Two recruiters can read the same senior engineer’s CV and reach opposite conclusions, and neither can fully explain why. A candidate screening scorecard fixes that by asking the same six questions of every CV, rating each answer from 1 to 5 against a written bar, and recording the result. The first decision in a technology hire then becomes something a hiring manager can inspect, question and compare.
This article sets out how the scorecard Masarat.ai uses works, what each dimension means for technology roles, and how to set the bar so the ratings mean something.
Why a candidate screening scorecard beats a quick CV read
A fast CV read is shaped by whatever the reviewer notices first: a familiar employer, a well-known university, a tidy layout. None of those tell you whether a backend engineer has run a service in production or whether a SAP consultant has led a full implementation rather than supported one.
A scorecard forces three things that a quick read skips:
- The criteria are agreed before anyone opens a CV, so the bar does not drift as the pile shrinks.
- Each rating needs a reason, which makes weak evidence visible.
- Candidates can be compared on the same terms, even when their CVs look nothing alike.
It also gives hiring managers something better than “the recruiter liked them”. When a shortlist arrives with ratings and short notes against each dimension, a CTO can disagree with a specific score instead of rejecting the whole list.
The six dimensions, rated 1 to 5
Our scorecard rates every CV on six dimensions. Each is scored against the job description for that role, not against an abstract ideal candidate.
Experience
Has the person done this kind of work, at this level, for long enough to be trusted with it? For a lead role that means evidence of owning outcomes, not only contributing to them. Years in a job title are a weak proxy; the scope of the work matters more.
Skills
Does the CV show the technical skills the role needs, used in real work? A list of frameworks in a skills box counts for little. A line explaining that the candidate migrated a monolith to Kubernetes and ran it on call counts for a lot.
Achievements
What changed because this person was there? Look for shipped products, systems stabilised, audits passed, platforms delivered. Vague claims such as “responsible for improving performance” earn a lower rating than a concrete result, even an unmeasured one.
Progression
Has the candidate grown in responsibility over time? Steady progression suggests people trusted them with more. A sideways move is not a red flag by itself, but the rating should reflect whether the career so far points towards the level you are hiring for.
Domain
Does the person know the environment the role sits in? A cybersecurity engineer from banking already understands regulatory pressure; a data engineer from telecom has probably worked with high-volume event data. Domain matters more for some roles than others, which is why it is weighted rather than fixed.
Application
How well does the candidate’s experience apply to this specific job, as written in the JD? This is the closest thing to an overall fit check, and it catches strong CVs that are strong for a different role.
Setting a clear bar for each rating
A 1-to-5 scale only works if everyone reads the numbers the same way. Before screening starts, write a short description of what a 3 looks like for each dimension on this role. A 3 means the candidate meets the bar. Anything above is evidence of more than you asked for; anything below is a gap.
A useful pattern for the anchors:
- 1: no evidence on the CV for this dimension.
- 2: some evidence, but below what the role needs.
- 3: meets the bar described in the job description.
- 4: clearly above the bar, with specific evidence.
- 5: exceptional for the level, and the evidence is hard to argue with.
Write the anchors in the language of the role. “Meets the bar for skills” for a ServiceNow developer should name the modules and integration work you expect. If the anchors are generic, the scores will be too. The work of writing a good technology job description pays off twice here, because the JD becomes the source for every anchor.
Weighting the score for the role
Not every dimension matters equally for every hire. For a head of data, progression and achievements usually carry more weight. For a specialist Oracle consultant, domain and skills may count for more. Agree the weights with the hiring manager at the brief stage, then combine the six ratings into a single weighted score.
The weighted score is a sorting tool. It helps decide who to look at first and where the evidence is thin. It is not a pass mark, and a candidate with one standout rating and a low weighted score can still deserve a conversation.
Where AI helps and where the recruiter decides
In the Masarat.ai process, AI-assisted screening parses each CV, matches it against the job description, proposes ratings and flags strengths, gaps and points to validate. That removes the repetitive part of reading a large pile of CVs. You can read more about how that stage works in our piece on AI CV screening.
A recruiter then reviews the scorecard. They adjust ratings where the AI has misread context, such as a job title that means something different in one company, and they validate every shortlist before it reaches the employer. The principle is simple: AI augments, humans decide. The scorecard is the record of that review, not a replacement for it.
After screening, the same candidate moves on to fit and technical interviews, and the candidate verification checklist confirms employment, work authorisation, qualifications, salary expectations and availability. The screening score travels with the candidate into the final report, so hiring managers can see how the first view held up. Our approach page shows each stage in order.
Common mistakes with a candidate screening scorecard
- Scoring against the last hire. The bar should come from the role, not from the person who left.
- Treating the weighted score as final. It ranks evidence on paper. Interviews and technical assessment test it.
- Skipping the notes. A rating without a reason cannot be challenged or learned from.
- Changing the weights mid-search. If the brief changes, rescore everyone, not just new applicants.
Frequently asked questions
Is a 1-to-5 scale detailed enough?
For screening, yes. The aim at this stage is to separate candidates who clearly meet the bar from those who do not, and to show why. Finer scales tend to create false precision, because reviewers cannot reliably tell a 6 from a 7 on a CV.
Can the same scorecard work across different technology roles?
The six dimensions stay the same; the anchors and weights change. A scorecard for a cybersecurity role and one for a product manager share the structure, but what “meets the bar” means is written for each role.
Do hiring managers see the scores?
They should. Sharing the scores and notes lets managers challenge a specific judgement and shows them how the shortlist was built. At Masarat.ai the scorecard sits inside the candidate report in the employer portal.
If you want to see a scorecard applied to a real vacancy, send us one technology role and we will run it through the full process, from screening to a structured report your hiring managers can act on.