How to Screen Resumes Effectively: 6 Steps, Sample Scorecard

To screen resumes effectively, decide what the role needs before you open the first CV, then score every CV against the same short list of criteria. Write 1 to 3 knockout criteria and 4 to 6 weighted criteria. For each score, copy the line from the CV that proves it. When the CV does not mention a criterion, mark it "Not stated" and ask about it in the phone screen. Have a second person score the same 10 CVs before you screen the whole pile. At the end, sort the CVs into Yes, Maybe and No, and check that each group of applicants passes at a similar rate.

Greenhouse reports an average of 244 applications per job in 2025, up from 116 in 2022, from data on more than 6,000 companies and 640 million applications (Greenhouse, The Hire Standard). At that volume, a first pass of 45 seconds per CV takes about 3 hours. When a role gets hundreds of CVs, Firstview's CV screening can read each one against your criteria and give a reason for every result, and a person still decides who moves on.

How to screen resumes effectively: 6 steps

The 6 steps below take you from the job post to a shortlist that someone else can check.

1. Write the criteria before you read a CV. Choose 1 to 3 knockout criteria that the job cannot work without, and 4 to 6 weighted criteria, marked as required work or nice-to-have. Write each one as something a CV can show, for example "led a team of 5 or more" instead of "strong leader". Agree on the list with the hiring manager first; the intake meeting questions help turn their words into criteria.
2. Build one scorecard with an evidence column. Give every criterion a weight, the 4 results below and a column for the CV line that supports the result.
3. Calibrate on 10 CVs. Two people score the same 10 CVs alone, then compare. Where they disagree on a criterion for more than 2 of the 10 CVs, rewrite that criterion so it names the evidence, and score the 10 again.
4. First pass: check only the knockouts. Spend about 45 seconds on each CV. In this pass, a CV goes to No only when it clearly fails a knockout. The second pass can still move a CV to No on points (see the bands below). A knockout that the CV does not mention is "Not stated", not "Not met".
5. Second pass: score the rest with evidence. Score each remaining CV against the weighted criteria and copy the line that supports each result.
6. Sort into Yes, Maybe and No, then check the pass rates. Move the Yes CVs to a phone screen, then Maybe CVs in order of points plus open points while the team has calls left. Ask the "Not stated" items as the first questions. Then compare the pass rates of applicant groups (see the bias section below).

Use 4 results for each criterion, and score them the same way every time:

Result	When to use it	Points
Met	A CV line shows it	Full weight
Partly met	A CV line shows part of it	Half the weight, rounded down
Not met	The CV shows the opposite or a clear lack of it	0
Not stated	The CV says nothing about it	0 now; the weight counts as open points

Keep "Not stated" apart from "Not met". Each CV gets two numbers: its points, and its open points, which are the weights of the "Not stated" criteria. A weak CV has few points and few open points. A short CV has few points and many open points, and the open points tell you what to ask before you decide.

Sample resume screening scorecard

This sample shows a filled scorecard for one CV. Firstview wrote it for this guide: the role, the candidate and every CV line are fictional, and the weights are an example, not a benchmark.

Role: Customer support team lead (sample). Candidate: Candidate A (sample).

#	Criterion	Type	Weight	Result	Evidence from the CV	Points
K1	1 year or more leading or coaching support agents	Knockout	-	Met	"Team lead, 8 agents, March 2023 to now"	-
K2	Can work weekend shifts	Knockout	-	Not stated	The CV does not mention shifts	-
1	Handled escalations or complaints	Required	30	Met	"Owned tier-2 escalations; reopened tickets fell from 14% to 9%"	30
2	Used a ticketing tool as an admin or power user	Required	20	Met	"Set up queues and macros for 3 teams"	20
3	Tracked team metrics such as CSAT or handle time	Required	25	Partly met	"Ran weekly QA reviews" (no metric named)	12
4	Wrote help articles or call scripts	Nice to have	15	Not met	No writing work in any role	0
5	Used a second language with customers	Nice to have	10	Not stated	The CV lists no languages	0
	Total		100			62, plus 10 open

Bands for this role:

• No: any knockout is "Not met", or the CV has fewer than 40 points and its points plus open points stay below 60.
• Yes: 60 points or more, and no knockout or required criterion is "Not met".
• Maybe: every other CV, including a CV with 60 points or more that misses a required criterion.

A knockout that is "Not stated" does not change the band; it becomes the first phone-screen question. The Yes line is 80% of the 75 points that the 3 required criteria carry, so a CV that meets most of the required work passes without any nice-to-have. Use the same rule to set the line for other roles.

Decision: Yes, with 62 points plus 10 open. The phone screen starts with the 2 "Not stated" items: "Can you work weekend shifts?" and "Which languages do you use with customers?"

Read the evidence column before the total. A recruiter who opens this scorecard a week later can see why criterion 3 got half the points and check it against the CV in less than a minute.

What to look for when you screen a resume

When you screen a resume, look for evidence of the work the role needs: what the person did, at what scale and with what result. A list of duties copied from a job post repeats the job's words and says little about what the person did. These signals are worth a weighted criterion when the role needs them:

• Scope and results: team size, volume, budget or a before-and-after number, such as "reopened tickets fell from 14% to 9%".
• Skills used in a role: a skill counts more when a job entry shows where it was used, and less when it appears only in a skills list.
• Progress over time: a promotion, more people to manage or harder work in a later role.
• A CV written for this job: the most relevant experience comes first, and the summary names this kind of role.
• Instructions followed: the application has what you asked for, such as a portfolio link or a salary range.

Do not score what the job does not need. A design-heavy CV is a plus for a designer and a neutral fact for a support lead.

Which resume red flags should stop a candidate?

A resume red flag should stop a candidate only when it breaks a knockout criterion, such as a missing licence that the job requires by law. Treat most items on common red-flag lists as questions for the phone screen:

Signal	Treat it as	What to ask in the phone screen
A gap between jobs	A question	"What did you do between these two roles?"
Several jobs of less than 1 year	A question	"Why did each of these roles end?"
A career change from another industry	A question, often a plus	"Which parts of your last job are closest to this one?"
Dates or titles that do not match	A question, then a check	"Can you walk me through these dates?"
Spelling mistakes	A score only if writing is a job skill	Nothing; score it under the writing criterion
Claims that are too large for the role	A question with a number	"What was your part, and what did the team do?"

Write the reason next to each question in the scorecard. When the answer is good, the CV moves up; when it is not, you have a reason that you can show.

How long should resume screening take?

For an average role, the first pass of resume screening alone takes about 3 hours. That figure is our arithmetic: 244 applications, the 2025 average per job in Greenhouse's data, at 45 seconds each is 10,980 seconds, or just over 3 hours, before the second pass starts.

Three habits keep the time down without losing people:

• Screen in fixed blocks, for example 2 blocks of 90 minutes a day, so the criteria stay fresh and the last CV gets the same attention as the first.
• Keep the first pass to knockouts. The full read is only for the CVs that pass it.
• Call in score order. When the Yes band holds more people than you can call this week, call the highest scores first and tell the others when you will contact them. Every CV still gets the same first pass, and every CV that passes it gets the same second pass.

How to reduce bias when you screen resumes

You reduce bias in resume screening when every CV gets the same criteria, every score has evidence and you check the results by group. Researchers have measured bias at this stage. In an experiment with more than 83,000 fictitious applications to jobs at 108 of the largest US employers, "Distinctively Black names reduce the probability of employer contact by 2.1 percentage points relative to distinctively white names" (Kline, Rose and Walters, NBER working paper 29053, 2021, revised 2022).

Three checks help:

1. Score only job criteria. Under the laws that the EEOC enforces, "it is illegal to discriminate against someone (applicant or employee) because of that person's race, color, religion, sex (including transgender status, sexual orientation, and pregnancy), national origin, age (40 or older), disability or genetic information" (EEOC). A criterion such as "recent graduate" or "digital native" points at age; write the skill instead.
2. Require evidence for every score. Write the CV line that supports each score next to it.
3. Compare pass rates. If applicants answer voluntary self-identification questions, compare the share of each group that reaches Yes. The US Uniform Guidelines say that a selection rate for any race, sex or ethnic group "which is less than four-fifths (4/5) (or eighty percent) of the rate for the group with the highest rate will generally be regarded by the Federal enforcement agencies as evidence of adverse impact" (29 CFR 1607.4). For example, if 40% of one group passes and 28% of another, 28 divided by 40 is 70%, which is below 80%. Then find the criterion that causes the gap and ask whether the job needs it.

When to use software for resume screening

Use software for resume screening when a role gets more CVs than one person can score with evidence in the time you have. The common tool types check different things:

Tool type	What it checks	Where it fails
ATS keyword filter	Whether the CV contains the words you chose	Passes CVs that copy the job post; drops CVs that use other words
ATS knockout questions	Answers on the application form	Only as good as the questions; candidates can answer "yes" freely
AI screening against criteria	What the person did in each role, per criterion	Needs clear criteria and a person who reviews the results

Firstview's CV screening uses written criteria and a reason for each result, as this guide does, but its results and points work in their own way. It reads each CV against your criteria, "counts only the experience that matches and gives a reason for every criterion". Each criterion gets one of 4 results, "Matched, Not matched, Needs review or Not applicable, with points and a one-line reason". When a CV is silent on a criterion, the result reads "The CV does not say. Left for you to decide.", the same job as "Not stated" in the sample above. If you change a criterion after the first interviews, every CV in the stage is evaluated again against the new version. Firstview ranks the candidates, and a person moves them to the next stage. It does not check references, verify identity or confirm what a CV claims, so the phone screen still does that work.

How we made this guide

We read the Greenhouse benchmark page, the NBER abstract of Kline, Rose and Walters, the EEOC page on prohibited practices and the four-fifths rule in 29 CFR 1607.4, and compared the steps with 5 of the top US search results for this question. The sample scorecard is fictional: Firstview wrote the role, the CV lines and the weights for this guide, and they come from no real candidate or customer. The 3-hour figure is our own arithmetic from the Greenhouse average. No named person has reviewed this guide yet.