How the match score works

Save a CV and every job shows your fit: how many of the skills the posting names are on your CV, which ones, and which are not.

One number: the share of skills on your CV

"Strong fit · 80%" with "8 of 10 skills this job asks for are on your CV" underneath means the posting names ten skills we recognise and your CV mentions eight of them. The skills come from the full posting, extracted when we stored it, or from the text we hold where that extraction is missing. The same percentage is on the result row, in the panel, and behind the sort and the filter, so they cannot disagree.

Common aliases are matched on both sides, so K8s on your CV matches Kubernetes in the posting, and Postgres matches PostgreSQL. One skill is counted once, whichever way the posting spells it.

When your saved CV has been indexed, we also measure how close the whole posting is to your whole CV, inside our own database. That is never shown as your fit. It is used for one thing: ordering jobs with the same count when you sort by Best fit first.

Strong, good and partial

• Strong fit: three quarters or more of the skills, such as 3 of 4 or 8 of 10.
• Good fit: 55% to 74%, such as 2 of 3 or 6 of 10.
• Partial fit: 40% to 54%, such as 2 of 4. Shown in grey: it is not a warning.

Hovering or tapping the indicator on a row lists the skills on your CV and the ones that are not. Strong fits only in the filter bar keeps the first band.

When there is no score

Below 40% nothing is shown, on the row or in the panel, and 0% is never shown. Nor is anything shown when the posting names fewer than three recognisable skills: a one-skill posting could only score 0 or 1 of 1, and a confident zero reads as "you match nothing" when the truth is that we cannot tell.

This is common. Several boards send only a few hundred characters of description, and no work on our side recovers text the source never sent. Nothing is shown when no CV is saved, either.