Interview load is the number of interviewer-hours a team spends per successful hire, and it is quietly becoming the most expensive line item in recruitment. Data from Ashby's analysis of over 100 million applications shows applications per hire have tripled since 2021 to more than 300 per role, while interviews per hire have climbed roughly 40% — from about 14 to 20. When a funnel asks human interviews to absorb that volume, hiring slows down, interviewers burn out, and the best candidates walk. The fix is not more interviews run faster; it is moving screening signal earlier in the funnel so interviews can do the one job they are actually good at.
The math nobody runs
Most teams track time-to-hire and cost-per-hire. Almost nobody tracks interviewer-hours-per-hire, which is strange, because it is usually the largest hidden cost in the funnel.
Run the numbers for a typical mid-level role. Twenty interviews per hire, at 45–60 minutes each, plus preparation, debriefs, and scorecard write-ups, lands somewhere between 25 and 35 hours of interviewer time per successful hire. Those hours come from your most senior — and most expensive — people: engineering leads, sales managers, team heads. At executive level it gets worse: searches now routinely involve 5–8 rounds with 8–15 stakeholders, totalling 20–40 interviewer hours per hire, and around 62% of executive hiring managers report significant fatigue from interview coordination and participation.
Multiply by open requisitions and the picture sharpens. A company making 100 hires a year at 30 interviewer-hours each is spending roughly 3,000 hours — about a person-and-a-half of senior capacity, annually — on interviewing alone. That figure appears in no budget, so it is never scrutinised.
The volume side is deteriorating too. With 300+ applications per hire, candidates are now roughly 50% less likely to receive an interview than five years ago. That sounds like selectivity, but look closer: if screening upstream of the interview is resume-based, you are being more selective on weaker signal — filtering harder on a document that AI tools now write, polish, and keyword-optimise at scale.
Interviews are doing a job they're bad at
The instinct behind interview sprawl is understandable: more eyes, more rounds, more confidence. The evidence says otherwise.
Interviews are excellent at a narrow set of things — probing depth behind a real work sample, testing judgment interactively, assessing mutual fit, and selling the role. They are poor at the thing volume funnels ask of them: reliably ranking large numbers of candidates on baseline capability. Unstructured conversations are noisy and idiosyncratic; even well-run structured interviews are expensive per unit of signal, because every data point costs an hour of a senior person's calendar.
Meanwhile, the marginal round adds less than teams assume. Once two or three structured interviews are on the record, additional rounds mostly add correlated noise — the same halo effects, the same shared impressions from debriefs — rather than independent information. What they reliably add is delay, and delay is lethal: 42% of candidates withdraw when scheduling drags, and at senior levels roughly 40% of top candidates drop out of processes that exceed four weeks.
So the 20-interview funnel fails in both directions at once. It over-spends senior hours restating what earlier rounds already established, and it under-invests in the top of the funnel, where 300 applicants are being sorted by the least reliable instrument available — the resume.
Rebalance: assess early, interview late
The structural fix is a division of labour. Standardised assessment handles volume and baseline capability at the top of the funnel; interviews handle depth, judgment, and fit at the bottom — fewer of them, better designed.
Move capability screening out of the interview
Anything with a right answer, a demonstrable skill, or a scorable scenario should be measured before a human calendar gets involved. Cognitive and role-knowledge tests, situational judgment scenarios, written communication samples, and job simulations can all be administered to hundreds of candidates in parallel, scored consistently, and completed by candidates on their own schedule — no coordination delay at all. This is precisely where AI-graded scenario assessments earn their keep: an open-ended customer-escalation response or a case write-up, graded against a rubric in minutes, yields structured evidence that used to require a screening call. Platforms like AssessAll pair this with AI proctoring that reports integrity bands, so the volume stage keeps its trust without adding human review load.
The economics are hard to argue with. A screening interview costs an hour of interviewer time — realistically ₹2,000–5,000 in loaded senior-salary cost, plus scheduling latency measured in days. A proctored, auto-graded assessment costs a fixed, small amount per candidate (on AssessAll's pay-as-you-go model, credits run ₹30 / US$0.50) and returns results the same day. Even if the assessment only replaces the first human round, a 100-hire year at 300 applicants per role reclaims thousands of senior hours.
Cut rounds, keep structure
With capability established upstream, the interview stage can shrink to two or three rounds that each do distinct work: one structured behavioural interview against defined competencies, one work-focused deep-dive anchored to the candidate's actual assessment output, and — where the role warrants it — a final values-and-fit conversation. Anchoring the deep-dive to the assessment artifact is the underrated move: the interviewer stops improvising questions and starts interrogating real work, which is where interviews genuinely outperform every other method.
Each round needs a scorecard filled in before debrief, and debriefs need to happen within 48 hours. That is what keeps the shortened process rigorous rather than merely short.
Instrument the funnel
Finally, measure the thing you are trying to reduce. Track interviewer-hours-per-hire alongside time-to-hire and quality-of-hire. Set a budget for it per role level. When a hiring manager wants to add a round, the question becomes explicit: what independent signal will this hour buy that the assessment and existing rounds have not already bought? Usually the honest answer is none.
What changes in practice
Teams that make this shift report the same pattern: top-of-funnel throughput rises because assessments run in parallel rather than serially through calendars; interview counts fall because interviews start from evidence instead of generating it; and offer velocity improves because the process stops leaking good candidates into competitors' faster funnels. The interviews that remain get better, too — interviewers arrive with a capability profile and a work sample in hand, and spend the hour on judgment and depth instead of re-verifying basics.
None of this diminishes the interview. It elevates it. An hour of a senior engineer's attention is one of the most valuable evaluation instruments an organisation owns; spending twenty of those hours per hire to compensate for a weak top-of-funnel is the real disrespect to the craft.
Takeaway: Interviewer-hours-per-hire is the metric your funnel is silently bleeding — measure it, then move baseline screening into standardised, proctored assessment so your two or three remaining interviews spend senior hours on judgment, not triage.