Your match score is not your chance of getting hired
It measures your document. It cannot measure the room.
· 5 min read · CV/Resume
You paste your CV and a job description into a tool, and it returns a number. 87% match. The number feels like a verdict — like a probability, like odds. It isn't either of those things, and treating it as one will cost you.
A match score measures one thing: how much the text of your CV overlaps with the text of the advert. That's a useful measurement. It is simply not the same measurement as how likely am I to get this job, and the gap between the two is where most of the real hiring decision lives.
What the number actually is
Every match score works roughly the same way. It extracts requirements, skills, titles and keywords from the job description, looks for corresponding evidence in your CV, and expresses the overlap as a percentage.
So a high score is a genuine signal — it means a screener searching for the obvious terms will find you, and that your document is talking about the same job the advert is. That's worth having. It just isn't a forecast.
Six things no score can see
1. Who else applied
Hiring is comparative, not absolute. There is no threshold you cross to get hired; there's a pile you get ranked within. An 87% is excellent against a field topping out at 70% and irrelevant against three people at 95%. Your score is calculated in isolation from the only thing that determines the outcome.
2. The unwritten requirements
Job descriptions are written by committee and rarely describe the actual job. The real requirement might be "can survive our head of product," or "has done this at our exact scale," or "starts in three weeks." None of that is in the text, so none of it is in your score.
3. Which requirements actually matter
Most scoring treats requirements as roughly equal, or weights them by crude signals like where they appear in the ad. Real screening doesn't work that way. One genuine must-have that you're missing outweighs ten nice-to-haves that you meet — and a score that averages across all of them will read 80% while a human reads "no."
4. Whether a person already has the job
An uncomfortable share of posted roles have an internal candidate, a former colleague, or a referral already in motion. You can score 99% against a role that was functionally filled before it was posted, and no tool can see that.
5. How you got there
A warm introduction changes your odds more than any edit to your CV. The same document, from the same person, converts at wildly different rates depending on whether it arrived through a portal or from someone the hiring manager trusts. Score is blind to distribution, and distribution is often decisive.
6. How you'll perform in the room
This is the big one. The score evaluates a document. Offers are decided by conversations — how you frame your experience, whether your stories hold up, whether you sound like someone they want to work with. A CV gets you the meeting. The meeting gets you the job, and nothing about the document predicts it.
The false precision problem
There's a deeper issue with the number itself. A score of 73% implies a precision that nobody actually has. It suggests the tool distinguished 73 from 74 — that there's a real, measurable difference between them.
There isn't. These systems are estimating semantic similarity between two pieces of text, and that estimate is nowhere near precise enough to justify two significant figures. The decimal point is doing rhetorical work, not mathematical work. It makes a rough judgment feel like a measurement.
Which is why honest ranges beat false precision. "You cover 6 of the 9 stated requirements, and the 3 you're missing include 2 marked required" tells you something true and countable. "68%" tells you something smooth, confident, and partly invented.
How to actually use a score
Stop reading it as a prediction and start reading it as a diff.
The number is nearly useless. The breakdown underneath it is where the value is, because it's a list of specific things the advert asks for that your CV doesn't currently evidence. That's a to-do list, and each item has exactly one of three honest answers:
- I've done this, it's just not on the page → add it, with a number attached. This is the most common case by a wide margin.
- I've done something adjacent → name the adjacent thing explicitly rather than hoping it's inferred.
- I haven't done this → decide whether it's central. If it isn't, apply anyway and address it directly. If it is, this is a stretch application and should be treated as one.
Read that way, a low score isn't a rejection and a high score isn't a green light. Both are just information about your document.
What we do about it
labor.quest produces a match analysis too, so this applies to us as much as to anyone.
Our position is that the percentage is the least interesting output. What matters is which specific requirements you address, which you don't, and which of the gaps are real versus cosmetic — countable things you can act on, rather than a smooth number that feels authoritative and can't be. We'd rather tell you "4 of 7, and here's what's missing" than hand you a 61% and let it imply a confidence the underlying analysis doesn't support.
The honest summary: a match score tells you whether your CV is speaking the right language. It says nothing about who else is in the room, what they're really looking for, or how you'll do when someone finally asks you a question.
Fix the document. Then go and prepare for the part that actually decides it.
All posts