CV Screening Criteria Matrix: Must, Nice and Exclusion

What's Included in This CV Screening Criteria Matrix

Most screening problems start before the first CV is read: the criteria are vague, or they change per candidate. This Excel workbook makes you decide first. It contains:

• A blank matrix with 12 criteria rows and five candidate columns, ready to copy per role.
• Four filled example roles: registered nurse, software engineer, warehouse operative with forklift licence and inside sales.
• Automatic scoring: must criteria not met, unclear, exclusions triggered, weighted nice points, fit score and a suggestion.
• A fairness sheet with ten notes and their legal basis.
• A guide sheet that explains the method in six steps.

Key Features That Save You Time

• Three criterion types. Must (cannot do without), nice (weighted 1 to 5) and exclusion (a factual reason that makes the hire impossible).
• Drop-down lists for type, weight and rating keep entries consistent.
• Adjustable thresholds for “invite” and “follow-up or pool”.
• “Unclear” as a rating. A must criterion you cannot judge from the CV leads to a question, not a rejection.
• Yellow input cells. Everything else calculates itself.

How to Use This CV Screening Criteria Matrix

1. Copy the blank sheet for your role and write down must, nice and exclusion criteria before you read any CV.
2. Give each nice criterion a weight from 1 to 5 and freeze the thresholds.
3. Rate each candidate: must with Yes, No or Unclear, nice with 0, 1 or 2, exclusion with Yes or No.
4. Read the suggestion as a sorting aid and look at every case yourself.
5. Record the result and delete data after your retention period.

Why This CV Screening Criteria Matrix Works

It separates what a candidate must bring from what merely helps, and it forces the weights to be set before the names are known. In our conversations with recruiting teams, the most common mistake in searching and screening is to set requirements as nice-to-have when they are must, or the reverse. The matrix makes that choice visible.

It is also built for fairness. Exclusions have to be factual, proxies such as photo, name or gaps in the CV are named as traps, and the decision stays with a person, in line with Art. 22 GDPR. Recruiting AI is classed as high-risk under Annex III of the EU AI Act, with deployer obligations from 2 December 2027. For AI-supported screening, read what AI resume screening can and cannot do.

Who Should Use This

• Recruiters who sort many applications and want comparable scores.
• Hiring managers who agree criteria with HR before the process starts.
• Teams that want a documented, fair first round.

If you want screening to run on your applications automatically, see CV screening.

A Worked Example: The Warehouse Operative

Take the example role in the workbook, a warehouse operative with a forklift licence. The must criteria are short: a valid forklift licence, availability for shift work and the right to work in the country. Everything else is a nice criterion with a weight, for example experience with a warehouse management system (weight 3) or a second language (weight 1). The exclusion criterion is a factual one, such as a start date that cannot work. The matrix then shows for each candidate whether a must criterion is missing, whether one is unclear and what the weighted score is. A candidate with an unclear forklift licence is not rejected. The suggestion reads "follow-up", and the next step is one short question.

Why Weights Must Be Set Before You Read CVs

Once you have read three strong CVs, your idea of "good" shifts. The weights drift toward the candidates you have already seen. Fixing weights and thresholds first removes that drift and makes the result comparable. If you later change a weight, change it for everyone and note the reason. That also makes your decision easier to explain to a hiring manager or, if needed, to the person who asks.

Typical Traps the Fairness Sheet Names

• Proxies. Photo, name, age, gaps in the CV or the address can stand in for protected characteristics. Do not rate them.
• Must criteria that are really nice. "Five years of experience" is often a habit, not a need.
• Exclusions that are not factual. An exclusion must be a reason that makes the hire impossible, not a preference.
• Scores treated as decisions. The suggestion sorts. A person decides on every case, in line with Art. 22 GDPR.

Sources

• Regulation (EU) 2016/679 (GDPR), Art. 5, 13 and 22
• General Equal Treatment Act (AGG), Sections 1 and 7
• Regulation (EU) 2024/1689 (AI Act), Annex III

FAQs

How many must criteria should a role have?

Usually two to four. Each extra must criterion removes more good candidates. Test each one: could it be learned after hiring?

Does the matrix decide who gets invited?

No. It suggests an order. A person decides on every candidate.

Can I use it for any role?

Yes. The four examples show different role types. Overwrite the criteria with your own.

Is this legal advice?

No. Criteria and exclusions must be genuinely necessary for the job. Check them with your HR or legal contact.

Can I use the matrix for an AI screening tool?

Yes, as the brief for the tool. Write the criteria in the matrix first, then give exactly those criteria to the tool and check its results against your own ratings on a sample. Keep a person in the final decision.

How do I handle candidates with the same score?

Look at the unclear ratings first and ask the missing question. If scores are still equal, use a work sample or a short call rather than a tie-break by gut feeling.

How long should I keep the filled matrix?

Treat it like the application file: keep it only as long as you need it for the process and for possible claims, then delete it. Agree the period with your data protection officer.

What if the CV does not show a must criterion?

Rate it as unclear and ask. Many facts, such as a licence or a notice period, are simply missing from CVs.

Can the matrix be used in a team?

Yes. Two people can rate the same candidates independently and compare. Large differences point to criteria that are written too vaguely.

Should I share the criteria with candidates?

Where you can, yes. A job ad that names the must criteria clearly saves unsuitable applications and makes the process more transparent.

Do I need to translate the matrix for international teams?

The workbook comes in English and German. Write your own criteria in the language of the job ad, so the matrix and the ad say the same thing.