Scoring methodology
The complete inner workings of the LinkedIn Profile Score: every dimension, weight, threshold and blind spot. If a signal is not on this page, the scorer does not use it.
The principles the scorer is built on
- Deterministic. The score is plain code: word lists, length bands, coverage arithmetic. Same input, same output, every time. No AI model grades your profile, so no model mood swings your number.
- Local. Parsing and scoring run in your browser. Your text is never uploaded, and nothing is stored. This is also why we publish no "average user" benchmarks: no data is collected from which an average could exist.
- Text in, text judged. The scorer sees exactly what you paste or upload. It cannot fetch your live profile, and things text cannot show (photo, banner, custom URL, Featured, connections, recommendations) are asked as checklist questions and marked self-reported in your report.
- No invented facts. Rewrites are assembled from your own lines. Where a fact is missing, the rewrite holds a [bracketed placeholder] instead of a fabricated number.
The ten dimensions and their weights
Weights total exactly 100. The rationale: headline, experience framing and keywords weigh most because first impressions and findability are where otherwise good profiles fail; the About pair carries the narrative; credibility, completeness and searchability carry the evidence and mechanics; the call to action is light but cheap to win.
| Dimension | Weight | Core question |
|---|---|---|
| Headline | 12% | Does the most visible line on the profile give anyone a reason to click? |
| About hook | 10% | Do the roughly 300 characters shown before "see more" earn the click? |
| About depth | 10% | Is there a full, structured, evidence-backed story? |
| Experience framing | 12% | Do roles read as owned outcomes or as pasted job descriptions? |
| Keywords | 12% | Do the terms you want to be found for actually exist in your profile? |
| Skills alignment | 8% | Does the skills list feed the filters recruiters actually use? |
| Credibility signals | 10% | Is there evidence a stranger could check or weigh? |
| Completeness | 10% | Are all the structural pieces of the profile in place? |
| Call to action | 6% | Does the profile tell an interested reader what to do next? |
| Searchability | 10% | Is anything mechanically blocking search from finding you? |
Every signal, dimension by dimension
This inventory renders from the same file the engine imports. Each dimension scores 0 to 10 from these checks, then contributes its weighted share to the 0 to 100 total.
Headline (12%)
Does the most visible line on the profile give anyone a reason to click?
What it checks:
- Length used of the 220 available characters (60+ scores best, under 25 is penalized)
- Multi-part structure using separators like | or commas
- The bare "Title at Company" pattern with nothing else (penalized, with a rewrite built from your own text)
- At least one concrete number
- Audience or outcome language such as "help" or "for"
- Filler-word density against our published buzzword list
- More than two emoji (small penalty)
- Stylized unicode pseudo-fonts (heavy penalty: they break search matching and screen readers)
What it cannot see: How the headline performs in real feeds; Your industry norms.
About hook (10%)
Do the roughly 300 characters shown before "see more" earn the click?
What it checks:
- Generic openers from our published list ("I am a passionate...", penalized, with your strongest sentence suggested as a replacement)
- Opening with a number or a direct question (rewarded)
- At least one concrete marker in the first 300 characters
- Whether the reader is addressed ("you", "your")
- A first sentence longer than 200 characters (penalized)
What it cannot see: Actual expand rates; LinkedIn does not publish them.
About depth (10%)
Is there a full, structured, evidence-backed story?
What it checks:
- Length bands: 800 to 2,000 characters scores best; under 400 reads as a placeholder; near the 2,600 limit suggests trimming
- Paragraph breaks (a wall of text over 600 characters is penalized)
- Three or more concrete markers such as numbers, percentages or amounts
- Heavy filler density (more than 3 buzzword hits) penalized
What it cannot see: Whether the claims are true; we grade specificity, not truth.
Experience framing (12%)
Do roles read as owned outcomes or as pasted job descriptions?
What it checks:
- Description volume (400+ characters across roles scores the base)
- Share of lines starting with action verbs from our published list (30%+ rewarded)
- Duty-speak phrases like "responsible for" (penalized, each with a verb-first rewrite of your own line)
- Concrete outcome markers (numbers) across the section
What it cannot see: Role seniority or gaps; we do not parse dates.
Keywords (12%)
Do the terms you want to be found for actually exist in your profile?
What it checks:
- You provide up to 5 target terms; without them this dimension scores a neutral 5 and says so
- Coverage per term, weighted by field: headline 40%, skills 25%, About 20%, experience 15%
- Terms missing from the headline specifically (flagged first)
- Stuffing: any term appearing more than 8 times across headline and About (penalized)
What it cannot see: LinkedIn’s real ranking formula; field weights here are our editorial model of relative importance, not LinkedIn documentation.
Skills alignment (8%)
Does the skills list feed the filters recruiters actually use?
What it checks:
- Count bands: 0 scores zero, 15 to 50 scores best, over 50 suggests curation
- Overlap between your target terms and the skills list (rewarded, missing overlap penalized)
What it cannot see: Endorsement counts or skill assessments.
Credibility signals (10%)
Is there evidence a stranger could check or weigh?
What it checks:
- Density of concrete markers across About and experience (5+ scores best)
- Presence of certifications, awards or publications text
- Presence of education
- Currency or percentage figures anywhere
- Visible recommendations, self-reported by you (we cannot see them from pasted text)
What it cannot see: Endorsements, recommendation contents, follower counts.
Completeness (10%)
Are all the structural pieces of the profile in place?
What it checks:
- Headline, About, experience, education and 5+ skills present in the text
- Photo, banner, Featured section and 500+ connections, each self-reported via the checklist because pasted text cannot show them
What it cannot see: Photo quality; we only know whether one exists, and only because you told us.
Call to action (6%)
Does the profile tell an interested reader what to do next?
What it checks:
- Call-to-action phrases from our published list in the About section
- A visible contact route such as an email address or booking link
- A filled Featured section (self-reported)
- "Open to" language in headline or About
What it cannot see: The Open To Work setting itself; LinkedIn does not expose it in text.
Searchability (10%)
Is anything mechanically blocking search from finding you?
What it checks:
- Custom profile URL (self-reported)
- Location present (a heavily used recruiter filter)
- A clean name field: no emoji, titles or pseudo-fonts
- Headline containing at least one of your skills or target terms
- Stylized unicode anywhere in headline or About (penalized again here: it breaks matching)
- 5+ skills and body text reinforcing your terms
What it cannot see: Your actual position in anyone’s search results.
Grades and tiers
Dimension scores map to grades on this scale:
| Grade | Score range (0 to 10) |
|---|---|
| A+ | 9.5 to 10 |
| A | 8.5 to 9.4 |
| B+ | 7.5 to 8.4 |
| B | 6.5 to 7.4 |
| C+ | 5.5 to 6.4 |
| C | 4.5 to 5.4 |
| D+ | 3.5 to 4.4 |
| D | 2.5 to 3.4 |
| F | 0 to 2.4 |
The weighted total maps to four named tiers:
| Range | Tier | Meaning |
|---|---|---|
| 0 to 44 | Starting out | The basics are missing or unclear. Fixing completeness and the headline usually moves this score fastest. |
| 45 to 64 | Building | The structure is there but the profile reads generic. Specifics, proof and keywords are the gap. |
| 65 to 84 | Strong | A clearly positioned profile. The remaining points live in credibility signals and sharper framing. |
| 85 to 100 | Standout | Specific, credible and easy to find. Maintenance mode: keep proof fresh and keywords current. |
Quick wins and time estimates
Every fix carries an impact tag (High, Medium, Low) reflecting the weight of the dimension it improves, plus a rough minutes estimate based on what the edit typically involves. The Top 3 quick wins are the fixes ranked by impact first, effort second. The estimates are deliberately rough and labeled as such; a headline rewrite is 15 minutes of typing and sometimes three days of thinking.
The Straight talk voice
Every finding ships with two phrasings: a professional one and a blunt one. The toggle swaps voice only; scores, findings and data are identical in both. It exists because some people hear "consider strengthening your opening line" and act on "your first line is a waiting room". Same diagnosis, two bedside manners.
Known limitations, stated plainly
- It grades presentation, not truth. A confident lie can score well on text signals; we cannot verify claims and do not pretend to.
- It cannot see LinkedIn internals: impressions, recruiter views, SSI, feed reach. No paste-based tool can, whatever their landing pages imply.
- Keyword field weights (headline 40%, skills 25%, About 20%, experience 15%) are our editorial model of relative importance, not LinkedIn documentation. LinkedIn does not publish its ranking math.
- Word lists are English-first. Profiles in other languages will score less accurately, especially on opener and duty-phrase checks.
- PDF parsing is heuristic; LinkedIn's export interleaves columns. The widget shows every parsed field for correction before scoring.
- Checklist items are self-reported. Answer them wrongly and that part of the score is wrong with you; the report marks those dimensions.
Who built this, and how
This scorer was built by the small team behind this site, written from scratch for it: the signals, weights, word lists and rewrite rules are original work, informed by publicly available LinkedIn documentation and years of writing and fixing professional profiles for real people. No LinkedIn insider data was used, because we have none; no competitor's scoring logic was copied, because the entire point of this site is that you can read ours.
We hold the method to the same rule as the marketing: claims must be checkable. That is why the tables on this page render from the same code the engine runs, why the limitations section exists, and why there are no star ratings under this paragraph. When we change a signal or weight, this page changes in the same commit.
Questions or disagreements
If a signal seems wrong, a threshold unfair, or a limitation missing, tell us through the contact page. Being argued with, with specifics, is how this method improves.