Body of Knowledge · Reference
Glossary
Every term of art this standard uses, defined as the standard uses it, with the units that develop each.
v3.1v2.0 archive
- Accountability without fear
- Clear ownership and honest consequences for patterns of unaccountability, combined with real safety for honest error. People must know exactly which of the two applies to them. Conflate them and both are lost.
- Admission criteria
- The written statement of how an initiative gets onto a department roadmap (expected contribution to stated outcomes, cost, risk, dependencies), so that a request is answered with the criteria rather than with a no.See: SV-3
- Affinity bias
- Favouring people who resemble oneself in evaluation. It is why interview panels should not resemble the manager, and one of the things structured interviewing is designed to counter.
- AI governance
- The explicit, proportionate policy for AI-assisted work. It says what the tools may be used for, what data may be sent where, what verification and disclosure each class of change requires, and who owns a change whatever produced it. It is scaled to the risk of the path and enforced in the pipeline where it can be.
- Aligned autonomy
- Direction, priorities and boundary standards set centrally, with methods left to teams. This is the balance an operating cadence strikes at organisational scale, and the answer to the question of what to standardise across teams.
- Architecture review forum
- A standing forum where significant designs get cross-team scrutiny early. It is most valuable when it advises, teaches and catches cross-team consequences, and least valuable as a gate every change must pass.
- Blameless postmortem
- An incident review that asks how the system allowed the failure and what to change, rather than who to blame. It makes honesty about failure safe, and so makes reliability improvable. Blamelessness does not mean an absence of accountability, which is handled separately.
- Bottom-up roadmap
- A department "strategy" assembled from team wish lists, with no connecting thesis. More detail in the list never fixes it. The fix is to define the outcomes and derive the list again from them.See: SV-1
- Brilliant jerk
- A high performer whose behaviour breaches the team's values. Tolerating one teaches everyone that results buy exemption.
- Brooks's law
- Adding people to a late software project makes it later, because the existing people must train the new ones and the communication paths multiply. It applies to work that cannot be partitioned.See: DE-2
- Build, buy, or AI service
- The three-way capability decision (build it, buy a product, or adopt an AI service), with each option weighed by the same questions: is this core, what is the lifetime cost, and where are we left if the provider or partner changes course?
- Burnout
- Exhaustion, cynicism and reduced efficacy caused by the conditions of the job (workload, control, reward, community, fairness, values) rather than by a shortage of individual resilience. It is treated at the cause, by changing the conditions.
- Bus factor
- The number of people who could leave before critical knowledge or capability is lost. It is a dependency to engineer away, and once the person has given notice it is an option with an expiry date.
- Calibration
- Aligning evaluation standards across interviewers, managers or organisations so that the same capability or performance earns the same judgment. Evidence counts over advocacy, and assessors who disagree review the case together.
- Capacity allocation
- A stated share of a team's capacity for platform, reliability and debt, held under pressure, so that foundational work is a decision rather than whatever is left over. Zero is also a decision, and it comes with a date on which it fails.See: SV-3
- Charter
- The explicit statement of what a team or department owns. Charter disputes are settled at the level that sets charters instead of being fought commit by commit.
- Coaching versus directing
- Developing someone's judgment through questions, as against supplying the answer. Coaching is the default for developmental moments. Directing is the named exception for emergencies, complete novices and stakes that leave no room.See: CC-2
- Commitment versus estimate
- An estimate is the best forecast. A commitment is what is promised outward, deliberately buffered, with its scope, known risks and variance protocol stated. Saying both aloud makes the buffer visible and builds trust.See: DE-2
- Conditions gap
- Underperformance caused by what the person was not given (clear requirements, access, time, a functioning team) rather than by a lack of skill or will. It is the third branch of the skill-and-will diagnostic, and the one managers most often miss.See: PL-4
- Cone of uncertainty
- Estimates are least accurate at the start of work, when least is known, and narrow as reality is discovered. What follows from it: give ranges, cut small slices, de-risk the fuzziest part first, and re-forecast as the cone narrows.See: DE-2
- Connecting thesis
- The statement of why these teams do these things, tying several teams' work to a few department-level outcomes. Without one, you have a bottom-up roadmap.See: SV-1
- Contract testing
- Automated verification of the agreement between a service and its consumers, which catches a breaking change at the boundary before integration does. It is the technical half of ownership and contract at the boundary.
- Core versus context
- Build what differentiates the business in its customers' eyes (core), and buy or adopt the commodity (context) so engineers are free for differentiating work. Core means differentiating, which is a different thing from important. Email is important and almost never core.See: SV-4
- Cost of delay
- The value lost for each unit of time a piece of work remains unshipped. Sequencing by cost of delay, rather than by effort, age or advocacy, is the economic core of prioritisation.See: DE-1
- Costly moment
- A situation in which a stated value and the pressure of the moment point in different directions, and people are watching the decision. Culture is set in these moments. A value that holds only when it costs nothing is decoration.
- Culture
- The accumulated pattern of what behaviour gets rewarded, tolerated and punished in practice. What leaders visibly do sets it far more than stated values, policies or announcements. Where behaviour and the stated values diverge, the behaviour is the culture.See: CC-5
- Cycle time
- How long work takes from start to done, as measured. It is the basis of historical forecasting, and it is a distribution rather than a single number. Quoting its average as a date repeats the original estimation error.See: DE-2
- Decision process
- The usual way a manager of managers intervenes in a cross-team technical decision. Before any opinion is issued, they install the criteria, the evidence required, a named decider and a date, so that the organisation keeps the capability.
- Decision rights
- The explicit statement of who may decide what, and at which level of authority. Where more than one person could reasonably claim a decision, the rights are designed and written down in advance instead of being discovered during the dispute.
- Default gradient
- The tilt of a distributed organisation toward wherever its decision-makers sit. Decisions, information, stretch work and promotions flow to the centre without anyone intending it. Design counters it; reminders never do.
- Delegation brief
- The written form of a delegation for anything larger than a day's work: outcome, constraints, authority level, who else is involved, and when the work is next reviewed together. If the brief cannot be written, the delegation is not ready.See: PL-1
- Developer experience
- The friction between an engineer and a shipped change (feedback loops, cognitive load, flow state), treated as measurable capacity. Investment in it is judged by the friction removed rather than the tooling installed.See: DE-5
- Disagree and commit
- Argue against a decision directly and with reasoning. If it goes the other way, commit to it fully instead of holding back effort while waiting to be proved right. This presumes that a decision was in fact made and that the disagreement was heard.See: SV-2
- Disengagement signals
- Withdrawal, cynicism and ideas that stop coming. A manager who is paying attention sees these early signs months before a resignation. On a distributed team they are fainter and arrive later.
- DORA metrics
- The four research-validated delivery measures (deployment frequency, lead time for changes, change failure rate, time to restore service). Together they describe a team's delivery system, and they move together in high-performing organisations. They are team-level measures, never individual ones, and never a leaderboard.
- Ego depletion
- The contested claim that self-control and decision quality draw on a limited resource that runs down through the day. A large multi-laboratory replication did not find the effect. The practical advice that survives rests on the recovery research instead.See: CC-6
- Engineering strategy
- The written translation of business goals into technical direction. It contains a diagnosis, a set of choices that includes what will not be done, and the reasoning. It is reviewed against reality and revised when signals call for it, not because the calendar does.See: SV-1
- Error budget
- The tolerable shortfall against a service-level objective, treated as a resource to spend that gates risk-taking. With budget available, ship; with the budget exhausted, stabilise. It is only as good as leadership's willingness to obey it when that is inconvenient.
- Executive docket
- The explicit, short list of technical decisions an executive owns personally because they cannot be separated from company strategy. These are bet-the-company technology, build-versus-buy for the differentiating capability, lock-in of a core capability, foundational infrastructure, and the security and reliability posture. Everything else is delegated with real authority.See: TJ-4
- Fair process
- A decision people accept even when it goes against them, because they were heard, the reasoning was explained, and what is expected next is unambiguous. Its three elements are engagement, explanation and clarity of expectation.See: PL-4
- Falsifiable scorecard
- A small, stable set of measures declared when a strategy is set, able to show that the strategy is failing. It covers the outcomes the strategy was meant to move, leading indicators chosen in advance, and trends in delivery and organisational health. Without one, success gets judged by the effort spent.See: SV-5
- First ninety days
- Treating a new role as a transition with its own work (learning, an early real contribution, the relationships the role depends on), with explicit outcomes at thirty, sixty and ninety days. For senior hires, most of the ninety days goes on integration.See: PL-3
- Glue work
- The coordinating, documenting and unblocking work that holds a team together and rarely shows up in the artifacts a promotion process values. The people who do it tend not to be promoted for it.
- Goodhart's law
- When a measure becomes a target, it ceases to be a good measure. It explains how healthy activity metrics can sit alongside flat outcomes, and it is the caution attached to every metric in this standard.
- Ground truth
- The real state of a piece of work, established by looking rather than by reading a status report. Above team scope, managing risk is largely a matter of managing the reporting system that carries status upward.See: DE-4
- GROW model
- A coaching structure (Goal, Reality, Options, Will) that gives "ask before telling" a shape a manager can hold in a one-on-one.See: CC-2
- Growth ladder
- An explicit description of what each level looks like for each competency, so that a promotion is judged against a shared description rather than against how assertive a manager is. Also called a competency matrix. It describes what each level involves, and sets no schedule.See: PL-5
- Halo effect
- One salient positive trait colouring the assessment of unrelated traits. In hiring, it is the impressive credential or strong first answer that lifts every later score. Independent scoring per criterion, before any discussion, counters it.See: PL-3
- Horizon mix
- The explicit split of an organisation's capacity between serving the current business, building the platform the strategy will need in a few years, and exploring what might disrupt everything. It is a decision owned at executive level, and should never be a residual.See: SV-3
- Improvement plan
- A formal, time-boxed, supported chance to meet a stated bar, with the outcome at the end defined when the plan is set. It is fair only if the person already knew the bar and had been given feedback. Paperwork for a decision already made does not count as one.See: PL-4
- Incident command
- A single defined role, assigned before the incident, that owns coordination and communication so that responders can focus on the fix. A commander who starts debugging has left the role.See: DE-3
- Innovation budget
- An explicit, bounded allocation for evaluating new technology, with pilots that have criteria, a contained blast radius and an honest review. Adoption is then neither prohibited nor stampeded, and the team that took the risk is not left holding the liability.See: TJ-3
- Intent-based leadership
- Replacing "may I…" with "I intend to…". The person states what they are about to do and why, the leader's silence counts as consent, and an objection has to be voiced. Decisions move to where the information is, and the leader stays informed.See: PL-1
- Interface discipline
- Versioning, a deprecation policy, contract tests and communicated change for a widely used interface. The owning team's definition of done includes its consumers' stability.
- Intervention threshold
- Step into a technical decision when being wrong would be costly and hard to reverse (one-way doors, safety, security, data integrity, customer impact in flight). Stay out when it is reversible and owning the outcome teaches more than the correction would. Apply the threshold consistently, and intervene by raising what was missed rather than by overriding.See: TJ-4
- Inverse Conway manoeuvre
- Designing the organisation to produce the architecture you want, on the basis that the system will mirror the organisation's structure either way.
- Judgment debrief
- The form a one-on-one with a manager takes. You walk through the hard call they face or have just made, probe the reasoning, and coach the thinking instead of issuing a ruling. This is how managers develop.See: PL-6
- Key-person risk
- Critical capability, context or ownership concentrated in one person, so that the organisation is destabilised when they leave or are courted. It is a structural property to design away rather than a personal failing to lament.
- Leading change
- What a structural change needs beyond its announcement. That means the why, what changes and what stays, the effect on every leader whose scope moves, presence through the transition, and consolidation so the change sticks.See: SV-6
- Learning organisation
- An organisation designed to turn experience into shared capability. Incident learnings travel between teams, postmortem actions are audited to completion, slack is funded for learning, and leadership reviews ask what was learned.See: CC-2
- Learning zone
- The combination of high psychological safety and high standards. High safety with low standards is the comfort zone. Low safety with high standards produces silence and hidden failure.See: CC-1
- Location-neutral
- Decision-making and information flows designed so that a team four time zones away has the same access as one down the corridor. In practice that means proposals with comment windows, meetings every location can attend, and decisions published in writing.
- Managerial leverage
- Grove's framing that a manager's output is the output of the organisation under them, and that activities differ in how much output they produce per hour spent. An hour spent making a delegation clear pays back across every hour the delegate then spends on it.
- Mirroring hypothesis
- The empirically tested form of Conway's observation. Products built by loosely coupled organisations are measurably more modular than equivalent products built by tightly coupled ones.See: TJ-1
- Office housework
- Necessary, low-visibility tasks such as notes, scheduling and cover, which fall along predictable and inequitable lines unless they are distributed deliberately.See: CC-3
- One-on-one
- The recurring, protected meeting between a manager and a report. It belongs to the report and is not for status. Blockers, frustrations, growth and the earliest signs of disengagement come up here. With a manager as the report, it becomes a judgment debrief.
- One-way door
- A decision that is hard to reverse and so deserves deliberation and senior involvement. A two-way door, by contrast, is reversible and should be decided quickly by the people closest to it. When in doubt, treat a door as one-way until shown otherwise.
- Operating cadence
- An organisation's rhythm of planning, review and escalation. It is sound when it is light and rolling against stable priorities, reviews evidence, has explicit escalation triggers and leaves time for correction. When it does none of that, it is planning theatre.See: DE-5
- Outcome versus output
- Framing work as the change wanted in the world instead of the thing to be built. Outcomes define success and leave the path open; outputs do neither. The test is to ask, of any item in a plan, why it is there.See: SV-1
- Ownership map
- The written record of every system, capability and decision and who owns each, checked before and after any structural change. Whatever has no owner on the map is where the next incident will come from.
- Paved road
- A well-supported default platform and toolchain that makes the good path the easy one. Teams may leave it if they take on the ownership cost themselves. A paved road nobody would choose is a mandate in costume.See: TJ-3
- Planning fallacy
- The systematic tendency to underestimate how long a task will take, even when similar tasks are known to have overrun, because the forecast reasons from the case at hand instead of from the distribution of similar cases. The outside view corrects it.See: DE-2
- Portfolio governance
- An explicit, ranked and deliberately small set of concurrent initiatives at organisational scale, reviewed against evidence on a cadence. The ranking settles contention for resources by strategy instead of by politics.See: DE-1
- Pre-mortem
- Before committing to a plan, assume it has failed and ask the team to write down why. This legitimises dissent and brings out the risks that optimism had suppressed. It works only where safety already exists.See: DE-4
- Privacy by design
- Privacy as the default, built into ordinary engineering decisions (what is collected, where it is copied, who can read it, what is logged, how deletion works in practice) instead of bolted on afterwards.See: TJ-6
- Proximity bias
- Opportunity and promotion going to the people a leader sees, regardless of performance. It was measured directly in the work-from-home experiment, where remote employees outperformed and were promoted less.
- Psychological safety
- A shared belief that the environment is safe for interpersonal risk-taking, such as admitting error, asking questions and challenging. It is one of the strongest predictors of team performance. It is independent of standards, and it is different from comfort or an absence of conflict.See: CC-1
- Reaching around
- A leader of managers overriding a manager's decisions with that manager's own engineers, fixing their team's problems directly, or letting their escalations become the leader's workload. It is the organisational form of taking delegated work back.
- Readiness
- Already exercising the judgment of a larger scope on real stakes before holding the title. Tenure, ambition and how much the person wants the role are no measure of it.
- Recovery
- The experiences through which people recuperate from work (psychological detachment, relaxation, mastery, control over one's time). Of these, detachment is the one most strongly related to well-being. A boundary is defined by the presence of detachment, which is a different thing from the absence of work.See: CC-6
- Rescuing
- Stepping in with the solution the moment someone struggles. It feels like help and is faster today, and it teaches people they cannot be trusted to think. It is the anti-pattern of both coaching and delegation.
- Risk-based triage
- Prioritising security findings by exploitability (whether the vulnerable path is reachable) and exposure (what it gives an attacker, and to how many), and fixing the whole class where several findings share a cause. It answers both failure modes, paralysis at the count and ignoring it.See: TJ-6
- Self-management
- The manager's judgment applied to the manager. It covers where their time and energy go, what boundaries they hold, and how they stay able to think clearly under load. It is the competency that sustains all the others.See: CC-6
- Service-level objective
- An explicit reliability target for a service, from which the error budget is derived. It settles how reliable is reliable enough, so the argument is not re-fought for every decision.See: DE-3
- Service ownership
- Every service has a named owner responsible for its operation, its interface stability, its documentation and its future. A critical service with no owner comes first, because everything about its health depends on somebody being responsible.
- Situation, behaviour, impact
- A feedback structure that names the specific situation, describes the observable behaviour instead of inferred character, and states the impact. It removes the two triggers of defensiveness, generalisation and diagnosed intent.See: PL-2
- Situational leadership
- Adjusting leadership style (directing, coaching, supporting, delegating) to a person's competence and confidence on the task at hand. The style follows the task, whatever the person's title.See: PL-1
- Skill and will
- The first diagnostic question about underperformance is whether the person cannot or will not, because the right responses are entirely different. This standard adds a third branch, the conditions gap.See: PL-4
- Skip-level
- A regular one-on-one between a leader and the people who report to their reports. It exists for unfiltered signal about team health and to make the leader visibly accessible. It is never a channel for decisions that belong to the manager.
- Span of control
- How many reports a leader can serve effectively. There is no constant. It depends on the reports' maturity, the stability of the work, and how much coordination each report needs. A span that works in steady state fails during change.
- Staff+ track
- The senior individual-contributor career path (staff, principal and beyond), of equal standing to management and with several distinct archetypes. A manager who cannot describe it cannot offer it honestly.See: PL-5
- Strangler pattern
- Replacing a legacy system incrementally by growing the new capability around it and routing traffic away piece by piece until the old system can be retired. It is the standing alternative to the big-bang rewrite.
- Structure follows strategy
- Changes in strategy require changes in structure, and structures that outlive the strategy that produced them become the obstacle to the next one. The organisation is how the strategy is implemented.
- Structured interview
- Consistent questions mapped to a rubric defined before the process starts, scored independently by each interviewer before any discussion. It is among the best-evidenced predictors of job performance, and fairer as well.See: PL-3
- Structured sensing
- How an executive keeps their technical judgment current. The means are architecture and operations reviews with evidence, principal engineers who tell the truth, periodic deep dives into the systems carrying the most risk, and samples of raw signal that keep the aggregates honest.See: TJ-4
- Succession
- A ready or nearly ready successor for every key role, the executive's own included, treated as an explicit expectation of senior roles and not as a personal virtue. It is the structural answer to key-person risk.
- Sunk cost
- What has already been spent, which argues for nothing in a decision about the future. Only the value remaining against the cost remaining, compared with the alternatives, counts. Changed fundamentals justify switching a bet; novelty does not.
- Sustainable pace
- A pace a team can hold indefinitely. Occasional, time-boxed crunch for a real reason may be acceptable. Chronic crunch is a management failure, and a run of heroics is a failure of the system rather than a badge.See: CC-4
- Systems understanding
- The knowledge that lasts after a manager stops coding and matters for leading, namely the system's shape, risks and dependencies and who understands each part. A manager keeps it by reading designs without approving them, staying in the hard rooms, and rotating deep dives.See: TJ-4
- Task-relevant maturity
- How much direction a person needs on a specific kind of work, set by their experience with that work and not by their general seniority. It is reassessed as tasks change.
- Technical debt
- The implied cost of future rework from choosing expedience now. A controlled amount is a sensible trade. Left unmanaged, the interest (slower delivery, more defects, more toil) compounds until movement stops.See: TJ-2
- Technical-debt quadrant
- Debt classified as deliberate or inadvertent, and as prudent or reckless. Deliberate, prudent debt is a trade taken knowingly; reckless debt is mess with a flattering name. The quadrant gives a team the vocabulary to classify debt before arguing about it.See: TJ-2
- Threat modelling
- A design discipline built on four questions (what are we building, what can go wrong, what will we do about it, did we do a good job). It makes security part of design instead of a scan at the end, and a manager can ask the questions without being a security engineer.See: TJ-6
- Time log
- Two weeks of recording every hour as it is spent, then comparing the record with what the team needs. It is the first practice of self-management, because people are reliably wrong about where their time goes until they measure it.See: CC-6
- Total cost of ownership
- The full lifetime cost of a capability, meaning build cost plus maintenance forever, or purchase price plus integration, lock-in and exit. It corrects reasoning from the sticker price. The discipline lies in naming the costs explicitly, not in computing them precisely.See: SV-4
- Trade-off transparency
- Saying no by showing the cost of yes ("we can do this, and that slips; which do you want?"), so the stakeholder makes an informed choice and the manager does not quietly absorb the demand.See: DE-1
- Watermelon status
- Reporting that is green on the outside and red on the inside. It points to a broken information system that has learned to optimise for reassurance, rather than to a broken team. The fix is objective status criteria, evidence-based review and visible safety for bad news.See: DE-4
- Work-in-progress limit
- A cap on how much is in flight at once, so that the most important things finish sooner. Starting less finishes more. Unless the limit is enforced at intake, it is decoration.See: DE-1