Job Scorecard
When To Use
The brief is "two jobs in one title", and everyone means something different by "senior". Use it after the intake meeting and before anything is written for candidates. It answers: what must this person achieve in year one, and what evidence would show they can do it?
When Not To Use
If the manager has not yet agreed what the job is, run Intake Meeting first; a scorecard built on a drifting brief only makes the drift look official. If you need the form interviewers fill in, that is Interview Scorecard, which uses these criteria.
Inputs
- The signed-off intake summary, or your intake notes
- The manager's answers on two or three times the job went notably well and badly
- Any existing job description, for reference only
If you have none of this, I start from the role title and three main tasks and mark the output as a first draft.
Approach
This is light job analysis as OPM describes it: list the tasks, then the competencies each task needs, and keep only criteria linked to a task. The evidence examples come from the critical incident technique (Flanagan, 1954, Psychological Bulletin): real moments when the work went well or badly, and what the person actually did. The role scorecard format is described generically. The failure it prevents: "senior" meaning years of service to one interviewer, scope to another and polish to a third, so three people assess three different jobs.
Workflow
- Ask at most three questions: the main tasks in order of time spent, two or three critical incidents (what happened, what the person did, the result), and who signs off the scorecard.
- Tasks to competencies: for each task, the knowledge, skill or behaviour it needs. Any criterion with no task behind it is dropped.
- First-year outcomes: 3 to 5 observable results, each with a measure as a [placeholder] the manager sets.
- Criteria: keep 4 to 6. For each, a one-line definition, what strong evidence looks like (drawn from the incidents), must-have or trainable, and the stage where it is assessed (screen, interview, work sample, reference).
- Wording pass: "culture fit", age-coded words and personality labels are replaced with the behaviour meant.
- Split test: if the outcomes pull in two directions, show both sets side by side and ask the manager to choose one job or open two.
Output Format
# Job Scorecard: [Role title]
## Purpose
[One sentence: why this role exists]
## First-Year Outcomes
| Outcome | Measure | By when |
|---|---|---|
| [observable result] | [measure the manager sets] | [month] |
## Criteria
| Criterion | Definition | Strong evidence looks like | Must-have or trainable | Assessed at |
|---|---|---|---|---|
| [criterion] | [one line] | [from a critical incident] | [must-have] | [stage] |
## Split Check
[One job, or the two competing sets of outcomes for the manager to choose between]
## Decision
[Hiring manager] approves the outcomes and criteria by [date]; [recruiter] uses them for every later step.
Done When
- Every criterion links to a named task
- There are 4 to 6 criteria, each assessed at exactly one main stage
- Every outcome is observable, with a measure or a [placeholder]
- No word describes personality, age or "fit"
Quality Bar
- The scorecard describes the job; it never rates, scores or compares a candidate.
- Strong evidence is written as what someone did, never as a trait.
- Trainable items stay off the list of must-haves and out of the screen.
- Measures are the manager's; I invent no targets or benchmarks.
- If two jobs hide in one title, the scorecard says so rather than blending them.
Next
Run recruit-salary-range (Salary Range Brief) to price the job before it is advertised.
About the makers
This pack is made by 90n (90n.ai): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).
1---2name: recruit-job-scorecard3description: Turns an agreed brief into a job scorecard with first-year outcomes, 4 to 6 job-related criteria with what evidence of each looks like, a must-have versus trainable split, and the stage where each criterion is assessed. Use for "run recruit-job-scorecard", "job scorecard", "role scorecard", "define the role", "two jobs in one title", "what does senior mean here", "hiring criteria for this role", "success profile", part of the AI for Recruiting Pack by 90n.4metadata:5 source: https://skillmd.com/skills/polar-bear-org/recruit-job-scorecard6---78# Job Scorecard910## When To Use11The brief is "two jobs in one title", and everyone means something different by "senior". Use it after the intake meeting and before anything is written for candidates. It answers: what must this person achieve in year one, and what evidence would show they can do it?1213## When Not To Use14If the manager has not yet agreed what the job is, run Intake Meeting first; a scorecard built on a drifting brief only makes the drift look official. If you need the form interviewers fill in, that is Interview Scorecard, which uses these criteria.1516## Inputs17- The signed-off intake summary, or your intake notes18- The manager's answers on two or three times the job went notably well and badly19- Any existing job description, for reference only20If you have none of this, I start from the role title and three main tasks and mark the output as a first draft.2122## Approach23This is light job analysis as OPM describes it: list the tasks, then the competencies each task needs, and keep only criteria linked to a task. The evidence examples come from the critical incident technique (Flanagan, 1954, Psychological Bulletin): real moments when the work went well or badly, and what the person actually did. The role scorecard format is described generically. The failure it prevents: "senior" meaning years of service to one interviewer, scope to another and polish to a third, so three people assess three different jobs.2425## Workflow261. Ask at most three questions: the main tasks in order of time spent, two or three critical incidents (what happened, what the person did, the result), and who signs off the scorecard.272. Tasks to competencies: for each task, the knowledge, skill or behaviour it needs. Any criterion with no task behind it is dropped.283. First-year outcomes: 3 to 5 observable results, each with a measure as a [placeholder] the manager sets.294. Criteria: keep 4 to 6. For each, a one-line definition, what strong evidence looks like (drawn from the incidents), must-have or trainable, and the stage where it is assessed (screen, interview, work sample, reference).305. Wording pass: "culture fit", age-coded words and personality labels are replaced with the behaviour meant.316. Split test: if the outcomes pull in two directions, show both sets side by side and ask the manager to choose one job or open two.3233## Output Format34```markdown35# Job Scorecard: [Role title]36## Purpose37[One sentence: why this role exists]38## First-Year Outcomes39| Outcome | Measure | By when |40|---|---|---|41| [observable result] | [measure the manager sets] | [month] |42## Criteria43| Criterion | Definition | Strong evidence looks like | Must-have or trainable | Assessed at |44|---|---|---|---|---|45| [criterion] | [one line] | [from a critical incident] | [must-have] | [stage] |46## Split Check47[One job, or the two competing sets of outcomes for the manager to choose between]48## Decision49[Hiring manager] approves the outcomes and criteria by [date]; [recruiter] uses them for every later step.50```5152## Done When53- Every criterion links to a named task54- There are 4 to 6 criteria, each assessed at exactly one main stage55- Every outcome is observable, with a measure or a [placeholder]56- No word describes personality, age or "fit"5758## Quality Bar59- The scorecard describes the job; it never rates, scores or compares a candidate.60- Strong evidence is written as what someone did, never as a trait.61- Trainable items stay off the list of must-haves and out of the screen.62- Measures are the manager's; I invent no targets or benchmarks.63- If two jobs hide in one title, the scorecard says so rather than blending them.6465## Next66Run recruit-salary-range (Salary Range Brief) to price the job before it is advertised.6768## About the makers6970This pack is made by 90n ([90n.ai](https://90n.ai)): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).