Candidate screening: how to filter applicants efficiently

Candidate screening is the set of filters between "everyone who applied" and "the few people worth interviewing." Its purpose is efficiency with fairness: spending interview time — the most expensive resource in hiring — only on candidates who could plausibly get the job, without filtering out good people for bad reasons. A screening process is working when the people who reach interviews are consistently interview-worthy; it is broken when interviews keep discovering basics a filter should have caught.

Layer one: resume screening that avoids the classic traps

• Screen against the written must-have list — not against gut feel. Every resume gets the same checklist; a candidate either shows the must-haves or does not.
• Time-box it — a few minutes per resume with a checklist beats long impressionistic readings, and produces more consistent results across screeners.
• Distrust proxies — prestigious employer names and universities predict less than the checklist does; they mostly predict who had access to prestige.
• Note claims to verify later — specific achievements, titles, and dates that will matter if this candidate reaches the final round. Screening is where verification targets get identified cheaply.

Layer two: the structured screening call

A 15–20 minute call, same questions for everyone, covering the eliminators that resumes hide: actual salary expectations against your range, notice period and start date, location and work-mode fit, work authorization where relevant, and a two-minute account of the most relevant experience. Half of late-stage hiring failures are mismatches that were knowable in this call. Score it immediately — a one-line verdict per criterion — so decisions a week later rest on notes, not memory.

Layer three: work samples where the role allows

For many roles, a small representative task — analyzing a dataset, drafting a reply to a difficult customer, debugging a broken function — predicts job performance better than any conversation about the work. Keep it short (an hour, not a weekend), relevant (the actual job, not a puzzle), and respectful (paid where substantial, and never free labor disguised as assessment). A candidate’s work sample plus a structured interview is the strongest signal pair screening can produce.

Layer four: verification — the filter fraud counts on you skipping

The final screening layer is checking that the winning story is true: employment history, credentials, and the specific claims you noted in layer one. Resume fraud persists because this layer is expensive when done by phone — days per candidate — so it gets skipped, and fraudulent candidates advance on well-rehearsed interviews. Verified employment records invert the cost: on EmployDB, a recruiter requests access and, with the candidate’s consent, reads history issued by verified employers in minutes — positions, dates, promotions, certifications, with evidence attached. When verification is cheap, it stops being skippable, and screening finally covers the one thing interviews cannot test: whether the past actually happened.