How Epimoni Scores a LinkedIn Profile
The score is a weighted average of twenty-three criteria out of 100, and it only covers what your document actually shows. That last clause is the important one: a criterion the export cannot support is not scored lowest, it drops out of the calculation.

The principle: three marks, five sections
Every criterion gets a mark from 1 to 3:
- 3 β the information is present, precise and verifiable;
- 2 β present but vague, partial or asserted without evidence;
- 1 β absent, or too imprecise to verify.
Criteria are grouped into five sections, each worth a share of the final score:
| Section | Weight | What it measures |
|---|---|---|
| Experience | 26% | Roles described, results, context |
| Headline | 20% | The headline and its use of 220 characters |
| About | 20% | The section, its first two lines, its evidence |
| Search visibility | 18% | What lifts you in LinkedIn Recruiter |
| Skills and credibility | 16% | Alignment, certifications, languages, education |
A sixth section, CV / profile consistency, is added when you supply a reference CV. It compares titles, the scope described and the chronology.
Why experience weighs most
Because it is the only section that can carry evidence. A well-written headline announces; a described role demonstrates. It is also the emptiest section on an average profile, where most roles come down to a title and two dates.
Headline and About together carry 40% because they decide two different moments: the headline decides whether you are found, the About decides whether anyone continues.
What we do not score, and why
This is what sets the method apart, and it follows from one constraint: a PDF export contains only text.
It never contains the profile photo, the banner, the Open to work badge, recommendations received, the connection count, skill endorsements, the Featured section or recent activity.
So we do not score them. Telling someone who already has a photo that they lack one is the kind of error that disqualifies a tool on its first screen. These things matter, and we list them β but on an unscored checklist, with the reason each matters, to verify on your own profile.
Two cases deserve naming:
Skills. An export shows exactly three skills, whatever your profile carries. So we score how well they match your headline, never how many there are. A tool that tells you that you have "too few skills" from a PDF does not know what it is reading.
Spelling. We deduct nothing for typos or accents. The LinkedIn export can mangle them, and nothing lets us tell your mistake from an export artefact.
The third state: not assessable
This is the mechanism that protects well-written profiles.
Two different documents reach us. LinkedIn's official export carries experience descriptions in full. A printed page of your profile captures what is displayed: long text is cut at "see more".
In the second case your About section may stop mid-sentence. Scoring it would be absurd β we would be judging a fragment. Those criteria therefore become not assessable, drop out of both numerator and denominator, and your score covers what remains.
The consequence is deliberate: a cut-off section never lowers your score, it only reduces the number of criteria it is computed over. And we write that down β "scored on 19 of 24 criteria" β rather than leaving you to infer it.
A genuinely empty section does score 1. That is a finding, not an uncertainty. The difference between "absent" and "cut off" comes from the document type, not from a guess.
The custom URL, the one computed criterion
One criterion is not judged by the model: the custom URL. It is evaluated by a rule, from the address you enter before the analysis.
linkedin.com/in/jane-doeβ customised, top mark.linkedin.com/in/jane-doe-847a3b2β the suffix LinkedIn generates automatically, lowest mark.linkedin.com/in/janedoeβ customised but hard to read, middle mark.
That address is used for nothing else. We never connect to LinkedIn and never visit your profile online. We ask for it because the export's two-column layout wraps the URL across two lines, which makes it unreliable to read from the PDF.
What the score does not say
A score is a starting point, not a verdict. It measures the quality of a document at a moment in time, against public criteria. It measures neither your professional worth nor your fit for a particular job β that is a different exercise.
A profile at 80 is not "finished": at that level it is precisely the things outside the PDF that make the difference, the ones on the checklist.
Try it
Export your profile to PDF, then run the LinkedIn profile analysis. You get the twenty-three criteria in detail, the list of what could not be assessed and why, and your rewritten text ready to paste.
FAQ
How is a LinkedIn profile score calculated?
Each criterion is scored 1 to 3, criteria are grouped into five weighted sections, and the overall score is the weighted average of those sections out of 100. A criterion that cannot be assessed drops out of the calculation rather than scoring lowest.
Why is the profile photo not scored?
Because no PDF export of a LinkedIn profile contains the photo. Scoring something absent from the document would mean telling someone who already has a photo to add one. The photo appears on an unscored checklist instead.
What happens if a section of my profile is cut off?
It drops out of the calculation. Printing a LinkedIn page cuts long text at "see more"; the affected criteria are marked not assessable and the score covers the remaining criteria, which is stated explicitly.
Put it into practice
Apply the advice from this article with the free Epimoni tool: Analyse my LinkedIn profile (free)