Performance management. Say it out loud in a room of people who do the work and watch what happens to their shoulders.
It’s a remarkable phrase. On the surface it’s meant to indicate that we’re not managing you, we’re only managing your performance. But that split is the whole problem. It takes a person, abstracts them into a stream of measurable output, and then manages the stream, optimizing for ever-increasing outcomes. The person is still standing there in all their messy, human glory. They just aren’t the thing under discussion any more. As if their performance were distinct from themselves.
“Perform” arrives in English around 1300 meaning to carry out, to accomplish, to finish a thing — but by about 1600 it also means to act or sing on a stage. Both senses are still alive in the word. If what we’re evaluating is your performance, then on some level we’re evaluating the show. And people are extremely good at giving us the show. Not because they’re dishonest. Because the show is what gets attention and rated more often.
Disclaimer, this is not based on any one place I’ve ever worked but is drawn from conversations with lots of people over the years who work at various places, research I’ve done into how performance is managed and measured, and my personal thoughts and views.
Where the phrase actually comes from
The earliest recognizable performance rating system in industry was Robert Owen’s, in the early 1800s, at his cotton mills in New Lanark. Owen hung a four-sided block of wood — a “silent monitor” — at each workstation. Each face was painted a different colour, and the colour turned outward reported yesterday’s performance. The first performance rating in industrial history was a public shaming device you could read from across the room. We’ve spent two centuries making it more private, not less humiliating.
Then, 1917. The US Army had a sorting problem: millions of men arriving, and someone had to decide who led and who was sent forward. So they hired a psychologist. Walter Dill Scott, of the Carnegie Institute of Technology, built a “man-to-man” rating scale for the Committee on Classification of Personnel. It worked well enough that industry imported it through the 1920s and 30s under the name “merit rating,” and the annual grade settled into corporate life and never left.
The term we actually use is much younger. “Performance management” was coined in the late 1970s by Aubrey Daniels, a clinical psychologist who’d become convinced that Skinner’s behavioural science worked as well in a factory as in a clinic. His own organization still describes what he meant by it as “a way of getting people to do what you want them to do and like it.”
That is, I think, the most honest sentence ever written about the field. Note especially the last three words. Compliance was never the whole deliverable. Manufacturing enthusiasm was in scope.
Shame the slow worker, sort the recruits, shape the behaviour, and get them to enjoy it. Nowhere in that lineage is anyone developing a person.
Deming spotted the statistical flaw forty years ago
W. Edwards Deming listed “evaluation of performance, merit rating, or annual review” as one of his seven deadly diseases in Out of the Crisis. He said: “The idea of a merit rating is alluring. The sound of the words captivates the imagination: pay for what you get; get what you pay for; motivate people to do their best, for their own good. The effect is exactly the opposite of what the words promise.”
His objection wasn’t sentimental, it was statistical. Most of the variation in any individual’s results comes from the system they’re working inside, meaning the tooling, the handoffs, the priorities that changed in week three, the dependency team that went quiet. Management owns that system. Rating individuals against each other assigns to the person what actually belongs to the system. You end up ranking noise, with a straight face, and attaching money to the ranking.
The inclined treadmill
Jack Welch’s version at GE was the vitality curve: 20-70-10. Celebrate the top 20%, keep the “vital 70,” remove the bottom 10% — every year, whatever the bottom 10% had actually done that year. GE abandoned it in 2015. Microsoft ran its own curve, and Kurt Eichenwald’s reporting on “Microsoft’s lost decade” found that every current and former employee he interviewed, every one, named stack ranking as the most destructive process inside the company. Microsoft killed it in November 2013 with a memo that opened, more or less, “no more curve.”
Both companies quit it. The shape survived anyway, in organizations. You don’t need a curve to build the machine. You only need to do one thing consistently: reward going above and beyond, and then quietly reset “beyond” as the new floor.
Do that for a few cycles and you have built an inclined treadmill. The belt keeps moving, the angle keeps steepening, and the apex, top box, exceeds, high performer, alpha, wizard, whatever you call it, recedes exactly as fast as anyone approaches it. Everybody climbing. Nobody arriving. Everybody somewhere on this silent mirror of the burnout spectrum.
I mean that fairly literally, companies reward employees who by good fortune, breeding, psychological makeup, can not burn out while being constantly stressed. The WHO added burn-out to ICD-11 in 2019 as an occupational phenomenon, not a medical condition, defined as “a syndrome conceptualized as resulting from chronic workplace stress that has not been successfully managed,” with three dimensions: exhaustion, cynicism or mental distance from the job, and reduced professional efficacy.
The long plateau of failure
If “exceeds” is the only rating that carries meaning, then “meets” becomes a euphemism. And “meets” is what the great majority of good, steady, load-bearing people will receive, most years, because that is what the word means. Everyone learns to read it as a soft failure. Which means you have taken the honourable centre of your organization, the people who actually do most of the work, and told them in writing, annually, that they are a problem because they are not in the top 10%.
People mostly don’t fight that. They slide. Quietly, off the lower end of the treadmill. Often not out of the company at all, just out of the part of themselves they used to bring to it. No complaint gets filed. Nothing shows up in the attrition numbers for another eighteen months. Turnover gets normalized.
Gallup’s most recent read has global engagement declining for a second consecutive year, with much of the recent downturn concentrated in managers. They put the cost of low engagement at roughly $10 trillion in lost productivity, about 9% of global GDP, and note that each single percentage point represents about 21 million fewer engaged employees. Whatever you make of the methodology, the direction is not ambiguous.
Variance isn’t a defect. It’s the resolution of a career.
The thing a rating structurally cannot represent: performance is not a property of a person. It’s a property of a person, doing a particular task, at a particular moment, inside a particular system. Move any one of those and the number moves with it. We should rate the system annually.
Someone brilliant at ambiguous greenfield work will look ordinary on a migration. A writer who carried a demanding, all out product launch for six months will have a slow quarter afterwards, and probably should. Your strongest contributor last year has a parent in hospital this year. A team that shipped something genuinely hard is going to be worse at everything else for a while, and that’s not a warning sign, that’s reality.
None of that is a performance problem. That’s what a human career looks like at normal resolution. The annual rating takes that waveform, samples it once, badly, and judges the person.
Two philosophies that already admit this
I keep coming back to two ideas that I think fit together pretty well.
David Marquet’s intent-based leadership or leading with intent, out of his time turning around the USS Santa Fe, sets two conditions on a team being able to run without you: competency and clarity. Competency is the craft, obviously, but it reaches further than that: it includes being able to connect with people and build trust quickly. Clarity is knowing how your work connects to the mission, the goals and expectations, and to the person on the other end of it.
The part that matters here is that the mix is different for every person, and it does not hold still. Nobody has perfect clarity all the time. With the ground moving the way it is, everyone has competencies that need work just to stay relevant. That’s not a ranking.
Situational Leadership, Hersey and Blanchard’s “Life-Cycle Theory of Leadership,” published in the ASTD Journal in 1969 and renamed in the mid-70s, comes at the same truth from the other side. There is no single best leadership style; you match yours to how ready the person is for the work actually in front of them. For work someone is ready to own, get out of the way, because hovering there signals nothing but distrust. For work that’s new to them, or ambiguous, or high-stakes, lean in and stay close until they find their footing, then back off again.
Competency and clarity tell you what a given person needs. Situational leadership tells you how much of you they need in order to get it. And the entire apparatus assumes the answer keeps changing per person, per piece of work, per month because readiness moves as reality changes.
It’s worth noticing who else this catches. If someone on your team appears to need no shoring up on either axis, that isn’t a five out of five, exceeds expectations, it’s a signal that you don’t yet know what they need. The rating system reads that same person as top box and moves on, which is how you lose them. They succeed into boredom. Your system rewards their ability to stay in a role that they find easy.
Neither model produces a number for the compensation spreadsheet. That’s the real obstacle, and it’s worth naming plainly: most organizations know the ratings are a fiction and keep them because the comp model needs an input, and nobody wants to redesign comp.
The rating is where failure gets priced
Every organization says it wants experimentation. Take risks. Fail fast. Try the new tool. And then, at the end of the year, it rates people on outputs and that rating is where the real price list gets published. A failed experiment costs the person who ran it, in a rating that follows them into next year’s calibration.
Everybody reads that price list. Nobody mentions it. And the following year you get proposals that were guaranteed to work before anyone started them, which is exactly the team that gets left behind when the ground moves. You can’t ask people to make failure cheap while running an instrument whose whole function is to make it personally expensive. You can remove risk from the equation but if you silently add it in as a tax during performance reviews, then people will catch on.
What I try to do instead
- Rate the work, not the person. “That project went sideways and here’s what I think happened” survives being said out loud. “You are a 3” does not.
- Make “meets” an honourable place to live. Most of the work in any organization is done by people who are reliably good and not currently on fire. If the only vocabulary you have for them is a euphemism for disappointment, fix the vocabulary before you fix the people.
- Ask what the year looked like, not what the peak looked like. A team where everybody peaks simultaneously is a team that’s about to lose several of them.
- Separate the growth conversation from the money conversation. The moment a development discussion has a compensation consequence sitting at the table, you stop hearing the truth about where somebody is struggling which was the only useful thing that conversation could have produced.
- When performance drops, look at the system first. Deming’s point, and it holds. Most of the time you’ll find a blocker nobody cleared, or something you changed. A leader whose job is removing blockers faster than they pile up has a very different relationship to a slow quarter than one whose job is grading it.
- Say the quiet part to your team, out loud. Nobody can be exceptional all the time. People are startled by how much room that sentence gives them.
Hold the bar still
Performance management is a 1970s name for a 1910s instrument with an 1800s ancestor: sort them, rank them, shape the behaviour, and get them to like it. It was built for a workforce that has been gone for a century, and the version most of us run today mostly measures who can keep climbing an incline that steepens by design.
The alternative isn’t the absence of standards. It’s a standard honest enough to admit that a career is a wave, that the same person is excellent and ordinary at different things in the same month, and that “reliably good” is not a diagnosis.