Haute Lumière · The reading

Your company says it values collaboration and promotes the people who skip it

How a performance system quietly overrules the values written on the wall, and what it takes to build one that tells the truth.

She has been at this long enough to stop performing it for anyone. Nothing in this frame is being scored.

THE CAGE

There is a moment — and most leaders can name it, if they are honest — when the performance system they trusted starts to feel like something other than a tool. It was supposed to be a map toward excellence. It measures, it ranks, it allocates, and somewhere along the way it began containing the very aliveness it was built to release. Nobody announced the change. The dashboard looks exactly as it did in the quarter when it was still helping.

This is an argument about how that cage was built, and it starts by clearing away the obvious suspect. No villain built it. It was assembled by earnest, intelligent people solving real problems with the best instruments available to them, and each generation of instrument was genuinely better than the one before. Understanding the history is not an exercise in blame. It is an act of compassion — for the organisations stuck inside it, for the leaders who sense something is wrong and cannot name it, and for the people whose full contribution has been quietly edited out of the equation.

A cage assembled entirely out of good intentions is still a cage, and it is far harder to see.

Underneath everything else, the story of modern performance management is a story about measurement: what was chosen for counting, what stopped being seen as a consequence of that choice, and what the selective attention has cost. Every performance system is a claim about what a person is for. Most organisations have never read their own claim aloud. When they finally do, they find it says something they would never put in a values statement, and something their best people worked out years ago.

The sequence here is deliberate. First the inheritance — a stopwatch, a factory floor, and an assumption that went underground and stayed there. Then the four great reforms that tried to correct it and left the assumption intact. Then the exclusion list: the contributions every organisation depends on and no organisation counts. Then the cost, which is not abstract and not only financial, because bodies keep the account. Then a case. Then the redesign, which is neither an abandonment of measurement nor a softening of standards, but something considerably more demanding than either.

One thing to say at the outset, because it changes how the rest reads. The trap is not evidence that an organisation is badly run. It catches the well-run ones hardest, because sophistication in measurement is precisely what deepens it. A company with a mediocre performance system distorts behaviour weakly and inconsistently. A company with an excellent one distorts it with precision, at scale, quarter after quarter, and calls the result discipline.

TAYLOR'S GHOST

The roots of contemporary performance management reach back to the late nineteenth century and a man named Frederick Winslow Taylor, standing on a factory floor with a stopwatch. Taylor's insight was revolutionary for its time and it is worth stating in its strongest form: work could be studied scientifically. Tasks could be broken into component parts, timed, optimised, standardised. Workers could be trained to perform each motion in the most efficient possible way, and the gains were not imaginary — output rose, costs fell, and the approach spread from manufacturing into every sector of organisational life.

Embedded in the breakthrough, though, was an assumption so foundational that it became invisible almost immediately: the human being performing the work is essentially a machine, and the machine's output can be maximised through better engineering. Taylor never had to argue for this. It was the water industrial-age thinking swam in. Assumptions of that kind do not get refuted, because nobody states them clearly enough to refute; they get inherited.

The consequences follow with a kind of logical inevitability. When you treat a person as a machine, you measure machine-like things — units produced per hour, error rates, time on task. You optimise for speed, consistency and predictability, because those are the qualities a machine can be improved on. You build management around surveillance, correction and incentive, which are the three levers that actually work on machines. And you exclude from measurement, without ever deciding to, everything that makes a human being not a machine.

Taylor did not claim a person was a machine. He simply built a system that could not tell the difference, and we kept the system.

That excluded list is long and it is not decorative: creativity, meaning-making, relational intelligence, moral sensitivity, developmental growth, and the capacity for care that no algorithm has yet reproduced. None of these were rejected. They were never in the frame to begin with, because the frame was designed around a different kind of thing entirely.

Taylor's ghost still haunts the modern performance review, and not in the stopwatch — that was replaced long ago by dashboards and analytics platforms considerably more elegant than anything he had. The ghost is in the underlying assumption that human contribution can be meaningfully captured through quantitative measurement alone. Every time an organisation reduces a complex person to a numerical rating, it is performing a Taylorist operation, whether or not it knows the name. The hardware changed completely. The premise did not change at all.

FOUR REFORMS

The century after Taylor produced four serious attempts to fix what he left behind. Each addressed a real limitation of the approach before it. Each was, by the standards of its moment, a genuine advance. And each preserved the reductionist assumption at the centre, which is why the trap survived four rounds of reform in better condition than it started.

Management by Objectives, introduced by Peter Drucker in the 1950s, added something important and humanising: goals should be agreed collaboratively between manager and employee, creating alignment between individual effort and organisational strategy. This was a real concession that workers are not merely executors of predetermined tasks but agents capable of understanding and engaging with purpose. What it still assumed was that the right objectives can be defined clearly in advance, and that progress toward them can be measured without ambiguity. Both assumptions hold well in a stable environment and poorly everywhere else.

The Balanced Scorecard, developed by Kaplan and Norton in the 1990s, made a larger move by insisting that financial metrics alone were insufficient. Organisations needed to balance financial performance against customer satisfaction, internal process efficiency, and learning and growth. The aperture genuinely widened. But it remained a measurement framework throughout: it broadened what was counted without ever asking whether counting was adequate to what mattered most. A wider net is still a net, and it still catches only the things shaped like fish.

OKRs — Objectives and Key Results, popularised by Intel and later by Google — brought a welcome insistence on ambitious goal-setting and transparent alignment, and the demand for measurable key results created real clarity and focus. It works beautifully for product launches and revenue targets. It works poorly for relational trust, ethical development, creative incubation and the slow, non-linear work of changing a culture, because it assumes that what matters can be expressed as a measurable outcome inside a defined timeframe. Anything with a four-year arc simply does not appear in a quarterly instrument.

Agile and continuous feedback models arrived largely as a reaction against the rigidity of the annual cycle, recognising correctly that yearly feedback is far too infrequent to support development. This mattered. But accelerating a feedback loop without changing the assumptions underneath it produces faster measurement of the same reductive metrics — a person can now be misunderstood weekly rather than annually, in a well-designed interface, with charts.

Four reforms widened the aperture. None of them questioned the lens.

Each of these contributed something genuine and none deserves the caricature it sometimes receives. Yet all four left the foundational premise standing: performance is fundamentally measurable output, and the job of a performance system is to maximise it. What none of them addressed is the question that sits behind the eyes of every thoughtful leader who has ever run a review — what about everything else?

The contribution no system counts tends to look like this from outside: a person, unhurried, paying close attention.

THE EXCLUSION LIST

“What gets measured gets managed” is usually offered as a celebration of measurement's power. It carries a shadow clause that is almost never said aloud: what does not get measured does not get managed, and over a long enough period it stops being valued at all. The maxim is true in both directions. Only one direction ever makes it onto the slide.

Start with what conventional systems do measure, and give it full credit: revenue generated, tasks completed, deadlines met, customer satisfaction scores, error rates, project milestones reached. These are real. They matter. Nothing in this argument suggests running an organisation without them, and any redesign that quietly drops rigour has replaced one blindness with another.

Now the list of what those systems typically do not measure, which is worth reading slowly, because most organisations have never seen it written down. The quality of attention a person brings to their work — not whether the task was completed, but how it was held: with care, with curiosity, with presence. The relational intelligence they exercise: how they navigate conflict, build trust, create psychological safety for others, repair ruptures, hold complexity in a room full of people who disagree. Their developmental growth: whether their capacity for perspective-taking is expanding, whether their tolerance for ambiguity is rising, whether their ethical reasoning is deepening, whether they are becoming more skillful at holding paradox rather than collapsing it.

The list continues past the point where most dashboards end. Their meaning-making contribution — how they help colleagues find purpose and coherence in the work, which has no line item anywhere and shapes organisational vitality more than most line items do. Their somatic intelligence: the capacity to read a room, to sense when something is off, to notice the unspoken tension blocking a team's creative flow, to recognise in their own body that they are approaching burnout before the calendar says so. Their ecological awareness: whether they weigh the impact of a decision beyond the organisation's boundary, on communities, on ecosystems, on people who will inherit the consequence.

These are not soft additions to real performance. They are the ground real performance stands on, and the ground is never on the invoice.

The dependency runs one way and it is not subtle. A team with extraordinary relational intelligence will outperform a team of individually brilliant people who cannot collaborate, every time, on any horizon longer than a quarter. A leader with deep somatic awareness detects organisational dysfunction months before it surfaces in lagging indicators. A culture rich in meaning-making retains talent that no compensation package alone can hold.

Because these capacities resist quantification, they are excluded from the performance equation by default rather than by decision. And the exclusion carries a message, delivered implicitly and absorbed completely: these things do not count. The organisation may have posters celebrating collaboration, innovation and integrity. If the system rewards individual output, speed and compliance, the posters are wallpaper — and everyone in the building has already worked out which document is load-bearing.

THE SHADOW SYSTEM

Every organisation runs two performance systems. The first is official. It has a name, a cycle, a software vendor, a set of published criteria, and a policy document somebody in human resources maintains with real care. The second has no name and no owner, and it is the one that determines promotions, bonuses, interesting assignments, invitations to the meeting where the decision actually gets made, and whose opinion carries weight when two people disagree.

The second system is the real one, and it teaches with devastating efficiency. People learn it in their first six months, usually without being told anything explicitly, by watching who rises and inferring backwards. They do not need the criteria. They need three examples, and every organisation supplies those within a year.

The values statement is what a company believes about itself. The promotion list is what it is.

The gap between the two is where the trouble lives, and mapping it is the most useful diagnostic an organisation can run on itself. Consider the most common version. A company states that it values collaboration. Its performance system measures individual deliverables. Time spent helping a colleague is therefore time not spent on one's own measurable objectives, which means collaboration is praised in the all-hands and penalised in the review — not by anyone's intent, but by simple arithmetic. The generous person's numbers are worse. Everybody can see it. Nobody has to say it.

The same structure produces every other variant. An organisation states that it values long-term thinking and measures quarterly results. It states that it values integrity and rewards the closed deal without asking what was promised to close it. It states that it values the whole person and evaluates a narrow slice of that person's output, then wonders why engagement surveys read flat while nothing specific is wrong.

What makes the shadow system so durable is that it is nobody's project. No executive decided that helping a colleague should damage a career. The outcome emerged from the interaction of a measurement choice and an incentive structure, and emergent outcomes have no author to appeal to. This is also, quietly, the good news: an unowned system can be owned. The gap is not a moral failure requiring confession. It is a design fault, and design faults respond to design.

The diagnostic itself is simple enough to run this month. List what the official system measures. Then list, honestly, the behaviours of the last five people who were promoted — what they actually did, not the citation that was read out. Where the two lists agree, the organisation is coherent. Where they diverge, the second list is the truth, and it is already teaching everybody else what to become.

THE BODY'S ACCOUNT

Here is something that rarely appears in a management textbook and that every person who has worked inside a misaligned performance system knows without being taught. It hurts. Not metaphorically — physically, in specific and locatable ways, and usually well before the person has language for what is wrong.

When someone's full contribution is reduced to a set of metrics capturing a fraction of what they bring, the body registers the incongruence directly. There is a somatic signature to it: a tightening in the chest, a constriction in the throat, a heaviness across the shoulders. It is the body saying something the person has not yet permitted themselves to say. I am more than this. What I offer is not being received. Something essential about me is being made invisible here, and I am cooperating with it.

Left in place, that incongruence becomes chronic and takes three recognisable forms. Burnout is the first, and it is widely misdiagnosed as a workload problem. The research of Christina Maslach and others has consistently shown that burnout has less to do with how much a person works than with the gap between their values and their work environment. A misaligned performance system widens that gap structurally, every cycle, for everybody at once — which is why organisations that respond to burnout with a wellbeing programme and no change to the review find the problem waiting for them the following year.

Disengagement is the second: the gradual withdrawal of creative energy and emotional investment that Gallup surveys have documented across decades. When a system communicates that only a narrow band of a person's capacity matters, the rest of their aliveness retracts. They still show up. They still meet their numbers. But the generative, relational spark that distinguishes a thriving organisation from a merely functioning one has quietly gone out, and no metric in the system is capable of detecting its absence.

Moral injury is the third and the least discussed. The term comes from military contexts, describing the wound that occurs when a person is required to act against their own moral compass. In organisational life it appears whenever a performance system incentivises behaviour that conflicts with a person's values — cutting corners on quality to hit a date, prioritising a short-term metric over a long-term relationship, performing enthusiasm in a culture review while feeling entirely disconnected from it. The body absorbs these contradictions. They accumulate. They erode trust in the organisation, then trust in leadership, and eventually trust in oneself, which is the expensive one.

You cannot innovate from a constricted nervous system, and no incentive has ever been designed that changes this.

Polyvagal theory, developed by Stephen Porges, gives this a physiological frame. The autonomic nervous system continuously scans the environment for signals of safety and threat, a process Porges calls neuroception. Where a performance system creates conditions of psychological threat — surveillance, comparison, the live possibility of judgment or humiliation — the nervous system shifts out of ventral vagal engagement, where social connection, creativity and collaboration live, into sympathetic activation, or further into dorsal vagal shutdown: numbness, withdrawal, the person who is present and absent simultaneously.

Most conventional performance systems operate in the sympathetic zone, whether or not anyone intended that. They generate productivity through activation — urgency, competition, anxiety about outcome. It works in the short term and it produces measurable output. It does so by spending exactly the capacities a values-based system exists to cultivate: creativity, relational depth, ethical sensitivity, sustainable engagement. You cannot build genuine trust while in fight-or-flight. You cannot reach your deepest judgment while your body is braced for evaluation.

Attention is not a soft quality. It is the condition under which everything countable gets made.

MERIDIAN SYSTEMS

Take a composite case, drawn from multiple consulting engagements and disguised to protect confidentiality, but faithful to a pattern that repeats with unsettling regularity. Call the company Meridian Systems: mid-sized, technology, and by every conventional measure a high-performing organisation. Revenue growth exceeded industry benchmarks. Product delivery timelines were consistently met. Customer satisfaction was strong. Leadership had invested heavily in a sophisticated OKR framework, quarterly reviews and a data-driven approach to talent.

Something was wrong anyway, and the chief executive could feel it without being able to name it. Employee surveys showed acceptable engagement scores. The qualitative comments told a completely different story from the numbers immediately above them. I feel like a cog. Nobody cares what I think, only what I produce. I used to love this work, and now I perform it.

Turnover was creeping upward, concentrated among the most experienced and creative people — precisely the ones the organisation could least afford to lose. Exit interviews produced a consistent theme, and it was not compensation. People were leaving because they felt unseen. Their most important contributions — mentoring junior colleagues, navigating complex stakeholder relationships, holding institutional knowledge, maintaining ethical boundaries under pressure — were invisible to the performance system. What got measured was output. What got valued was speed. What got rewarded was individual achievement.

The irony was precise and painful. Meridian's values statement, displayed in every conference room, read: we value collaboration, integrity, innovation, and the whole person. The performance system measured none of those four. The real system, the one determining promotions and bonuses, rewarded individual deliverables, speed to market, and revenue contribution. Collaboration was celebrated at all-hands and penalised in practice, because every hour spent helping someone else was an hour missing from one's own objectives.

The metrics were excellent. The soul was departing. Both statements were true on the same dashboard, in the same quarter.

This is what the measurement trap looks like from inside a company that is, by any external standard, succeeding. Not malice anywhere in the chain. Not incompetence. Just the invisible logic of measurement doing what it does: when you can only count certain things, those things become the things that count, and everything else — however essential, however explicitly celebrated in the rhetoric — gets gradually deprioritised, defunded and devalued without a single meeting ever being held about it.

Meridian had optimised itself into hollowness. It is worth being exact about what that means, because it is the outcome every well-run organisation is closest to. Not failure. Not a scandal. A company that hits every number it has chosen to watch while the substrate underneath those numbers thins out, year over year, until one bad quarter arrives and the resilience that would have absorbed it is no longer in the building.

THE ADEQUACY CLAIM

Beneath all of this sits a single assumption that functions almost as an axiom of organisational life: what matters about human performance can be adequately represented through quantitative measurement. This is the reductionist assumption, and it deserves to be named precisely because it is so rarely examined. The objection is not to measurement, which is essential and not going anywhere. The objection is to the adequacy claim — the tacit belief that with the right metrics, enough data and sufficient sophistication, the full reality of human contribution can be captured.

It cannot, and this is not a failure of measurement technology awaiting a better instrument. Certain dimensions of human contribution are constitutively resistant to quantification. Their nature is such that reducing them to numbers does not measure them imperfectly; it substitutes something else and then reports on the substitute.

Consider trust. Proxies for trust can be measured — survey scores, retention rates, willingness to speak up in a meeting — and those proxies are informative. But trust itself is not a number. It is a lived, relational, embodied reality that emerges between specific people in specific contexts and shifts from one conversation to the next. A trust score is not a compressed version of trust. It is a different object with a similar name, and organisations routinely manage the second while believing they are managing the first.

Consider wisdom. A leader's capacity to hold several perspectives at once, to sense the right timing for a difficult conversation, to know when to push and when to wait, to distinguish productive discomfort from harmful pressure — these are among the most valuable capacities in any organisation, and they are unmeasurable in any meaningful sense. Their effects can be evaluated after the fact. The capacity itself lives somewhere numbers do not reach, which is why it is so often possessed by the person the system rates third.

Consider beauty, which looks like an odd entry on a performance list and is not. There is a quality of beauty in work done with care, in a meeting facilitated with grace, in a product designed with genuine attention to the person who will use it. That beauty is not decoration. It is functional: it produces engagement, loyalty and meaning, and it cannot be captured on a dashboard by anyone at any budget.

The reductionist assumption does not merely fail to see these things. It removes their structural support and lets them starve.

That last point is the sharp one. When a reward system runs entirely on quantifiable output, people redirect their energy toward what is counted, rationally and without cynicism. The unmeasured dimensions — trust, wisdom, beauty, relational care, developmental growth — are never explicitly discouraged. They simply receive no structural support at all. And in the competitive ecology of organisational attention, whatever receives no structural support withers, on a timeline slow enough that nobody connects the cause to the effect.

THE SCAN

Diagnosis before redesign, and the diagnosis is seven steps that any leadership team can run on itself without a consultant, a licence or a budget line. The point is not to produce a score. It is to make the second, unofficial system visible in writing, because a system nobody has written down cannot be argued with.

First, identify the key metrics. Review what the organisation currently measures and make a comprehensive list — revenue, task completion, customer satisfaction, cycle time, every quantitative output that carries weight. Be complete rather than flattering. Include the informal metrics that nobody publishes and everyone tracks.

Second, evaluate the missing contributions. Beside the first list, build a second: the contributions that are essential to performance and appear nowhere in the first. Relational intelligence. Creativity. Ethical behaviour. Mentoring. Institutional memory. The gap between the two lists is the assessment's central finding, and most teams are startled by how wide it is when the second list is written by the people doing the work rather than the people designing the system.

Third, assess somatic responses. During performance reviews and feedback sessions, attend to physical sensation — tension, constriction, heaviness, the urge to leave the room. These are data, not distraction. A manager who notices their own chest tighten while delivering a rating has learned something about that rating that the form will never record.

Fourth, analyse the shadow system. Name the unofficial behaviours and dynamics that are actually rewarded, and set them against the stated values. Fifth, reflect on recognition: recall the occasions when people felt fully acknowledged for what they brought, and identify the conditions that made that possible. Those conditions are reproducible, and almost nobody tries to reproduce them because nobody asked the question.

An organisation that cannot describe its own shadow system is being run by it.

Sixth, consider the redesign, and begin in the right place. Envision a performance system that honours every dimension of human contribution, including those that resist quantification — and start from what matters most to the organisation and its people, not from what the current software can report. Seventh, document the insights: gather the findings, the gaps and the recommendations into one coherent document, because a diagnosis held only in conversation evaporates and a diagnosis on paper becomes a mandate.

There is a companion practice worth naming, the Embodied Performance Scan, which teaches leaders and teams to read somatic signals of alignment and misalignment as a routine competence rather than an occasional insight. Constriction in the chest during a performance conversation, chronic tension in a team's collective body, the heaviness that accompanies values incongruence — all of it is vital information that conventional metrics miss entirely. Performance is not only what people do. It is how their bodies participate in the doing, and the body reports first.

OPENING IT

If the descriptions here have been recognisable, that recognition is not evidence of failure. It is evidence of having inherited a paradigm built for a different era and a narrower understanding of what a human being is. The trap is ordinary. What is not ordinary is seeing it clearly, and seeing it clearly is most of the work, because a system nobody can see is a system nobody can change.

The way out is not the abandonment of measurement, which would substitute one blindness for another and would make the organisation worse in ways that show up fast. It is three moves held together: expanding what counts as performance, diversifying the methods of assessment so that qualitative and somatic evidence sit alongside quantitative evidence with equal standing, and redesigning the system so it honours the full spectrum of contribution rather than the portion that fits a column.

In practice that means several things at once. Broaden the criteria to include creativity, relational intelligence and ethical consideration, and give them real weight in real decisions rather than an advisory paragraph at the end of the form. Assess coherence between stated values and actual performance on a regular cycle, not once during a rebrand. Align every practice — incentives, feedback, recognition — with the values the organisation claims, and review them when they drift, because they will drift. Honour the non-measurable explicitly: make trust, collaboration and the quality of working relationships visible in the organisation's own account of how it performs.

Two conditions make the rest possible. The first is safety: a genuine space for open dialogue about performance and values, where people can describe their experience without calculating the cost of describing it. The second is continuous learning — feedback and adaptation treated as the normal state rather than an event, so the system stays responsive instead of hardening into the next inheritance.

Design for what matters and then find the measurements. Every trap in this argument comes from doing it in the other order.

The redesign draws on more than one tradition, and deliberately so: Kegan's constructive-developmental theory on how adult capacity grows, Spiral Dynamics on how value systems evolve, Porges on what the nervous system is doing during evaluation, Appreciative Inquiry on starting from what already works, the Balanced Scorecard on widening the aperture, Self-Determination Theory on what actually motivates a person, and indigenous wisdom traditions on reciprocity and relational accountability. No single framework is adequate to a human being. Several, held together with judgment, get considerably closer.

Five questions carry this further than any framework will, and they are best answered in writing. What does your performance system actually measure, and what matters most to you that is absent from that list. Where do you feel performance measurement in your body — in your chest, your gut, your shoulders, your jaw — during the last review you gave or received. What is the shadow system in your organisation, and where does it diverge from the stated values. When have you felt most fully seen in your contribution, and what conditions made that possible. And if you could redesign the system from scratch with no constraints, what would you want it to honour.

That last question carries an instruction inside it. Do not start with what is measurable. Start with what matters, and then build the instrument that can see it. The full argument — the history, the somatic dimension, the assessment, and the redesign in detail — is in the volume, and the reading is free.

Free to read

Free to read, and free to hear. Every chapter of every book in this house, and every narration of it, is open to anybody. No account, no card, nothing to cancel.

What is in it

• Values-Based Performance Systems™open this in the house search

What cannot be counted is not therefore optional; it is usually the ground the counted things stand on.

A values statement is a hypothesis. The promotion list is the result.

Measurement does not describe an organisation. It selects one.

The body files its report before the survey does, and it is rarely wrong.

Keep looking

Every phrase on this page opens into the house search. The shelf holds The Developmental Canon and six other shelves, and the reading is free.