"Stack, Rank & Yank": The Performance System That Refuses to Die
Why a discredited practice keeps coming back under softer names, what it costs the people inside it, and how to think differently about performance.
How Stack, Rack & Yank (SRY) Breaks High Performers
A few weeks ago, I was on a call with some old colleagues. The conversation drifted, the way these calls do, into what each of us was doing now, and then one of them mentioned, almost in passing, that he had come to consider stack, rank & yank, or SRY, as I will call it from here on, the single biggest reason he had left his last organization. What he added next has stayed with me. He had grown steadily in every role he held since, he said, which was a relief, because by the time he left, he had started doubting himself due to the verdict the system had delivered on him.
The moment he said it, others on the call began sharing similar experiences. Quietly at first. One by one. The longer I listened, the clearer it became that SRY has done more damage to people I consider rockstars than anyone has been willing to admit.
What SRY Actually Does
SRY forces managers to distribute their team along a fixed performance curve, regardless of how the team actually performed. The most famous version is Jack Welch’s vitality curve at GE: the top 20% celebrated, the middle 70% retained, and the bottom 10% fired every year. Welch called the practice differentiation, though the shop floor knew it as rank and yank.
The original case for the model has been steadily overturned. Microsoft, Adobe, Deloitte, and GE itself eventually moved off it. The practice persists anyway, under softer names: calibration, talent differentiation, performance distribution, top grading. Amazon still operates one of the most aggressive versions through its annual “unregretted attrition” target, reported at around six percent of the workforce per year. Meta returned to harder differentiation under Mark Zuckerberg’s “year of efficiency.” Tesla operates a famously demanding variant. Most of management consulting and large parts of investment banking still run on up-or-out models that function the same way. The vocabulary keeps evolving while the underlying mechanics stay intact.
Why It Goes Against Everything We Know About Growth
Carol Dweck’s distinction between a growth mindset and a fixed mindset is one of the most-cited frameworks in organizational psychology. People who treat ability as developable take on harder problems, recover from setbacks faster, and stay in their craft longer than people who treat ability as essentially fixed. SRY is the most efficient mechanism a company can deploy to install the fixed version across its entire workforce. When the score that determines your job security is comparative, your incentive shifts from learning faster than yesterday to looking better than the colleague sitting next to you. You stop volunteering for the hard project that might fail. You stop sharing the play that is working. You stop coaching the new hire who might one day outperform you.
Amy Edmondson’s research on psychological safety identifies the same condition from the team angle. The single most reliable predictor of team performance is whether members feel safe to take interpersonal risks. A forced ranking system is architecturally opposed to that condition.
Why It Survives Anyway
If the science is this clear, the obvious question is why the practice persists. The honest answer, the one most consultants will not put on a slide, is that SRY was never designed to serve the employee. It was designed to serve the organization, and it does so in five specific ways.
It makes compensation budgets more predictable because a forced curve flattens the payout distribution.
It makes terminations legally defensible because a documented bottom ten is harder to challenge in court than a narrative judgment.
It substitutes for managerial courage by allowing managers who have not built the muscle to give honest feedback to hide behind the calibration committee.
It feeds a Wall Street narrative in which ruthlessness is packaged as a proxy for performance culture, despite a generation of contrary evidence.
It defends the comforting story that the company hires only the best, positioning the bottom of the curve as a mistake to be corrected rather than as evidence that the comparative frame itself is the problem.
When the Mask Came Off
The version of this practice that has always bothered me most is its relationship to the recruitment funnel. We tell candidates the process is brutal because the bar is high, welcome them with a celebratory email, and six months later, put them on a curve where 10% must be marked unfit by the same company that just spent six figures recruiting them. Ranking in the bottom 10% every year is an implicit confession that one in 10 hiring decisions was wrong, and a choice to address that failure by punishing the hire rather than fixing the process or possibly punishing the hirer — why not?
There is something uniquely cynical about telling a person that surviving the recruitment process makes them special, then constructing a system whose architecture guarantees that one in ten of them will feel the opposite by Christmas.
The larger confession came when artificial intelligence (AI) arrived, and the layoffs began. After years of telling employees that the calibration committee was the rigorous arbiter of performance, the same companies conducted some of the largest workforce reductions in their history without leaning on those ranking systems in any meaningful way. The cuts were not “we are letting go of the bottom ten percent of our calibrated talent.” They were “we are eliminating this team,” “we are deprecating this function,” “we are anticipating which roles AI will absorb and removing them in advance.” Many of the people let go in those waves had been rated highly on the very systems that cost their companies millions of hours a year to administer.
That is the moment the mask came off. If your performance apparatus had produced rankings rigorous enough to fire the bottom slice of the workforce every year for a decade, those same rankings should have been the obvious instrument for any larger reduction. They were not used because they were never really designed for that. The system was always about manageable, defensible, plausibly meritocratic turnover during normal times. The moment real cost decisions arrived, performance disappeared from the conversation and was replaced by org chart math, role obsolescence math, and the political calculus of which functions had stronger executive sponsorship.
What the performance review apparatus was actually built for was something more modest by comparison: managing the workforce in a way the institution and its legal department could defend.
What It Does to People
The mental health cost of these systems is well-documented and rarely discussed in the rooms where the systems are designed.
Chronic workplace stress is a well-studied driver of sleep disruption, immune dysregulation, and cardiovascular load. The American Institute of Stress estimates the cost of workplace stress to U.S. employers at more than $300 billion a year due to absenteeism, turnover, and healthcare. The World Health Organization has formally recognized burnout as an occupational phenomenon in the ICD-11, defining it as a syndrome resulting from chronic workplace stress that has not been successfully managed.
I have lost count of the coaching clients who came to me with what looked like generalized anxiety and turned out to be a perfectly proportionate response to the system they were inside. People stop sleeping well in October (replace this with whatever cycle you are on!). They start rehearsing every interaction with their manager. They begin to interpret neutral feedback as catastrophic. They develop what I have come to call calibration insomnia, the condition of lying awake replaying months of small interactions, trying to predict a verdict that has already been written somewhere they cannot see.
The harder cost is identity. Years of being slotted into a curve teach a professional to evaluate themselves through the eyes of an institution that was never designed to see them whole. Many of the most capable people I work with have spent their first few coaching sessions just untangling the system’s voice from their own.
What looks like weakness in those early conversations is almost always the residue of a system that worked exactly as designed.
A Coaching Lens
My coaching and mentoring rest on a few foundational propositions. The client is the expert in their own life. Awareness is evoked through powerful questions rather than imposed through a verdict. Trust and safety are necessary conditions for any movement forward. The coach holds the space and the client owns the choice.
SRY inverts every one of these. The employee is told they are not the expert; the calibration committee is. Awareness is delivered as a number rather than evoked through inquiry. Safety is replaced with comparative threat. Choice is replaced with a verdict.
The contradiction that results is unresolvable as long as both elements coexist inside the same organization. You cannot credibly ask people to operate from a growth mindset while running a system whose architecture punishes growth behaviors, ask them to take risks while ranking them on outcomes, or ask managers to coach when their formal job is to sort.
Why I Built RISEUP@work
RISEUP@work began as my answer to that contradiction and as a place to put what I have learned over decades of building, leading, and coaching within performance cultures. It rests on the conviction that people who survive these systems carry knowledge the systems were never designed to surface, and that the work of the next decade is to help individuals reclaim agency that the system was structured to take away.
The program is built on coaching first principles rather than management theatre. It treats the professional as the expert in their own trajectory. It teaches the framing and narrative skills that give people language for what was done to them, vocabulary for what they want next, and the tools to walk into rooms designed to evaluate them as someone with a story rather than a score.
I built it because the alternative was to keep watching capable people internalize verdicts that said more about the system than about them.
A Closing Thought
If you run a company, the honest question is whether the people who leave you list your performance system as the reason. Ask it in your next round of exit interviews. Ask it of the people who left two years ago, not the people in their notice period. The answer you get will tell you more about your culture than your engagement survey ever has.
If you are inside one of these systems, the question is different. It is worth asking whether the verdict you have absorbed about yourself was earned, or whether it was structurally guaranteed to land on someone in your seat.
A Self Check Worth Sitting With
Three questions for either side of the table.
Whose voice are you listening to when you evaluate your own year?
When was the last time you heard a colleague describe themselves in language you suspect they borrowed from a calibration committee?
If the curve disappeared tomorrow, what would you do differently next quarter, and why are you not already doing it?
Either way, the conversation starts with naming the system honestly. We can stop pretending that ranking the bottom 10% of human beings as the bottom of a curve every year amounts to performance management, when it is actually a comfortable lie the organization tells itself to avoid the harder work of building something better.