How to Write a Job Description

Almost every employment problem we get called into turns out to sit downstream of the same gap. Nobody had written down clearly enough what the job was.

The job description is the obvious place that gets written down, and it is the document nobody trusts. Leaders treat it as paperwork. Candidates skim it. HR files it next to the handbook. So it sits there, doing none of the work it could do, while the company argues about performance problems that were never defined in the first place.

Built right, though, the job description is the sharpest tool a leader has, because it is the only written agreement about what the job is. It is the standard you hire against, onboard against, review against, pay against, and part ways against. Five jobs, one document, and most companies retire it after the first.

Why job descriptions fail

Five failure modes, and most descriptions have all five.

They list activities instead of outcomes. Most read like a task dump. Manages subcontractors. Oversees scheduling. Prepares reports. None of that says what has to be achieved. You can manage subcontractors all day and still have late schedules, safety violations, and budget overruns.

They are written in isolation. The role reads as a part in a machine, not a position in a story. Candidates notice. So do employees, especially the ones trying to work out whether they are succeeding.

They sell without unselling. Fast-paced environment. Exciting growth. Every role has rewards and costs, and leaving out the costs is a promise you cannot keep past the first month.

They never define success. Manage the schedule. Be detail-oriented. That is not a scoreboard. Without one, a strong performer can feel like they are failing, and a weak one can be certain they are doing fine.

They do not age. Written once, forgotten. The company grows, the systems change, the projects change, and the description stays where it was. A living business needs a living document.

Design the role before you describe it

A description is the record of a decision, so the decision has to come first. Most companies skip to the document because the document is what the job board asks for, and then spend a year managing a role nobody designed.

Designing the role means answering four questions before you open a template. Why does this role exist. What must it achieve. How will anyone know whether it did. What will the work be like on a Tuesday in February.

Answer those and the description writes itself. Skip them and you get a task list, because a task list is what you fall back on when you have not decided what the job is for.

Connect the role to what the company is for

Start with the question that puts the role in context: if this role disappeared tomorrow, what part of our mission or our reputation would suffer most?

Scheduling trades is what a superintendent does. What a superintendent is for is safety, craftsmanship, and client trust, which together are most of what the company has promised anyone. An accounting manager balances books in order to safeguard resources, so that excellent work does not arrive with a financial surprise attached.

Write that connection into the description in a sentence or two. Three prompts get you there:

• Mission. How does this role help fulfill why we exist?
• Vision. How does it move us toward the future we are building?
• Values. How should this role embody the way we work?

If the company's mission is delivering high-end homes that honor both design intent and community trust, the superintendent's connection statement is: this role ensures every build meets design intent, safety expectations, and community standards, protecting both the client's dream and the firm's reputation. If one of the company's values is hungry for humility, the project manager's is: this role models humility by resolving conflict constructively and owning mistakes early, setting the tone for the project team.

A description without that connection reads as transactional, because it is. The role is a set of duties in exchange for money. With it, the reader can see what the work is for.

Write outcomes, not activities

Tasks describe what someone does. Outcomes describe what has to be achieved. The difference is the difference between measuring effort and measuring contribution.

Outcomes are also what keeps the interview finishable. If you intend to interview to the description, and you should, a legalistic mountain of possible activities gives you a twenty-hour interview process. Outcomes give you something you can test.

Every outcome fits one formula: contribution, then result, then impact.

Task: Supervise subcontractors.

>

Outcome: Subcontractors complete work on schedule with zero preventable safety incidents, protecting client satisfaction and profitability.

The second version is measurable, and it tells the person in the role what they are for.

Five kinds of contribution

Most descriptions describe one kind of contribution, usually operational, and leave the other four to chance. Write a success statement in each:

• Strategic. How does the role advance long-term goals? Projects strengthen the firm's reputation for craftsmanship and repeat business.
• Operational. What has to be delivered day to day? Projects are consistently delivered on time, on budget, and documented with zero preventable safety incidents.
• Relational. How should people experience this role? Clients receive proactive communication and trust that issues surface early.
• Cultural. How should this role model the company's values? Gives credit, takes responsibility, reinforces humility in team interactions.
• Developmental. How does this role leave the business better than it found it? Mentors assistant project managers so they are ready for the next step.

Take any line from an old description and run it through all five. One vague task becomes five layers of clarity, and you find out quickly which of the five you have never thought about for this role.

Sell the role, then unsell it

Most descriptions are written like brochures, and a brochure attracts people who are excited about the brochure.

Sell the role honestly. Name what makes it worth doing, which is usually the visible projects, the growth, and the real effect on the company. As a project manager, you will deliver homes that stand as a visible legacy of craftsmanship in this community.

Then unsell it just as honestly. Name the stressors, the constraints, the parts nobody enjoys. Demanding clients with high expectations. Budget pressure that fights design intent. Leading a team through setbacks and schedule crunches. This role requires navigating competing priorities under tight deadlines while protecting quality and client trust.

Three things happen when you do both. Candidates trust a leader who is upfront. The people who are put off by the honest version were not going to thrive in the real version. And the person who accepts knows exactly what they signed up for, which is most of what commitment is made of.

Skip the unsell and you have written false advertising. The new hire finds the hard parts in month two, concludes they were misled, and either checks out or leaves. That is an expensive way to learn that honesty at the front saves pain at the back.

Define what winning looks like

People cannot hit a target they cannot see, and most descriptions never say what winning is.

• First 30 days. What foundational wins have to be in place? Relationships built, systems learned, processes understood.
• First 90 days. What measurable contribution should be visible? Accurate reporting delivered, client trust established, team direction clear.
• First 365 days. What long-term impact should be undeniable? Projects completed profitably, safety record intact, repeat business secured.

An outcome says what has to be true. A success signal adds how you will know and when. So the same line grows one step: outcome, then evidence, then impact.

Outcome: Project schedules are consistently reliable.

>

Evidence: Milestones tracked with fewer than two missed deadlines in the first year.

>

Impact: Client confidence and repeat business are protected.

This is also the most useful thing you can give a candidate, because it lets them answer the only question that matters in a career move. Can I win here? The ones who cannot picture it filter themselves out, which spares you both the year it would have taken to find out.

The one-page template

Every element above fits on one page, and one page is the target. If the finished description runs past two, you have started listing activities again. Copy this structure for any role in the business:

• Role title
• Reports to
• Mission, vision, and values connection statement
• Outcomes by contribution category: strategic, operational, relational, cultural, developmental
• Sell the role
• Unsell the role
• Definition of success: 30 days, 90 days, 365 days
• Performance review criteria
• Accountability anchor

Every number you see in the examples below is a placeholder. A 3% budget variance is not an Ambassador Group benchmark, it is a stand-in for whatever your business can carry, and adopting someone else's threshold is how a scoreboard stops meaning anything.

The last line of the template is the one most companies leave off. An accountability anchor is a short statement of the role's non-negotiables, written at the start, so that a hard conversation two years later has something to reference other than memory.

The template's real job is to stop the document drifting into either of the two places job descriptions go to die: HR compliance paperwork, or marketing copy. It forces four answers at once. Why the role exists, what it must achieve, how success is measured, and what the work is really like.

Two worked examples

Here are two roles from high-end residential construction, written the way this guide argues for. Notice that the software shows up as evidence rather than as the outcome. Nobody was ever hired to use Procore.

Project Manager, high-end residential

Reports to: Director of Construction

Why the role exists. Design intent and client trust either survive a project or they do not, and the project manager is where that gets decided.

• Strategic. Finished projects bring the next one. Architects who specified this build ask for the same team again.
• Operational. Projects land on schedule and inside budget, and the project record is complete enough that a decision made in March can be reconstructed in December. Evidence: Procore current within the week, MS Project matching field reality rather than the original plan, Bluebeam carrying the markups the design team worked from.
• Relational. The client hears about a problem from the project manager first. Architects and consultants get an answer inside a day, including the answer that says we do not know yet.
• Cultural. Mistakes surface while they are still cheap, and credit goes outward.
• Developmental. One assistant project manager is ready to run a job alone by the end of the year.

Definition of success. At 30 days: proficient in the company's systems, trusted by the client, first financial and schedule update delivered clean. At 90 days: fewer than two missed milestones, budget variance inside the company's tolerance, client updates generated from the record rather than assembled by hand. At 365 days: delivered on time and on budget with design intent intact, no preventable safety incidents, and a junior who has visibly grown.

Accountability anchor. On-time and on-budget delivery. Zero preventable safety incidents. A project record a stranger could audit. Client confidence held through the hard weeks, not only the easy ones. Missing these is what makes the role a poor fit, whatever else is going well.

Superintendent, high-end residential

Reports to: Project Manager, Director of Construction

Why the role exists. The jobsite is the only place the company's standards are visible to a client, an inspector, or a neighbor. The superintendent sets what they see.

• Strategic. The site itself is the sales asset. A clean, organized, safe job is why the next client tours this one.
• Operational. Work goes in to plan, code-compliant, without rework, and the trades stay out of each other's way. Evidence: daily logs entered the day they happened, inspections logged, RFIs answered before they become schedule.
• Relational. Inspectors, neighbors, and HOAs get treated as people whose cooperation the client is going to need again.
• Cultural. The site is clean because the superintendent cares, not because a walkthrough is scheduled.
• Developmental. Assistant superintendents get coached in the moment on the thing that just happened, and field insight makes it back into how the company plans the next job.

Definition of success. At 30 days: control of the site established through safety, order, and cleanliness, systems proficient, first client walkthrough delivered against accurate logs. At 90 days: milestones hit with fewer than two slips, no preventable safety violations, the record current without being chased. At 365 days: delivered on time and on budget with design intent realized, coordination steady, safety record clean, and an assistant ready for more.

Accountability anchor. Safe, organized, compliant jobsites. On-time and on-budget execution. A field record leadership can trust without verifying it. Client, inspector, and neighbor confidence intact. Any one of these going missing is what ends the match.

Interview to the description

A strong description is worth nothing if the interview is run on instinct. The full mechanics of that belong to designing the interview process; what follows is only the part the description itself decides.

Name the interview team before the first candidate. Every role deserves a named set of interviewers, not whoever is free that Thursday. The scramble is what produces four people asking the same three questions and nobody asking the hard one.

Give each interviewer a lane. Divide the five contribution categories across the team instead of asking everyone to cover everything, so no risk is owned by everyone. For a project manager, that might be an executive leader on strategic alignment, an operations director on budget, schedule, and technical capability, a client-facing leader on relational and cultural fit, and a future peer on developmental contribution and team compatibility. Each interviewer has a category, and the description supplies the questions.

Test for alignment, not just capability. The description already says how the role connects to mission, vision, and values, so the interview has something specific to probe. Does the candidate care about the purpose or the paycheck. Do they have the energy for where the company is going. Do their instincts match how this company expects people to treat each other. If humility is a value, ask how they handle a mistake and who gets credit. If stewardship is a value, ask how they have managed someone else's resources.

Prepare in advance. Give every interviewer a packet holding the description with outcomes highlighted, the categories they own, and the evidence they are responsible for bringing back. An interviewer who first reads the description in the parking lot is having a conversation, not running an evaluation.

The payoff is evidence you can compare, which is what makes an honest debrief possible at all. Four interviewers who each own a category come back with four parts of one picture. Four interviewers who each formed a general impression come back with an argument.

Onboard to the description

You already wrote the onboarding plan. The definition of success was written to tell a candidate what winning looks like, and that is also the new hire's first-year map.

So use it as one. Walk through it in week one, with the person, out loud. The outcomes they own, the evidence that will show they own them, and when you will look at it together. It is a short conversation, and it is the one that stops month four being a surprise to either of you.

What you are avoiding is the common version, where the description does its job during recruiting, disappears at the offer, and the new hire spends ninety days inferring the standard from what their manager reacts to.

Manage to the description

Most job descriptions stop working the day someone signs one.

The document took weeks. You argued about outcomes, cut the task list twice, and got honest about the parts of the role nobody enjoys. Then the candidate accepted, you closed the file, and it went into a folder nobody opens again until the position turns over.

Six months later you sit down for a review and work from memory. One recent win. Two irritations. A general read on attitude. That conversation reveals more about you than about the person across the table. Without the standard they wrote in front of them, a leader is working from recall and calling it measurement.

What fills the vacuum

Watch what replaces a written standard. Recency comes first, so whatever happened in the last three weeks outweighs the ten months before it. Visibility follows, because the project manager down the hall gets read more generously than the one running a job forty minutes out. Then temperament, which scores well on a document that never mentioned it.

None of that is malice. It is what a leader reaches for when the measuring stick is missing. We sat in a review where a superintendent was marked down for communication. Asked what the standard was, the leader said he did not return calls quickly enough. His job description said nothing about response time. He had been graded against a rule nobody had given him.

Turn the outcomes into a scorecard

If the role was written with outcomes instead of activities, the conversion is mechanical. Give each outcome three things. A measure, so it can be observed rather than argued. A cadence, so it gets checked on a rhythm instead of at renewal. An owner, usually the person in the role and occasionally you.

Projects close within budget and on schedule becomes a line with a number attached, reviewed monthly, owned by the project manager. An outcome about developing the field team becomes a count of foremen ready for their next step, reviewed quarterly, owned by both of you.

This needs no new system. The definition of success already set the cadence. Keep running it past the first year, because the first year sets the rhythm the rest of the job follows.

Coach against the gap, not the person

A scorecard changes the shape of a hard conversation, because two documents are on the table instead of one opinion.

The question moves from "are you doing well" to "here is the outcome you own, here is where the result sits, what is in the way." A person can answer that without defending their character. Sometimes the gap is skill and the fix is training. Sometimes a resource never arrived. Sometimes the outcome was written for a role that no longer matches the work. All three are findable. None of them surface in a conversation about attitude.

This is also where a problem stays cheap. A gap you name in month four is a coaching conversation. The same gap in month fourteen has hardened into a reputation.

When it does not work out

No leader enjoys a termination conversation. The hardest part is not the decision, it is having no anchor for it. If expectations were vague from the start, the decision feels subjective to you and unfair to them, and both of you are right.

A description written to outcomes changes that, in four ways. Everyone knew what winning looked like from day one. The conversation is about outcomes not delivered rather than personal shortcomings. The employee can see that expectations were transparent and measurable. And the company's legal and relational exposure drops, because there is a written record instead of a recollection.

Start with support. Point back to the outcomes as benchmarks early, while there is still time to change the result. The description defined safety as a non-negotiable. Let us talk about this quarter's incidents and what support you need to correct them.

Document the gaps. Keep a record of outcomes consistently missed. This is not building a case. It is making the performance story visible to both of you, which is the only way the person gets a real chance to change it.

Anchor the decision in the standard. If improvement does not come, the conversation references the document rather than your patience. The description required on-time and safe delivery. Despite coaching and support, those outcomes were not achieved, and that makes clear the role is not a fit.

Termination should never be a surprise. A description written properly turns the worst conversation in the job into an honest alignment check.

Rewrite it as the role grows

The document should not survive three years unchanged. If it does, either the company stopped growing or nobody is reading it.

When someone clears an outcome consistently, raise it or retire it and write the next one. That edit is the most honest development plan available, because it is specific, it is written down, and it changes what gets measured next quarter. A promotion becomes a new set of outcomes, not a new title and a vague expectation of more. Compensation gets easier for the same reason. When pay moves against outcomes the person can see, the conversation stays on scope instead of drifting into worth.

The job description is the only written agreement about what the job is. It belongs on the table every time you hire, onboard, evaluate, coach, promote, or pay someone.

You wrote the standard. Whether you hold yourself to it is the one part nobody else can decide.