Career development and growth paths

Career development is understanding where a person wants to go and creating the experiences, not only the conversations, that get them there. Its instrument is the one-on-one: the recurring, protected meeting that belongs to the report rather than to status, where blockers, frustrations and the earliest signs of disengagement surface if anyone is listening. Its structure is the growth path: levels and tracks that make "what does the next level look like?" concrete and comparable, including the senior individual-contributor track that lets strong engineers grow without leaving the work, and the honest choice between that track and management. The competency matters because retention is largely decided here, months before any resignation, and because a growth path that only exists in the manager's head is distributed by the manager's attention, which is not the same as fairly. The examinations test it through the promotion request that is early, the strong engineer who is bored, the ambitious manager who wants your job, and the succession call with one seat and two ready people.

1. The one-on-one belongs to the report. Status can travel by other channels. This meeting is for what does not surface otherwise: what is in the way, what they want, what has changed. Cancelling it whenever things heat up signals exactly the wrong priority.
2. Growth is built from experiences. Conversations set direction; stretch assignments, real responsibility and visible work are what move a career. A development plan with no experiences in it is a wish.
3. Make the next level concrete and comparable. Ladders or competency matrices describe what each level looks like, so that a promotion is a judgment against a shared description rather than against the manager's assertiveness in the calibration room.
4. Promotion follows demonstrated judgment, not tenure or desire. The evidence is that the person is already exercising the next level's judgment on real stakes. A promotion to retain, or to reward waiting, teaches the wrong lesson to everyone who watched.
5. The senior individual-contributor track is a real path, and management is not a prize. The choice between them is about the work the person wants to do, and a manager's job is to make it honestly, including by describing what management actually involves.
6. Watch for disengagement early. People rarely leave suddenly. Withdrawal, cynicism and a stopped flow of ideas appear first, in one-on-ones, for a manager paying attention.

Career development is practitioner territory. The models below are widely adopted and clearly attributed; none is a controlled finding, and the unit grades them as practice.

The one-on-one as the report's meeting practice

Grove's treatment in Andrew S. Grove, High Output Management (1983) of the one-on-one as the manager's highest-leverage recurring activity, whose agenda belongs to the subordinate and whose purpose is mutual learning about the work and the person. The practical consequences have held for four decades: it is protected, it is regular, the report sets much of the agenda, and the manager listens more than they talk. The failure that has also held: it becomes a status meeting the moment the manager needs it to.

Growth ladders and competency matrices practice

Explicit descriptions of what each level looks like per competency, making progression concrete and promotion decisions comparable across people and teams. The engineering-management version is set out in Camille Fournier, The Manager's Path: A Guide for Tech Leaders Navigating Growth and Change (2017), which walks the arc level by level. A ladder describes the what, not a schedule; it decays into checkbox collection when levels are treated as tenure milestones rather than demonstrated judgment, and it is only as fair as the calibration that applies it.

The Staff+ track practice

The senior individual-contributor path — staff, principal and beyond — as a career of equal standing to management, described from the inside in Will Larson, Staff Engineer: Leadership Beyond the Management Track (2021) and Tanya Reilly, The Staff Engineer's Path: A Guide for Individual Contributors Navigating Growth and Change (2022). Two ideas from that literature matter for managers: that the track has distinct archetypes rather than one shape, so "senior engineer but more so" is not a description of it; and that the invisible coordinating work that holds teams together, which Reilly names glue work, tends to be done by people who are then not promoted for it. A manager who cannot describe the track cannot offer it honestly.

Readiness practice

Not a named model but a rule this standard adopts from the literature above and from practice: someone is ready for a larger scope when they are already exercising that scope's judgment on real stakes, before holding the title. Readiness is not tenure, not ambition, and not how much the person wants it. The rule matters most at succession, where the pull toward the person who has waited longest or asked loudest is strongest.

The protected one-on-one

Weekly or fortnightly, never cancelled for a status need, with the report's items first. The manager keeps a running note per person — what they said they wanted, what they said was in the way, what changed — and reads it before each meeting. The note is what makes the pattern visible: three consecutive meetings with no ideas offered is a signal the individual meetings did not show.

The growth conversation, quarterly

Separate from the weekly one-on-one and from the review cycle: where does the person want to be in two years, what would the next level require of them, and which experience in the next quarter moves them toward it. The output is one concrete assignment, not a plan document.

Answering a promotion request

The request is answered with evidence against the level description: where the person already shows the next level's judgment, where they do not yet, and what would show it. Neither a yes to retain nor a no without reasons. If the answer is not yet, the path is specific enough that the person can see the gap closing.

Describing management honestly

When an engineer considers management, the manager describes the actual job — the meetings, the ambiguity, the distance from the code, the accountability for others' work — and asks what the person wants to spend their days doing. The strongest engineer who has said they do not want to manage is not pressured into it; the reluctant manager is worse for everyone than the empty seat.

Promotion fairness across teams

Where several teams share a ladder, promotions go through shared criteria, written evidence and a calibration that compares people against the description rather than against each other's managers. The check is simple: if one team's promotions are consistently faster, either its people are consistently stronger or its manager is consistently more assertive, and the second is more common.

At team scope the object is an individual's growth and next step, and the one-on-one is where it happens. At organisational scope the object becomes the growth paths available across several teams and who is getting them: promotion mechanics that are fair across managers, moves between teams, and the stretch roles that exist only at department scale. At executive scope the object is the career architecture itself — levels, tracks and the leadership pipeline — and succession as a structural property: every key role, including the executive's own, has a ready or nearly-ready successor, and readiness is judged on demonstrated next-level judgment. Developing the managers who report to you is its own competency, PL-6 Leading through leaders; the pattern is in How Judgment Scales.

• Holds one-on-ones consistently, lets the report set much of the agenda, and listens more than they talk. team
• Builds growth from real experiences: the stretch assignment, the visible piece of work, the responsibility not yet held. team
• Answers a promotion request with evidence, a named gap and a path. team
• Describes the senior individual-contributor track and management honestly, and lets the person choose the work they want. team
• Notices withdrawal, cynicism and a stopped flow of ideas as the early signals they are. team
• Failure mode — cancelling one-on-ones under pressure. team
• Failure mode — hijacking the one-on-one for status. team
• Failure mode — discussing growth only at promotion time. team
• Failure mode — promoting to retain, or promoting for waiting.
• Failure mode — drafting a reluctant engineer into management because it is the only visible path to more pay or status.
• Creates growth paths across teams — moves, stretch roles, the senior track — and sees who is and is not getting them. org
• Runs promotion fairness across teams with shared criteria, evidence and calibration. org
• Decides compensation on value and fairness: fixes an underpaid engineer because it is wrong, not because they threatened to leave; corrects a package that inverts a tenured colleague's through the compensation process, honestly. org
• Failure mode — letting each team invent its own promotion bar, so each team's ceiling is its manager's assertiveness. org
• Failure mode — the reactive counter-offer, which teaches everyone that threats work and rarely retains anyone for long. org
• Failure mode — a director who hoards strong engineers and blocks internal moves to protect local output; talent belongs to the organisation. org exec
• Designs the career architecture — levels, tracks, the leadership pipeline — and reads it by what it produces. exec
• Makes succession an explicit, non-optional expectation of every senior role, including their own. exec
• Judges readiness by demonstrated next-level judgment on real stakes, not tenure or ambition. exec
• Tells the runner-up for a senior seat the truth about their path, invests in it genuinely, and accepts they may leave. exec
• Failure mode — inventing a title to avoid the succession conversation. exec
• Failure mode — a pipeline one name deep, and critical initiatives that destabilise when one leader is courted. exec
• Failure mode — equating readiness with tenure, or with how much someone wants the role. exec
• Failure mode — counter-bidding for a leader whose growth was neglected for years. exec

The person's ambition versus the evidence. A talented person wants the next level before consistently showing its judgment. Saying no risks losing them; saying yes teaches the team that asking works. The resolution is the specific path: a no that names exactly what would make it a yes, and an assignment that lets them show it.

Retention versus fairness. The counter-offer retains one person this quarter and distorts equity for everyone who did not threaten to leave. Fixing genuine underpayment because it is wrong, before it is used as leverage, is the only version that survives.

Local output versus organisational growth. A team's manager loses their strongest engineer to an internal move and the team's delivery dips. The organisation gains a person who grows and stays. Talent belongs to the organisation, and the manager who blocks the move is optimising the wrong thing, but the dip is real and the judgment is in how it is absorbed.

Ladders versus judgment. A detailed ladder makes promotion fair and comparable; it also invites people to collect its boxes. The resolution is to treat the ladder as a description of judgment to be demonstrated, not a list of activities to be completed, and to say so in every promotion conversation.

Honesty versus hope. The runner-up for a senior seat, the engineer who will not reach the next level on this team, the manager who wants a role that is not coming: each is owed the truth, and the truth may cost the relationship or the person. Manufacturing hope costs more, later, and is noticed by everyone.

An engineer with ten months in the team asks their manager for promotion to senior. They are talented, fast, and have shipped two features that mattered. The manager's honest reading is that the engineer is not yet consistently showing senior judgment: they still need scope handed to them, they have not yet led anything that involved another team, and their design reviews are sharp on their own code and quiet on everyone else's. The engineer has mentioned, twice, that a friend at another company was promoted at the one-year mark.

The two easy answers are to approve the promotion to keep a talented person, or to decline and say "not yet" with a vague reference to time. The first teaches the team that promotion follows asking and tenure; the second teaches this engineer nothing and leaves them to conclude the manager is the obstacle.

The manager answers with the level description. In the one-on-one they acknowledge the growth specifically — the two features, the pace, the quality of the engineer's own work — and then walk through the senior description alongside the evidence: senior engineers define their own scope, lead work that crosses a team boundary, and raise the quality of others' work through review. On the first two there is no evidence yet, because the engineer has not had the opportunity; on the third the evidence points the other way. The manager says plainly that the promotion is not yet, and that the reason is evidence, not time.

Then they build the path from experiences rather than advice. The engineer is given the lead on an upcoming change that touches a neighbouring team's service, with the manager's support at the boundary and an explicit expectation that the engineer defines the scope. They are asked to review two pull requests a week from outside their own area and to give the feedback in the review, not privately. A date three months out is set to look at the evidence again together, and the manager says what would make the answer yes.

The engineer may still leave for the company where the friend was promoted. The manager has judged that a promotion the engineer cannot yet sustain would cost them more than the risk of leaving costs the team, and has given them the honest path and the opportunity to walk it. What the manager does not do is promote to retain, or decline without giving the engineer anything they can act on.

• PL-1 Delegation, ownership, and decision rights — the stretch assignment as delegation: scope matched to capability plus a deliberate stretch.
• PL-6 Leading through leaders — developing the managers who report to you, and the leadership pipeline as a system.
• CC-2 Coaching versus directing — coaching as the form growth takes when the goal is judgment.
• CC-4 Motivation, recognition, and sustainable pace — motivation and the early signals of disengagement the one-on-one exists to catch.
• CC-3 Inclusive practices and equitable opportunity distribution — who gets the stretch work and the visibility; growth paths distributed fairly.

1. Why should the one-on-one belong to the report, and where do the earliest attrition signals appear?
   Answer

   Because status can travel by other channels and what cannot — blockers, frustration, disengagement — surfaces only in a meeting the report controls. The earliest signals are withdrawal, cynicism and a stopped flow of ideas, visible across several one-on-ones to a manager who keeps notes.
2. A talented engineer asks for promotion before consistently showing the next level's judgment. What is the response, and what are the two failures?
   Answer

   Acknowledge the growth, walk through the level description against the evidence, name the gap specifically, and build a path from real experiences with a date to look again. The failures are promoting to retain and declining without giving anything actionable.
3. What distinguishes a real senior individual-contributor track from "senior engineer but more so"?
   Answer

   Distinct archetypes with their own scope and influence, equal standing to management, and recognition for the coordinating and quality-raising work that otherwise goes unrewarded. A manager who cannot describe the track cannot offer it honestly.
4. An engineer holds a competing offer and your manager wants to match it. What should the decision turn on? org
   Answer

   Value and fairness. If the person is underpaid, fix it because it is wrong; if they are not, a reactive match teaches everyone that threats work and distorts equity for those who did not threaten. Retention bought under duress rarely lasts.
5. One team's promotions are consistently faster than the others'. What are the two explanations, and which is more common? org
   Answer

   Either its people are consistently stronger or its manager is consistently more assertive in calibration. The second is more common, and shared criteria with written evidence are the check.
6. What is the strongest signal that someone is ready for a larger scope, and what are two signals that are often mistaken for it? exec
   Answer

   That they are already exercising the larger scope's judgment on real stakes. Tenure and ambition are the usual substitutes, and neither is evidence.
7. Two directors are ready for one vice-president seat. What do you owe the one who does not get it? exec
   Answer

   The truth about their path, genuine investment in it, and acceptance that they may leave. Inventing a title to avoid the conversation weakens the organisation and deceives nobody.
8. What is the quiet measure that an executive's leadership development is working? exec
   Answer

   Decisions that used to need the executive get made well without them, and the leaders they developed are developing leaders of their own.

• Andrew S. Grove, High Output Management (1983) — the one-on-one as the report's meeting; managerial leverage.
• Camille Fournier, The Manager's Path: A Guide for Tech Leaders Navigating Growth and Change (2017) — the management career arc, level by level; ladders.
• Will Larson, Staff Engineer: Leadership Beyond the Management Track (2021) — the senior individual-contributor track and its archetypes.
• Tanya Reilly, The Staff Engineer's Path: A Guide for Individual Contributors Navigating Growth and Change (2022) — senior individual-contributor leadership; glue work.