The trouble with stacked rankings
“Stacked ranking,” also known as “forced distribution,” was once the standard way to evaluate employee performance. It was formerly used by major brands like Accenture, IBM, Adobe, Deloitte and Microsoft. Despite the move away from this system, stacked ranking is still a go-to performance management approach for many companies that are unsure of how to evolve their performance management processes and systems.
Originally conceived as a way to force managers to make tough choices and create competitive, fast-paced environments, stack ranking mandates assigning a certain percentage of employees into predetermined performance levels — typically 1 through 5, with 5 denoting “exceeds expectations” and 1 rated as “does not meet expectations.” This style of ranking was popularized by GE under esteemed former CEO and management guru Jack Welch. Now GE has jettisoned annual performance reviews altogether, opting for a performance app instead, in response to a changing workforce as millennials came on board in greater numbers.
Stacked ranking is designed to help businesses allocate annual raises and promotions. It also, in the same vein, makes it easier to terminate employees for poor performance. Unfortunately, stack ranking has become a black box of stress for employers and employees.
Why stack rankings don’t work
While employees may not understand the exact algorithm behind their company’s stack ranking system, they will likely remember this concept from their days as a student when they were “graded on a curve.” This means that the actual breakdown of results has to ensure that a certain number of employees are below average and a certain number are above average.
Since this system requires managers to assign their subordinates into buckets using a predetermined formula, every manager is forced to put a certain percentage of their workers in a bottom bucket. Those employees are put on performance-improvement programs and “managed out” of the company. While the few employees at the high end of the curve may be motivated to continue coming out on top, most employees are demotivated by receiving a 3 — meets expectations — on a scale of 1 to 5. Demotivating the majority of your workforce is clearly not a desired outcome of a fundamental performance management process.
Additionally, in the case of learning curves in school, stack ranking makes sense because students share the same learning objectives, materials and test questions. But that’s precisely where the problem with stack ranking lies in the workplace: Even on a team with the same objectives, people play different roles and perform a range of tasks. Forced distribution evaluations don’t recognize these variations since they put everyone in the same distribution stack. And the technique takes away manager discretion — by design — replacing it with a rigid, anti-progressive method.
Worse yet, stacked ranking systems bring out undesirable behaviors in employees: “The system was meant to encourage hard work and weed out underperformers, but it soon produced the exact opposite,” wrote Nicholas Carlson in The New York Times in late 2014. “Because only so many 4s and 5s could be allotted, talented people no longer wanted to work together; strategic goals were sacrificed, as employees did not want to change projects and leave themselves open to a lower score.”
In the last year, large firms, including Amazon and Yahoo, have come under fire for their stacked ranking systems. Both companies have been subjected to criticism for systems that are said to require managers to place a certain percentage of their employees in the ominous bottom box of performers, even if all team members are meeting or exceeding expectations.
A better approach
If stacked ranking systems don’t work, what’s a better approach? We recommend empowering the manager to determine who on their team should be rewarded and who needs to be placed on a performance improvement plan. There should never be a specific number of employees who have to be placed in the lower bucket, or the top bucket, for that matter.
Giving managers control of raises, promotions and terminations at scale may sound like a difficult feat, but many corporations are moving away from stacked ranking systems and toward modern, real-time feedback solutions that are focused on manager-centric performance management. Ratings on their own aren’t the enemy of good management, but when companies empower managers to spend more time developing employees with real-time coaching, managers are more in tune with performance. And that means when it’s time to determine raises and promotions, they are already experts on the topic. Data-driven reviews and check-ins help tell a story over time of individual and team performance.
In order to create a manager-centric performance system, companies first must ensure that managers have the tools needed to track employee contributions and provide constructive criticism throughout the year as well as note opportunities for improvement. When raises are provided, whether that’s annually or at more regular frequency, these are determined by the manager based on their own data collected throughout the period since the last raise opportunity. With a system that integrates into business applications like Gmail, Outlook and Slack, managers can tag communications for immediate feedback and recognition and/or for use during frequent one-to-one discussions, weekly meetings and monthly/quarterly check-ins.
Stacked rankings continue to fall out of favor across the best companies in America. Surveys from the Arlington, Va.-based consultancy CEB show that only 27% of companies in the Fortune 1000 in 2015 reported using a forced ranking to measure the performance of some part of their workforce. That's down from 44% in 2013.
However, in some exceptional situations, companies still do find it necessary to use stacked rankings. These scenarios often include difficult times where decisions on mass-scale corporate layoffs must be made via some quantitative measure that can be accessed across an entire workforce.
Ed Lawler, a professor at the University of Southern California's Marshall School of Business, noted to The Washington Post that Yahoo's current situation could be one of those scenarios. “The most defensible place to use a forced distribution is in a turnaround, when you desperately need to make some tough decisions and don’t trust managers to make them,” he said.
However, in most cases where a company is performing well and not facing a massive restructuring, stacked ranking systems often cause more harm than good. With a real-time feedback strategy, executives and managers can use both quantitative and qualitative data to fairly calibrate the systems used to determine compensation and promotions. And with this approach, they can harness the immediacy of real-time feedback to remove the barriers that clunky, outdated annual review processes and stacked ranking systems raise against agile performance development. More companies are learning this lesson in modern management every day.