Methodology
The Six-Dimension Fit Score: how to know if a job is worth applying to
This page documents how ScoredWell scores a job posting: the six dimensions, their exact weights, the verdict bands, and the checks that stand between a strong score and the documents you send. Everything below describes the system as it runs — the numbers come from the code, not from marketing.
1. What a job fit score is
A job fit score is a 1–10 measure of how well a specific posting matches your actual experience across six weighted dimensions. It answers one question before you spend an evening on an application: is this role close enough to what you have actually done to be worth pursuing?
The score exists to serve a fit-first job search. Fit-first job search means scoring whether a role is worth applying to before you spend a single evening tailoring for it. The order matters: most job-search tools help you produce applications faster, which optimizes the wrong step. Volume does not fix aim. A score forces the decision — apply, stretch, or skip — before any writing happens.
The score is computed against your real profile: the résumé you uploaded and the preferences you set. The same posting produces different scores for different people, because fit is a relationship between a role and a person, not a property of the role.
2. The six dimensions
The Six-Dimension Fit Score weighs six aspects of a posting against your profile: skills, location and work eligibility, seniority, growth potential, compensation, and company prestige. Each dimension is scored 1–10, multiplied by its weight, and summed into the overall score. The weights total 100.
Two scoring rules apply across all six. First, 5 means neutral or unknown — it is not a politeness floor. Second, every scored role gets a named watchpoint: the one specific risk or stretch, even on a strong fit, because a score that only flatters is a score you stop trusting.
| Dimension | Weight |
|---|---|
| Skills Match | 30 |
| Location, Hours & Eligibility | 25 |
| Seniority Match | 15 |
| Growth Potential | 12 |
| Compensation | 10 |
| Employer Standing | 8 |
| Total | 100 |
Every dimension is scored in five defined bands
A number between 1 and 10 is easy to write down and hard to mean. So each dimension now has five written bands, and the scorer picks the band before it picks the number — and takes the lower of the two unless the posting earns the higher one. Two rules bind all six: a score of 5 means “the posting didn’t say” and is never held against the role; and a dimension can only be marked into its bottom bands on something the posting actually states — never on a hunch, and never on silence. Anchoring the scale this way cut the run-to-run movement of the score by about a quarter, and by half on the dimension that was worst.
2.1Skills Match · weight 30
Whether you can actually do this job — judged on capability, not vocabulary. The scorer is told that a posting’s list of fourteen “requirements” is a wish list, and that three to five of them are what the role really exists to do. It scores against those, and it is explicitly forbidden from scoring by keyword count: doing the same craft with a different tool is not a gap. The top of the scale means you have done this job; the middle means the posting doesn’t say enough to tell.
Example: Someone whose experience transfers into an adjacent role scores in the middle, not the basement — because the alternative is a score that punishes career switchers for not writing their CV in the job ad’s words. We will not do that, and our CV generator is built so it cannot invent the missing keyword either.
2.2Location, Hours & Eligibility · weight 25
Whether you can actually live this role — legally, geographically, and daily. Three things: your legal authorization to work in the country the posting is actually in (worked out from your profile, not guessed from the ad); whether remote, hybrid or onsite suits you; and the hours. Hours count as much as geography, because a role that pays well by running on someone else’s clock is a trade, not a free upgrade — you tell us in your profile which side of that trade you want to be on.
Example: A remote role you are eligible for, but which runs a graveyard shift to cover US hours, scores lower for someone who has said they can only work day hours — and is not marked down at all for someone who has said nights are fine. Where a posting says nothing about sponsoring, we score a neutral 5 and tell you the question is open, rather than assuming a no.
2.3Seniority Match · weight 15
Whether the role's level matches the seniority your experience evidences — in both directions. A role below your stated minimum is flagged, and a role clearly above what your track record supports is scored low too, because over-reach is the more common reason a recruiter screens an application out.
Example: A staff-level posting scores low for a candidate whose résumé evidences mid-level scope, and the score names that gap instead of flattering you.
2.4Growth Potential · weight 12
Whether this role moves you forward — in scope, market, or craft — measured against the targets you set, not a generic ladder. A step up on two of those scores at the top; one is strong; a genuine lateral move sits in the middle, because flat is not a failing. A smaller company is not automatically less growth, and a famous name is not growth at all.
Example: A lateral title with a broader remit, a bigger market, or a skill you are deliberately building scores higher than the same job under a different logo — and a prestigious company that would narrow your remit scores lower, however good the logo looks.
2.5Compensation · weight 10
Whether the listed pay aligns with what you are looking for — alignment, not generosity. A range far above your level is not a win; it usually means the role is more senior than the title says. When a posting lists no salary at all, pay is not scored at all: the dimension is dropped and the other five weights are rescaled, so an employer’s silence is never counted against the role.
Example: A posting with no salary range is scored out of the remaining five dimensions — you will see “not posted” rather than a half-empty bar, because we will not manufacture a weakness out of missing information.
2.6Employer Standing · weight 8
How solid and real the employer is — not how famous. This is the only dimension whose main job is to score DOWN, and it may only do so on something the posting actually shows: an unnamed company, a deterministic agency signal, or visibly recycled boilerplate. Never on a hunch. An employer you have simply never heard of sits in the middle — obscurity is not a defect, and most real employers are not globally famous. Standing is judged in your own market, not only in San Francisco.
Example: A national bank or a leading local fintech scores as well here as a household-name multinational. A posting with no named employer and no trace of a real business scores low — and at 8% weight, no logo can rescue a role that does not fit your skills.
3. The verdict bands — and why Skip is the point
A verdict is a plain-language band on top of the numeric score: Strong Fit, Moderate Fit, Stretch, or Skip. It turns a decimal into a decision. The bands are fixed, and each is inclusive at its lower bound — a 6.0 is a Moderate Fit, an 8.0 is a Strong Fit.
| Verdict | Score |
|---|---|
| Strong Fit | ≥ 8.0 |
| Moderate Fit | 6.0 – 7.9 |
| Stretch | 4.0 – 5.9 |
| Skip | below 4.0 |
A Skip verdict is not the product failing to help you. It is the product working. The alternative is the tailoring treadmill: the cycle of rewriting your CV for every posting — spending hours per application with no idea whether the role was ever a realistic fit. Every Skip is an evening you did not spend on the treadmill, and a Skip costs you nothing — you can still add the job to your tracker and apply anyway. The score advises; it never blocks.
A tool that scored everything Moderate-or-better would be easier to sell and useless to you. The verdict bands only mean something because the bottom one gets used.
4. What happens after a strong score
A strong score unlocks an application kit: a CV and a cover letter tailored to the posting, both as editable Word files. Each document must pass its own critique gate — a scored quality check — before the kit is marked Ready. If either fails, the whole kit does. A pre-apply company snapshot and a deeper interview-stage brief are available on demand, held to the same critique discipline.
Each gate scores a document against a fixed rubric, cheapest checks first:
- A deterministic code floor. Plain code — not AI judgment — checks the mechanical criteria: no banned filler phrases (“I am excited to apply”, “results-driven”, “team player”), CV bullets that lead with a real action verb and keep the source bullet’s metric, a curated skills list with no fluff entries, and a cover letter inside its word-count band. These checks cost nothing, give the same answer every time, and the AI cannot argue with them — a code result overrides anything a model says.
- The no-fabrication check. A code check rejects any number or claim that isn’t already in your résumé. Every rewritten CV bullet is traced to a specific source bullet, and every figure in the rewrite — digits or spelled-out quantities like “tripled” — must already exist in that source. An invented number fails the gate outright, regardless of how good the rest of the document is.
- A model judge for the subjective half. Criteria code cannot measure — does it sound like a person, is it specific to this company — are scored by an AI judge, and any borderline or failing result is re-adjudicated by a stronger model before it can fail. The judge grades; the code floor decides the non-negotiables.
The pass bars are fixed points on each rubric: 11/14 for the CV, 15/18 for the cover letter. A failing document is regenerated with the critique attached — at most twice. If it still cannot pass, the kit is marked Error and you are told so. Nothing below the bar is quietly shipped to you, because everything in a kit is something you might put in front of an employer.
5. What the score can’t tell you
The score measures fit, not luck. It reads the posting and your profile; it cannot see inside the company. Several things that decide real outcomes are invisible to it, and pretending otherwise would make the number less trustworthy, not more.
It cannot see an internal candidate the role was written for, a hiring freeze that leaves a posting up after the budget is gone, or the advantage a referral carries over any cold application. A Strong Fit means your experience genuinely matches what the posting asks for — it is a statement about the match, never a prediction about the outcome. Use the score to decide where your effort goes. What happens after you apply still depends on people and timing no rubric can read.
If you want this run against your own résumé, ScoredWell is free during the invite-only beta.