Stacked Ranking: A Pay For Performance Model You Need to Avoid

Ann Bares recently wrote a predictive article here on TLNT about the potential end of merit pay (How Will We Pay With Open Salaries and No Performance Reviews?).

In her post, Ann argues that because “open salaries” and “blowing up performance appraisals” are becoming more popular, merit pay cannot be long for the world. She ends by asking:

What will we do instead? Strictly market-based wages with “hot skill” premiums as appropriate? More emphasis on variable pay plans designed to reward specific, pre-determined individual or group metrics? Will recognition and non-cash rewards step into the void to provide the necessary differentiation for key talent?”

Is “Meets Expectations” good enough?

I sincerely hope she’s right. Any of those outcomes are better than many of the constructs at play today, and certainly far better than the most egregious pay for performance models.

I’m looking at you, stack ranking. Consider this Fistful of Talent article on a major government contractor taking this to the extreme:

A major government contractor in Washington, DC is doubling down on the bell curve. But wait a minute; this is not just your typical bell curve, but a new extreme level for assigning managers’ performance quotas.

The new rating system looks like this:

  1. Exceeds expectations: 15 percent;
  2. Meets expectations: 45 percent;
  3. Partially meets expectations: 30 percent;
  4. Does not meet expectations: 10 percent.”

The article goes on to describe why this is so bad. Don’t we want all employees striving to perform better? The natural (and observed) outcome of this, however, is the rapid egress of truly top talent who failed to make the 15 percent Level 1 cut.

The common argument is that “Meets expectations” is perfectly fine and a truly desirable state for such a large chunk of employees — but not when you know you perform well and deliver the goods on a regular basis. In that case, “Meets expectations” is a slap in the face and a clear signal that it’s time to take your talents elsewhere.

What does forced ranking look like in reality?

A member of my team related this story to me last week of her husband’s performance review (I’ll call him Jim). Jim filled out the self appraisal form, only to have his words returned to him verbatim in his manager’s “appraisal” of his performance.

But that’s not the real story here. This Fortune 100 firm conducted several rounds of layoffs in the last year, resulting in Jim’s team being diminished by about half, with only the true stars remaining.

Jim was informed by his manager that, since the team was so much smaller now, there could only be one (1) “exceeds expectations” employee on the team. All the rest would be marked “meets expectations” because senior management was also not allowing any “needs improvement” out of fear even more employees would walk out. Even worse, the lack of managerial skills on this team results in each employee simply keeping his or her same ranking designation year after year.

Article Continues Below

So, despite the fact that every member of the remaining team is a true star in the organization, each pulling double the work after the rounds of layoffs, only one could be marked as exceeding expectations for the year.

Stack Ranking drives fear, reduces productivity

Another fall-out of such a strict system can lead a significant proportion of employees to believe their jobs might be at risk. Research reported in Quartz shows the fear induced by this scenario leads to even more unintended consequences:

Studies find that workers who fear being laid off are less safety-conscious, more likely to get injured, and less likely to report injuries … Recent research underlines the other flaw with forced ranking: its basic assumption that pressure will boost productivity. Making people insecure may make them work harder to try to be on the right side of the bell curve, but that will come at the cost of health, happiness, and future performance. Unless companies are confident that they can constantly replace the people who burn out with new employees that are just as good, it looks like a losing formula.”

I certainly don’t need to tell readers of the high costs associated this amount of turnover, much less the costs of lost productivity and safety challenges. But stack ranking hangs on.

My theory as to why is because it’s the easy way out. Managers operating under this system can more easily allocate limited merit pay budgets and blame “the system” for the outcome.

How does your organization measure performance? Do you see the end of merit pay on the horizon? What do you think is the best long-term solution?

This was originally published at the Compensation Café blog, where you can find a daily dose of caffeinated conversation on everything compensation.

Derek Irvine is one of the world’s foremost experts on employee recognition and engagement, helping business leaders set a higher vision and ambition for their company culture. As the Vice President of Client Strategy and Consulting at Globoforce, Derek helps clients — including some of world’s most admired companies such as Proctor and Gamble, Intuit, KPMG, and Thomson Reuters — leverage recognition strategies and best practices to better manage company culture, elevate employee engagement, increase retention, and improve the bottom line. He's also a renowned speaker and co-author of Winning with a Culture of Recognition. Contact him at


5 Comments on “Stacked Ranking: A Pay For Performance Model You Need to Avoid

  1. I agree strongly with Derek. The HR community should be pushing back and telling senior management its counterproductive. It is not legally defensible and has no basis in our understaniang of human performance or in proven performance management practices. I am well aware ratings are often badly inflated but there are better answers.

    1. I absolutely agree with Derek as well! I work in an organization where we use this ridiculous bell curve with the forced percentages and I am definitely seeing the consequences of the ranking. Employees are less inclined to perform better because they feel their efforts are not reflected correclty during their evaluation, supervisors are not managing “out” the poor performers because they would rather wait until evaluation time so that they can go ahead and place those poor performers at the bottom as opposed to having to place their good employees at the bottom because they actually did their job and now have a great team. Then to make matters worse, HR is tasked to put out fires because when employees don’t understand why they were ranked low. their first place to turn to is HR! Overall, everyone feels disengaged for a system that no one sees any value in and is beyond archaic.

  2. I think we need to be more clear about distinguishing between stack ranking – indicating differentiated levels of performance, and FORCED ranking, where pre-determined percentages are established and managers are required to fit their people into those quotas.
    The reality is that people do contribute at different levels and do meet, exceed or even fail to meet the expectations of the role. We need to be able to formally acknowledge those different levels of performance and results delivery.
    Where the rubber should be meeting the road is in how we properly define the expectations of the role and in how we evaluate and communicate what that looks like.
    Maybe I am “old school (hint – I am. I believe in a strong work ethic, going above and beyond, making things better than you found them, etc., but I digress), but the concept of stack ranking is valid. The system has merit. The failure lies in the execution by managers and in how senior leaders value assessing the performance of the people in their organizations.

  3. Understanding who your best employees are across the organization is a valuable exercise, and stack ranking is one way to facilitate that discussion. I’ve had functional heads rank all their employees, and then get together with each other and debate where everybody fits in the company. This process, over time, gives visibility across departments who the real stars are, and where development action seems warranted. The goal of the process is NOT to allocate who gets how much of the salary budget. Rather, its designed to develop a shared view of what good performance looks like, and produces, over time, more consistent performance metrics throughout the organization.

    Using stack ranking for merit allocation is a bad practice, and destroys any credibility associated with the salary adjustment process. It’s an artificial exercise at best, and demotivates employees, even those at the top. Managers will go transparent and blame the system, which only reinforces its artificiality. Bell curves may work in theory, but fail miserably in practice.

  4. In general, I don’t like strict stacked or “forced” ranking although there are certain times when it can add value. The following is one of the few actual empirical studies I’ve seen examining the impact of forced ranking on organizational performance and identifying those situations when it does seem to work, as well as the many conditions when it hurts performance:
    S. E., Bergey, P. K., & Aiman-Smith, L. (2005). Forced distribution
    rating systems and the improvement of workforce potential. Personnel Psychology, 58, 1-32.

    I believe the main problem with forced ranking is it is over-used by executives who don’t want to do the hard work of actually defining what they mean by “high performance” or spending any time discussing this with their managers and employees. In essence, they want the HR process to force certain results without having the conversations needed to ensure these results are accurate.

    It is also important to distinguish forced ranking from calibration sessions that use what I called “justified rating”. In these sessions managers can rate employees however they want, but they must justify their ratings to their peers. So a manager can have a team consisting entirely of high performers, but they must be able to justify this to others. What I’ve found is managers tend to be far more realistic in their performance evaluations when they know they will have to share it with others. Also, somewhat counter intuitively, the calibration sessions I’ve been in spend far more time discussing who the company’s high performers are and relatively little time debating about who falls into the under-performance area. If done well, managers usually respond very favorably to these sessions as they help them create and apply a consistent definition of high performance across the company.

Leave a Comment

Your email address will not be published. Required fields are marked *