees and managers preferred the new approach. At a meeting last month, about a dozen senior executives finally decided to dispense with the past practice. Scrapping ratings “led to more meaningful, richer conversations that were not getting distracted by…a label,” said Janice Semper, a GE human-resources executive. She adds that the changes apply to GE’s 200,000 salaried employees. Hourly workers may eventually be included if labor contracts allow.
GE executives had hinted at this last month, saying at the time that they were even thinking about doing away with annual raises. The decision to abandon performance ratings comes amid a trend of major employers from Goldman Sachs to IBM to the Pentagon shaking up their performance review systems to de-emphasize and simplify ratings. Other organizations have done away with ratings altogether, like Accenture, which set this trend in motion last year. As a gigantic employer with outsized influence in the corporate world, GE’s performance management systems have always invited copycats—longtime CEO Jack Welch’s “rank and yank” practice of firing the bottom 10 percent of performers was widely imitated—so this change is likely to resonate beyond the confines of GE itself.
However, GE’s move also comes just as we are finding that eliminating performance ratings isn’t necessarily a good idea.
In a short but insightful article (PDF), the Talent Strategy Group pushes back on the trendy skepticism of ratings-based performance management systems, observing that many of the “scientific” arguments leveled against ratings aren’t really all that scientific:
Science Claim #2: Ratings aren’t accurate, so don’t use them
In “Reinventing Performance Management”, Marcus Buckingham and Deloitte’s Ashley Goodall write about Deloitte’s former performance management system which, based on their descriptions (“creating the ratings consumed close to 2 million hours a year”), sounds ridiculously complex and bureaucratic. It’s understandable why they felt a redesign was necessary but the science they say inspired them to eliminate ratings offers no support for that choice.
In their article they cite, under the heading “The Science of Ratings,” research that they say shows that rater bias (anything in a rating not directly related to one’s actual performance) accounts for most of the differences in performance ratings. They state that the research says that “actual performance” accounts for only 21% of a performance rating. They wanted to redesign their performance management process to avoid this type of error.
The Reality: The article they cite, “The Latent Structure of Ratings,” in no way supports their argument. In fact, that research had nothing to do with actual performance ratings and didn’t involve a real company’s performance management process!
Permanent, Non-Senior Executive Service Employee Performance Rating Outcomes (All Rating Systems,
Calendar Year 2013) (GAO.gov)
Lisa Rein at the Washington Post examines a new report from the Government Accountability Office finding that the federal government has either a superhumanly high-performing workforce or—more likely—a broken performance management system:
In the ranks of the federal government, 99 percent are really good at their jobs — and almost two-thirds exceed expectations or do outstanding work. … [The report] also found that 78 percent of high-level civil servants — those in GS grades 13 through 15 — were given top performance scores of outstanding or fully successful. Senior executives were not covered in this data. The glowing picture of everyone in calendar year 2013, the most recent data available to auditors, is, on one level, good news for federal agencies. But it underscores a more likely reality to many in and outside of government. Rather than so many federal workers being exceptional, the system for rating them isn’t working right.