What actually happens
When you paste a CV and a job ad, both go to a large language model with instructions to do three things: pull out the key requirements and terminology the ad uses, check which of them your CV evidences, and return a score out of 100 along with the terms it could not find.
The concrete part of that is the count. When the free checker tells you it matched 11 of 19 key terms, those are 19 terms extracted from your ad and 11 found in your CV, and you can see the missing 8 for yourself. The score out of 100 is the model's judgement of how well the CV answers the ad overall, weighted toward the requirements the ad treats as essential. It is not the percentage.
What moves the number
- ▸Essential requirements. Terms the ad marks as required, or repeats, count for more than nice-to-haves.
- ▸Credentials and registrations. AHPRA numbers, licence classes, White Card, CPA status. In Australian hiring these are often checked first, so a missing one costs more than a missing skill word.
- ▸Terminology match. Using the ad's own words. A CV saying 'peripheral IV insertion' against an ad asking for 'cannulation' describes the same skill and reads as a gap to software.
- ▸Evidence, not assertion. 'Strong communicator' scores worse than a sentence showing what you communicated, to whom, and what happened.
- ▸Structure a parser can read. Text in tables, headers and images is frequently dropped by the software employers use.
How to read your score
Treat it as a band, not a measurement. Most untailored CVs land between 35 and 65 against a specific ad, which is not a judgement on your career, it is what happens when a general CV meets a particular role.
Above 75
The ad's language is well covered. Remaining gaps are usually nice-to-haves.
55 to 75
The core is there and specific requirements are missing or buried. This is the band where an hour of editing changes the most.
Below 55
The CV is not yet answering this ad. Often it is a good CV aimed at a different role.
What it cannot tell you
This is the part most scoring tools leave out, so it is worth being blunt about.
- ▸It is not a simulation of any particular employer's applicant tracking system. Workday, PageUp, Taleo and the rest behave differently, most do not publish how they rank, and none of them expose a score to a tool like this one.
- ▸It does not predict whether you will be shortlisted. Nothing can. Who else applied, whether someone internal was always going to get it, and whether the role survives the quarter are all invisible to your CV.
- ▸It does not judge whether you are good at the job, only whether this document evidences what this ad asked for.
- ▸It cannot tell whether what you have written is true. That part is on you, and it is the reason every suggestion is offered for you to approve rather than applied for you.
On the widely quoted statistic. You will read that 75% of CVs are rejected by ATS before a human sees them. It traces back to vendor marketing rather than published research, and CVelo does not repeat it. Applicant tracking systems are mostly databases with search over them. Wording matters because it decides whether you are findable and legible, not because a machine is silently binning you.
Common questions
Is the role match score an ATS score?+
It is the same idea as what other tools call an ATS score, but the name is deliberately different. An ATS score implies a number produced by a particular employer's applicant tracking system. CVelo has no access to any employer's system, and neither does any other tool that sells you a score. What is being measured is how well your CV evidences what this specific job ad asks for.
Does a high score mean I will get an interview?+
No, and be sceptical of anything that suggests otherwise. The score reflects the overlap between your CV and one job ad. It knows nothing about how many people applied, the internal candidate who was always going to get it, your referees, your interview, or the budget being pulled in week three.
Why did my score change when I re-ran the same CV?+
The score is produced by a language model reading your CV against the ad, not by a fixed formula counting words. Two runs on identical text can differ by a few points. Treat a score as a band rather than a precise measurement: the move from 45 to 70 is meaningful, the move from 68 to 71 is not.
Can I game the score by stuffing in keywords?+
You can raise the number that way, and it will hurt you. A CV packed with terms you cannot speak to falls apart in the first interview, and human readers notice keyword lists that do not match the work described underneath. Every suggestion CVelo makes is meant to be true of you; if one is not, do not use it.
What if the job ad is vague?+
The score is only as good as the ad. A short or generic ad gives the model little to extract, so the key terms will be thin and the score less meaningful. Where an ad is vague, the position description, if one is attached, is usually the better thing to paste.
See it on your own CV
One free check a day, no account and no card. Nothing you paste is stored.
Score my CV free