Performance management

Performance management is setting expectations people can see, supporting them to meet those expectations, and following through with honest consequences when they do not. Fair process is the fourth element and the one v1.0 left implicit: the person knows the bar, has been told where they stand, has been given a real chance, and can see that the same rules apply to everyone. The competency matters because its signature failure — avoiding the hard conversation — is paid for by the rest of the team, who see the gap and quietly recalibrate their own standards, and because its opposite failure, the consequence delivered without the expectation or the support, is unfair in a way that everyone also sees. The examinations test it at every scope: the engineer whose slow delivery had a real cause, the manager who brings a formal plan for someone who was never told the bar, the leader whose results are excellent and whose methods are hollowing out the organisation.

1. People can only meet a bar they can see. Expectations are explicit, written where it matters, and revisited when the role or the work changes. A performance conversation should refer to something the person already knew.
2. Early, private and specific. A gap is raised in the week it is noticed, one to one, with instances. Waiting for the review cycle turns a correctable problem into a verdict.
3. Diagnose before responding: skill, will, or conditions. The person does not know how, is not applying themselves, or was not given what the job needed. The responses differ entirely, and the third is more common than managers like to admit.
4. A formal plan is a real chance, or it is a lie. A time-boxed, well-supported improvement period with a clear bar is fair. Paperwork for a decision already made is not, and the team knows the difference.
5. Follow through. When the period ends and the bar is not met, the decision was made when the plan was set. Unwinding it teaches everyone that the bar is negotiable and that the plan was theatre.
6. How results are achieved is part of performance. Output that is produced by fear, by burning out a team, or by breaching the organisation's values is a performance problem, and the results do not offset it.

Performance management is well supplied with process and thinly supplied with evidence. The models here are practitioner tools with clear provenance; the fairness claim rests on a body of organisational-justice research that the cited article summarises for managers.

Skill and will practice

The diagnostic from Max Landsberg, The Tao of Coaching: Boost Your Effectiveness at Work by Inspiring and Developing Those Around You (1996): before responding to underperformance, separate whether the person cannot (a skill gap, met with teaching and support) from whether they will not (a will gap, met with expectations and consequences). This unit adds a third branch the original does not name: the conditions gap, where the person was never given what the job required — clear requirements, access, time, a functioning team. A skill gap treated as a will gap produces a demoralised person; a will gap treated as a skill gap produces endless coaching of something coaching cannot fix; a conditions gap treated as either produces an unfair review that the whole team recognises as one.

Fair process practice

The three elements set out in W. Chan Kim and Renée Mauborgne, Fair Process: Managing in the Knowledge Economy (1997) — engagement, explanation, and clarity of expectation — drawn from the authors' research on why people accept decisions that go against them. Applied to performance: the person is heard, the reasoning for the judgment is explained, and what is expected next is unambiguous. Its importance is that a consequence people regard as fairly reached is accepted and learned from, while the same consequence reached unfairly is resisted and remembered by everyone who watched. It is graded practice here because the management version is a synthesis, though the underlying justice research is substantial.

Output and leverage practice

Grove's framing in Andrew S. Grove, High Output Management (1983) of a manager's output as the output of the organisation under them, which is the reason a manager's performance cannot be read from their own report: it is read from sustained team health, sustained delivery and how the results were achieved, over time. The same framing explains why a leader who hits every number by leading through fear is a performance problem — the organisation's future capacity is what is being spent to produce the present number.

Calibration practice

The mechanism, shared with hiring, by which ratings across teams are made to mean the same thing: shared standards, evidence over advocacy, and a joint review when two managers disagree about the same person. Its purpose is fairness, and its characteristic decay is into horse-trading, where ratings are negotiated rather than evidenced. A calibration meeting in which the most assertive manager's people are consistently rated highest has stopped being one.

Expectations, written

For each person, a short written statement of what the role expects at their level and what the next quarter is meant to produce, agreed in a one-on-one and revisited when either changes. This is the artifact a later performance conversation points at. If it does not exist, the first act of performance management is to write it, not to raise the gap.

The early conversation

The first performance conversation is small: the gap, an instance or two, the question of what is getting in the way, and an agreed next step. It happens the week the gap is noticed and it is not called a warning. Most performance problems that reach a formal stage did so because this conversation was skipped.

The diagnosis before the plan

Before any formal step, the manager writes one paragraph answering three questions: were the expectations explicit, was the feedback given and documented, and is this skill, will, or conditions? If the first two answers are no, the formal step is premature and the manager's own gap comes first.

The improvement period

A formal plan states the specific bar, the specific support — pairing, reduced scope, a named reviewer — the duration, and what happens at the end. It is reviewed weekly. At the end the outcome is the one the plan defined, delivered with the same directness the plan was set with.

Reading a manager's performance

For someone who leads others, performance is read from three things over at least two quarters: their team's health, measured by attrition, engagement and what skip-levels reveal; their team's delivery; and how the results were produced. A single bad quarter with a cause outside the manager's control is judged by the quality of their decisions and their response, not the raw outcome. A single good quarter produced by heroics is not evidence of much.

At team scope the object is an engineer's performance against expectations the manager set, and the manager acts directly. At organisational scope the manager of managers does two jobs: quality-controlling how their managers manage performance — declining a formal plan built on absent expectations and coaching the manager first — and judging the managers themselves, on evidence over time rather than on their own status reports. At executive scope the object is leaders whose results are entangled with context, politics and the executive's own public bets, and the fairness discipline becomes calibrating across organisations by the delta a leader produced rather than the health of what they inherited. The pattern is in How Judgment Scales; the managers whose performance is being read are the subject of PL-6 Leading through leaders.

• Sets expectations the person can see, and revisits them when the work changes. team
• Raises a gap early, privately and with instances, before it becomes a crisis. team
• Separates skill from will from conditions, and responds to the one that is actually there.
• Treats a formal plan as a real, supported chance with a defined end, and follows through at the end.
• Protects the team by making a hard decision about one person when the damage outweighs the output, including the high performer whose behaviour breaches the team's values (see CC-5 Values, integrity, and ethical judgment under pressure).
• Failure mode — hoping the problem resolves itself. team
• Failure mode — vague feedback that never names the gap. team
• Failure mode — endless coaching of a will problem. team
• Failure mode — a formal plan for someone who was never told the bar.
• Declines a manager's formal plan when expectations were never explicit, and coaches the manager to set them first. org
• Judges managers on sustained team health, sustained delivery and how results were achieved, over time and in context. org
• Runs calibration as a fairness mechanism — evidence over advocacy, joint review when managers disagree — not as a ranking ritual. org
• Names an equity problem when a manager gives all the high-visibility work to favourites; opportunity is part of compensation (see CC-3 Inclusive practices and equitable opportunity distribution). org
• Owns a headcount-reduction selection with consistent, defensible criteria after hearing the manager's case in full. org
• Failure mode — approving a plan built on absent expectations because "the manager owns their team". org
• Failure mode — judging managers by their own status reports or by likeability. org
• Failure mode — punishing outcomes without reading context, or excusing patterns because of it. org
• Failure mode — letting calibration decay into horse-trading. org
• Holds leaders accountable for how results are achieved; a leader who delivers through fear while capable people leave is a performance problem, full stop. exec
• Gives a loyal, long-tenured leader the organisation has outgrown honesty and a reshaped role or a dignified transition, not quiet marginalisation. exec
• Acts on the evidence when a leader the executive publicly backed is failing; doubling down to protect the bet loses both the organisation and the reputation. exec
• Calibrates leaders across organisations by the delta they produced and the quality of their decisions, not the health of the hand they were dealt. exec
• Failure mode — tolerating a feared-but-effective leader until the pipeline beneath them is hollow. exec
• Failure mode — letting loyalty freeze the organisation chart. exec

Care versus clarity. The manager wants to protect the person from a hard message and the person needs the message to have any chance of meeting the bar. Clarity delivered with care is the resolution; care delivered instead of clarity is the failure the whole unit describes.

Fairness to the individual versus fairness to the team. A long improvement period is fair to the person and costly to the team carrying the gap. The resolution is a period that is genuinely time-boxed and genuinely supported, so that both the person and the team know when it ends.

Context versus accountability. Reading a bad quarter in context is fair; reading every bad quarter in context is how patterns get excused. The judgment is in the timescale: one quarter is weighed by decisions and response, a pattern across quarters is the performance.

Results versus methods. The leader whose numbers are excellent and whose people are leaving forces the question of what performance is. This standard's answer is that the methods are part of the result, because the organisation's future capacity is what is being spent; but the pull toward the number is strong, and the examination tests whether the candidate resists it.

Loyalty versus scope. Tenure and past contribution earn respect and candour; they do not earn permanent scope. Honouring the first while refusing the second is uncomfortable and is the only fair option, because the alternative is either an organisation that cannot grow past its early leaders or a quiet demotion nobody was honest about.

A manager who reports to a senior manager arrives with a formal improvement plan for one of their engineers, drafted and ready to issue. The engineer's delivery has been slow for two quarters. Reading the plan and the history, the senior manager finds that the engineer's objectives were never written down, that the feedback in the period consists of two informal comments in one-on-ones, and that the requirements for the engineer's main project were, by the manager's own account, unclear until six weeks ago.

The two easy responses are both wrong. Approving the plan because the manager owns their team issues an unfair consequence to someone who was never given a fair chance, and everyone on that team will read it correctly. Taking the case over — rewriting the plan, or having the conversations with the engineer directly — solves the case and teaches the manager nothing, while telling their team that the manager is not really their manager.

The senior manager declines the plan and names why, to the manager, as feedback about the manager's own performance: expectations were never explicit, feedback was not given or documented, and the conditions for the work were poor for most of the period. The engineer has not yet been given the chance the plan claims to be. The diagnosis for the engineer is, on the evidence, at least partly a conditions gap, and the diagnosis for the manager is a skill gap in performance management.

They then coach rather than do. The manager is asked to write the engineer's expectations for the coming quarter this week and agree them in a one-on-one; to give specific, documented feedback at least fortnightly; and to fix the requirements problem with the product partner, which is the manager's job regardless of the engineer. A date is set eight weeks out to look at the evidence again together. If, with clear expectations and real support, the delivery is still short, a formal plan will then be fair, and the senior manager says so now, so that the manager does not hear the decline as protection of the engineer.

What the senior manager keeps for themselves is the observation about the manager. A manager who reaches for a formal plan before setting expectations has shown a gap of their own, and the senior manager records that as an item for the manager's development, and for their own reading of the manager's performance over the next two quarters.

• PL-2 Feedback and difficult conversations — the feedback that precedes any performance conversation, and the difficult conversation itself.
• PL-6 Leading through leaders — managing the performance of managers, and quality-controlling how they manage performance.
• CC-5 Values, integrity, and ethical judgment under pressure — the high performer whose behaviour breaches the team's values, and the costly right call.
• CC-3 Inclusive practices and equitable opportunity distribution — opportunity as part of compensation; the equity problem inside performance decisions.
• PL-3 Hiring and onboarding — the hire that is not landing, and the early intervention that precedes any formal step.

1. Why is skill, will or conditions the first diagnostic question?
   Answer

   Because the responses differ entirely: teaching and support for a skill gap, expectations and consequences for a will gap, fixing what the person was not given for a conditions gap. Responding before diagnosing produces the wrong one, and the team sees it.
2. What makes a formal improvement plan fair rather than paperwork?
   Answer

   A clear bar the person already knew, specific support, a defined duration, a stated outcome at the end, and a real chance of meeting it. A plan written after the decision has been made is the opposite of fair, and everyone can tell.
3. A manager brings you a formal plan for an engineer who was never given clear expectations. What do you do, and what does it tell you about the manager? org
   Answer

   Decline the plan, coach the manager to set expectations and give documented feedback first, and set a date to revisit with evidence. It tells you the manager has a skill gap in performance management, which becomes an item in their own development and in your reading of their performance.
4. How do you read a manager's performance, and why not from their status reports?
   Answer

   From sustained team health, sustained delivery and how results were achieved, over at least two quarters, in context. Their own reports are the most filtered source available, and a manager's output is the output of their team.
5. Two of your managers cannot agree on the rating for an engineer who worked across both their teams. What is the resolution, and what are the two wrong ones? org
   Answer

   A joint review of the evidence against the shared standard. The wrong resolutions are seniority — the more senior manager's view wins — and splitting the difference, which rates the negotiation rather than the person.
6. A vice-president hits every number and leads through fear; capable leaders keep leaving beneath them. What exactly is the performance problem, and what does delay cost? exec
   Answer

   The methods are the problem: the organisation's future capacity is being spent to produce the present number, and the results do not offset it. Delay costs the pipeline of leaders who leave, and teaches everyone that results buy exemption from how they are achieved.
7. A leader you publicly backed is clearly failing. What is the credibility move? exec
   Answer

   Acting on the evidence. Executives who double down to protect the appointment lose both the organisation and the reputation they were protecting; acting is what credibility looks like.
8. A long-tenured director whom the organisation has outgrown is loyal to you personally. What are they owed? exec
   Answer

   Honesty: an explicit conversation, a role reshaped to their real strengths, or a transition with dignity. Tenure earns respect and candour; it does not earn permanent scope, and quiet marginalisation is a form of neglect.

• Max Landsberg, The Tao of Coaching: Boost Your Effectiveness at Work by Inspiring and Developing Those Around You (1996) — the skill and will diagnostic.
• W. Chan Kim and Renée Mauborgne, Fair Process: Managing in the Knowledge Economy (1997) — engagement, explanation, clarity of expectation.
• Andrew S. Grove, High Output Management (1983) — a manager's output as the output of their organisation.