Why this page exists
Most resume-scoring tools (Jobscan, ResumeWorded, EnhanCV) don't publish how they calculate their score. You upload your resume, you get a number, you decide whether it's credible. That's a problem: an ATS score without a methodology is just an opinion.
We chose the opposite. CVpass publishes here the exact weighting, the origin of the data, the biases we've identified and the limits we own up to. If you disagree with a sub-score, you can know precisely which criterion is responsible.
How we calculate the score (exact formula)
The ATS score is a whole number out of 100, calculated by this formula:
0.40 × score_motscles +
0.20 × score_miseEnForme +
0.20 × score_sectionsClaires +
0.20 × score_chiffres
)
Each sub-score (keywords, formatting, sections, numbers) is itself a number between 0 and 100, calculated from measurable and deterministic signals (no AI "feeling"). Details of each sub-score below.
When you accept an AI suggestion, the score is recalculated locally using this formula: score_before + (accepted / total_suggestions) × (100 - score_before). It's a linear approximation that assumes each suggestion contributes proportionally to the missing points. Not perfect but transparent.
The 4 scoring pillars
1. Keywords (40%)
Comparison of the resume's terms with the terms expected for the role (or with the pasted job posting if you use "Match posting" mode). We check:
- Presence of the exact terms (no synonyms: ATS don't make the connection)
- Density (1-3% of the resume's length, otherwise keyword stuffing is detected)
- Placement (resume title, summary, experience = high weight; interests = low weight)
- Lexical variation (presence of "JavaScript" AND "JS" if both are expected)
2. Formatting (20%)
- Single-column vs multi-column (ATS read columns in a zigzag)
- Standard font (Arial/Calibri/Helvetica/Geist) vs unusual
- Size (10-12pt body, 14-18pt headings)
- No floating text boxes
- No icons in section headings
3. Clear sections (20%)
- Presence of the 5 standard sections (Experience, Education, Skills, Languages, optional: Certifications)
- Exact headings recognized by ATS (not "My journey", "Who am I", "My talents")
- Reverse-chronological order for experience
- Dates in a recognizable format (MM/YYYY, MMM YYYY)
4. Concrete numbers (20%)
- Number of bullet points containing at least one figure
- Diversity of units (€, %, duration, team, volume)
- Consistency (revenue of €10M for a junior = inconsistent signal, negative scoring)
- Penalty for resumes with 0 figures (very common: 64% of the sample)
Where the numbers we cite come from
We cite several statistics on the site. Here is the origin of each, distinguishing external sources from our own internal measurements:
| Statistic | Source |
|---|---|
| Qualified candidates screened out by automated filters | Harvard Business School & Accenture "Hidden Workers" study (2021): more than 88% of employers acknowledge that qualified candidates are screened out because their resume doesn't exactly match the job posting's criteria |
| Canva resumes often poorly read at extraction | CVpass internal measurement: share of resumes detected as "Canva format" whose extraction loses a significant part of the content |
| 64% of resumes with no figures at all | CVpass internal measurement on the sample |
| Average score before: ~56/100 | CVpass internal measurement on the first analyses (before suggestions) |
| Average score after: ~65/100 (up to 87 when all suggestions are applied) | CVpass internal measurement; the after-score depends on how many suggestions are applied |
| Average gain: +8 points; up to +31 | Difference between before and after score (internal measurement); the maximum applies to the most reworked resumes |
| Most large French companies use an ATS | Market observation: large groups and mid-caps rely on ATS platforms (Workday, Taleo, SmartRecruiters, Teamtailor) to process applications; executive hiring is largely intermediated by APEC |
Our sample of analyzed resumes
When we say "1 900+ resumes analyzed", we mean all the resumes that were submitted to the tool and produced an ATS score. This number is updated in real time via our database and displayed dynamically on the site. You can reload any page: the counter reflects the current state.
Approximate composition of the sample:
- Geography: ~95% metropolitan France, ~3% Belgium/Luxembourg/French-speaking Switzerland, ~2% other
- Experience level: ~30% juniors (0-3 years), ~50% mid-level (3-10 years), ~20% seniors (10+ years)
- Sectors: tech (35%), sales/marketing (20%), HR/management (15%), finance/accounting (10%), other (20%)
- Resume format received: 75% PDF, 22% DOCX, 3% other converted formats
Acknowledged limit: this sample is biased in favor of active candidates seeking to optimize their resume. It does not represent the entire job market but rather the population interested in ATS optimization, which matches our product target.
The biases and limits we own up to
No scoring tool is perfect. Here is what we can't (yet) measure correctly:
- Writing quality: we don't grade style. A grammatically shaky resume can have a good score if the content is there.
- Career coherence: we don't judge whether a career path makes sense. That's the job of a career coach, not an ATS scanner.
- Role relevance vs posting: without a specific job posting provided, the score is based on the declared role and may over- or under-estimate depending on the actual position targeted.
- Specific ATS: our scoring is calibrated on the average behavior of the major ATS (Workday, Taleo, SmartRecruiters). A niche ATS may have different rules.
- Change over time: ATS change. Our scoring is a snapshot at the most recent date, not an eternal truth.
External sources cited
- Jobscan, annual studies on ATS behavior in the US/Europe
- APEC, data on the recruitment practices of French companies
- France Travail (formerly Pôle Emploi), job-market statistics
- INSEE, demographic and salary data by occupation
- Glassdoor, salary ranges corroborated with French data
- Indeed, analysis of job postings to calibrate the keywords expected per occupation
How AI is used (and what it doesn't do)
AI (OpenAI's GPT-4 model) is used to generate rewrite suggestions, not to calculate the score. Concretely:
- ATS score: 100% deterministic, calculated by our code from the 4 pillars above. No AI.
- Problem detection: 100% deterministic. We use rules (regex, structured parsing) to identify Canva resumes, poorly named sections, multi-columns, etc.
- Rewrite suggestions: AI. For each detected problem, we ask GPT-4 to propose a new wording that solves the problem while keeping your original meaning.
- Cover letter, interview coach, LinkedIn optimizer: 100% AI, but with constrained and auditable prompts.
This design is deliberate: if you disagree with your score, you can know exactly which criterion the scoring penalized you on. If you disagree with an AI suggestion, you can reject it without it changing your score (unless you rewrite it yourself to solve the detected problem).
More details on the confidentiality of your data during the AI analysis: how CVpass protects your resume.