How to screen resumes effectively, one decision at a time

When we pulled six months of our own screening data in August 2026, one number stood out. Nearly half the applications we scored for relevance (49.5 percent of 33,753) covered under 30 percent of their role’s requirements, and just 1.7 percent covered over 70 percent. Those figures are from the India hiring data we published, and they changed how I think about the question of how to screen resumes effectively.

Most advice on resume screening is about reading faster: skim the top third, look for keywords, check for gaps. Recruiters already read fast. Ladders’ 2018 eye-tracking study found an average of 7.4 seconds on the initial screen of a resume. That advice assumes the pile is mostly worth reading. At volume it is not. Half of a typical pile can be set aside by checking whether the experience is relevant to the role at all, and the real skill is in the narrow band at the top and the messy band in the middle. This is the method we would hand a recruiter facing 800 applications on a Monday, step by step.

How do you screen resumes effectively?

Screen resumes effectively by deciding before you open the first one what ends a review and what earns a closer look. Write three to five must-haves from the job, check each resume for those first, then judge relevant experience rather than matching keywords. Record one piece of evidence for every score, and send unclear cases to a short structured interview instead of guessing.

The order matters more than the speed. A recruiter who reads every resume top to bottom with the whole job description in mind can make a different call at 4 PM than at 10 AM on the same kind of profile, simply because the bar was never written down.

Step 1: Write the must-haves before you open a resume

A job description is usually a mix of three things: real requirements, nice-to-haves, and text copied from the last posting. Screening against all of it means every resume fails something, so the recruiter falls back on instinct.

Before the first resume, write down three to five must-haves in plain sentences, the kind you could check in ten seconds:

• “Has a graduate degree” (if that is genuinely required, not habit)
• “Has at least one year in a customer-facing role”
• “Lives in or can relocate to Pune”
• “Has used Excel for reporting, not just data entry”

Everything else is a nice-to-have and never rejects anyone on its own. If the list grows past five, some of it is a wish. We wrote more about this upstream problem in how a job description sets up the screen.

Step 2: Check relevance, not keyword matches

The most common resume screening mistake is treating a keyword as evidence. A resume that says “customer success” five times may describe two years in a call centre or six months of an unrelated internship with the phrase in the title. A resume that never uses the phrase may describe exactly the work.

Ask a different question of each resume: how much of this person’s experience is relevant to the work this role does? Years in an unrelated field count for little. A strong degree in the wrong discipline does not outweigh a year of directly relevant work. Over-qualification is a signal too: a candidate with twelve years of experience applying for an entry-level role usually leaves soon, or was never really applying.

This is also where blunt filters fail quietly. In Harvard Business School’s Hidden Workers study, 88 percent of employers agreed that qualified high-skills candidates are vetted out because they do not match the exact criteria in the job description, and 94 percent said the same of middle-skills workers (quoted in The Register). A rule like “reject any gap over six months” removes people who were caring for family, studying or recovering, and none of them were less able to do the job.

Step 3: Score every resume the same way, with evidence

Pick a small scale and use it for every resume in the batch. For a first pass, 0 to 3 works: 0 fails a must-have, 1 passes but has little relevant experience, 2 has clearly relevant experience, 3 has relevant experience plus something the role values (a domain, a tool, a result with numbers).

Next to every score, write the line from the resume that justifies it. “2: 18 months as a collections agent at an NBFC” takes five seconds and does three things. It makes you actually find the evidence. It lets a colleague check your call without re-reading the resume. And when a hiring manager asks why a candidate was rejected, you have an answer that is not “they didn’t seem right.”

A worked example: three resumes, one role

Take a customer success associate role in Pune that needs one to three years in a customer-facing job and regular use of Excel for reporting. Three illustrative resumes, scored with the method above:

A keyword filter ranks A first and B last. Reading for relevance reverses that. C is not a rejection, but it is not a fast-track either: the over-qualification is a question for the interview, not a reason to skip it.

Step 4: Decide the middle band before it becomes a pile

In any real batch, a large share of resumes are neither a clear yes nor a clear no. In our data, 48.8 percent sat between the two marks above: more than 30 percent relevant, but not more than 70. Without a rule, they sit in a folder and nobody returns to them, or they get read last, at the end of a long day, by the most tired person on the team.

Set the rule in advance. For example: score 3 goes straight to the interview list, score 0 gets a polite rejection, and every score 1 or 2 gets one more check, either a second reader or a short structured interview. The second option is usually better, because a resume is a document the candidate controls. A ten-minute conversation tells you more about a borderline profile than a second skim of the same page.

Step 5: Watch for resumes written by AI, without rejecting them for it

Applicants now use AI to write and polish resumes. Treat that as a reason to look for evidence, not as a reason to reject. A polished resume with no numbers, no dates and no named projects deserves a probing question in the interview; a polished resume with specific results is still a strong resume. We go deeper on the signs in how to tell if a resume is AI-generated.

Step 6: Hand the shortlist on with its reasons

The output of a resume screen is not a list of names. It is a list of names with a score, the evidence line and anything the interviewer should probe. That turns the next round from a repeat of the resume (“tell me about yourself”) into a check of the specific claims that earned the candidate a place.

Where software helps, and where it should not decide

Everything above can be done by hand, and for 40 applications it should be. At 800, the method is the same but the arithmetic is not. At a minute per resume, 800 resumes is more than thirteen hours of reading before the first interview is booked.

HireQwik does this first pass against each job’s own criteria. Every resume gets a score out of 100 based on how relevant the experience is to that role, not how many keywords it repeats; experience in unrelated work counts for little, and over- and under-qualified profiles are capped. Deal-breakers are written as plain sentences, the same way as in Step 1. Every score comes with a “why this score?” panel quoting the resume lines behind it. A recruiter can choose, per job, to reject below one score and invite above another, so people only review the middle band, and rejected candidates can be pulled back into review. Resumes the system cannot read are flagged for a person, never rejected automatically, and a resume that looks AI-written gets a flag, not a lower score.

For the middle band, the next step is a structured voice interview, 10 to 20 minutes long, that the candidate takes at a time they choose, scored against the same job. That is where Step 4’s “short structured interview” stops being a recruiter’s afternoon. HR still makes every decision on who moves forward.

If you are comparing tools for this step, our guide to the best resume screening software for high-volume hiring judges ten of them on the same five questions. Once the shortlist exists, the next problem is getting candidates booked, which is what our guide to interview scheduling software covers.