Stack Ranking is a Problem
- Stephen Dawkins
- Apr 3
- 4 min read
During my time as a manager at Amazon, I witnessed firsthand the harm caused by stack ranking—an employee evaluation system that forces managers to rank team members against each other. While it may sound like a meritocratic way to reward performance, in practice it fosters fear, discourages collaboration, and damages morale.
Incentivized to Avoid Hiring the Best
At Amazon, stack ranking created a perverse incentive: as a manager, I found myself hesitant to always hire top-tier talent. That sounds counterintuitive—why wouldn't you want the best people on your team? The problem was simple, stack ranking required every team to have a bottom performer, regardless of overall team quality.
Bringing in exceptional talent raised the bar across the team, which might sound great in theory, but in practice meant that someone who was already performing well might be pushed into the "least effective" category. That could mean damaging someone's career or even putting their job at risk. For employees on work visas, this was an even greater concern as losing their job due to an arbitrary ranking could mean losing their ability to stay in the country. The stakes were incredibly high, and it forced managers into impossible situations where we had to consider not just performance, but also the very livelihoods of our employees.
Adding to the dysfunction, stack ranking wasn't just applied at a large organizational level, it was often enforced at the team level, sometimes on teams of fewer than 10 people. Managers were instructed to hit distribution targets within their small teams, no matter what. This created an artificial ranking system where even high-performing teams had to designate someone as underperforming. The reasoning was that when teams were later rolled up into larger groups, the distribution would already be "corrected" at the lower level.
Worse still, discussions about adjustments to where employees should be placed in the ranking distribution often happened without their direct manager in the room. Decisions were frequently made by managers two or three levels above the employee in question—people who had little to no direct interaction with them. This led to situations where employees were labeled as "low performers" based on secondhand impressions rather than actual work output. It was a deeply flawed process that stripped direct managers of their ability to advocate for their team members, leaving employees vulnerable to arbitrary decisions.
I've seen managers intentionally leave roles unfilled or choose "safe" hires to avoid upsetting the internal balance. And when you're focused on protecting your existing team from a broken system, innovation and excellence take a backseat.
Biases and Infighting
Another major issue with stack ranking is how it magnifies human bias. In theory, evaluations are based on measurable impact. In reality, managers are human, and subjectivity creeps in. People who were more vocal or visible often got more credit than quiet high-performers.
I recall one situation where two employees—both doing excellent work—ended up in direct competition for a promotion. One of them was more assertive, frequently spoke in meetings, and regularly shared status updates with senior leadership. The other, quieter but equally effective, often worked behind the scenes and was deeply trusted by peers. When rankings came around, the louder employee was rated higher, not because of better outcomes, but because they were more visible.
Stack ranking also turned performance reviews into a zero-sum game. Managers had to argue about which employees were better or worse relative to each other, even when everyone was meeting or exceeding expectations. I sat in numerous calibration meetings where leaders fought to defend their team members, often turning collegial relationships into adversarial ones. The emotional toll of those discussions lingered long after the meetings ended.
Misleading and Meaningless Rankings
One of the most frustrating aspects of stack ranking is how misleading it can be. I personally was rated "Top Tier"—supposedly indicating I was in the top 5-10% of the company—my last four years at Amazon. Despite that track record, I was laid off in 2023.
When it came time to decide who to lay off, performance wasn't even part of the conversation. It didn't matter how many high ratings I had received, how many projects I had led, or how many people I had mentored. Layoff decisions were made based on organizational reshuffling and budget cuts. The same system that once used performance ratings to determine promotions and raises was suddenly irrelevant. It was a stark reminder that stack ranking is often more about optics than value.
No Fairness for Those on Leave
Amazon also lacked clear guidelines for evaluating employees on extended leave. I had team members who went on maternity leave or took time off for health reasons. When they returned, they were at a disadvantage simply because they hadn't been as visible during the evaluation period.
Without a structured way to account for this absence, returning employees were often slotted into lower rankings. It was a demoralizing and unfair process that penalized people for life circumstances. Worse, as a manager, I had little to no ability to advocate effectively for them within the rigid structure of stack ranking.
A Damaging Demoralizing System
At its core, stack ranking breeds fear. It shifts the focus from team success to individual survival. I saw people withhold information, avoid collaboration, and even sabotage others to protect their standing. It also led to burnout. Employees were constantly trying to "prove" their worth, even when their contributions were obvious. They weren't just working hard; they were working scared. And scared employees do not do their best work, they do whatever will get them noticed by the right people at the right time.
Stack ranking may have been implemented with the goal of driving high performance, but the reality is that it often drives high anxiety, low trust, and poor decision-making. It penalizes collaboration, rewards visibility over substance, and fails to account for life outside of work. At Amazon, I worked with some of the smartest, most dedicated people I've ever known. But the stack ranking system created an environment where even the best people were vulnerable, and good managers were forced to make bad decisions.
If companies truly want to unlock the potential of their teams, they need to ditch stack ranking in favor of evaluation systems that reward real impact, support employee well-being, and foster long-term growth.
Sources:
Comments