The recruitment funnel, stage by stage, with real numbers

At the end of August 2026 I pulled six months of our screening data into one place: 66,873 resumes screened, 8,096 voice interviews created and 2,589 completed, almost all of it high-volume hiring in India. The textbook recruitment funnel is a neat triangle narrowing evenly from applicants to hires. The one in our data looked nothing like that. Half of it fell away at the very first step, before any human had looked, and one of the biggest leaks further down had nothing to do with candidate quality at all.

This guide walks through the funnel stage by stage, with the numbers we measured (published in full in what 66,000 resumes taught us), outside benchmarks to compare against, the four places high-volume funnels leak and how to measure your own this week.

What is a recruitment funnel?

A recruitment funnel is the path candidates take from first seeing a job to joining the company, broken into stages: attraction, application, screening, interviews, offer and joining. At each stage some people drop out or are turned down. Measuring how many move from one stage to the next, and how long each step takes, shows where hiring actually stalls.

The point of the funnel is not the diagram. It is the two numbers at every stage: the conversion rate (what share moved on) and the wait (how many days they sat there). A team that knows those numbers for its main roles can usually find its biggest problem in an afternoon.

The stages of a recruitment funnel

In most teams the middle row, screening, is where the most people and the most recruiter hours sit, so it is where we will spend most of this guide.

What a high-volume funnel looks like in practice

Here is what the screening stage looked like across the pipelines in our data, March to August 2026. Every figure is aggregated; no company or candidate is identifiable.

Resume relevance. Of 33,753 resumes that went through relevance scoring, 49.5 percent scored below 30 percent relevance to the role they applied for. Only 1.7 percent scored above 70 percent. Half the funnel, in other words, was never a fit for that job. It was facility managers applying for customer success and mechanical engineers applying for frontend roles.

Getting to the interview. Of 8,096 voice interviews created, 2,589 were completed. The biggest single leak we can name inside that gap is the no-show: 21 percent of booked interviews expired without the candidate ever joining, even with slots the candidates chose themselves, calendar invites and reminders.

Verdicts. Of 1,414 fully screened candidates, 54.2 percent were No Go, 23.0 percent On Hold, 15.0 percent Go and 2.7 percent Strong Go. Fewer than three in a hundred cleared the bar that means “move fast on this person.”

Three lessons came out of that. The top of a high-volume funnel is mostly noise, so the first filter has to be cheap and fast. The thin end is very thin, so the people who reach it deserve speed. And some of the leaks that hurt most, like the no-show, are about logistics, not talent.

Benchmarks from other recruitment funnels

Your numbers will differ from ours, and that is the point of measuring them. Two public benchmarks are useful for orientation.

CareerPlug’s 2025 Recruiting Metrics Report, based on more than 10 million applications to over 60,000 small businesses in 2024, found that 3 percent of applicants reached an interview, 27 percent of interviews led to a hire, and employers received about 180 applicants for every hire. That is a small-business, mostly US picture, with hourly and local roles.

Gem’s 2025 recruiting benchmarks, drawn from more than 140 million applicants between 2021 and 2024, found hiring teams running 20 interviews per hire, up from 14 in 2021, with 84 percent of offers accepted. That is a picture of larger, mostly tech-heavy employers, where the interview stage has grown heavier, not lighter.

Use benchmarks to spot a stage that is badly out of line, not as targets. A campus drive where 3 percent of applicants reach an interview might be perfectly healthy, or might be rejecting good students at the resume stage. Only your own trend over time can tell you which.

Where high-volume funnels leak, and what fixes each

1. The relevance leak (application to screening). When half the applicants are not a fit, recruiters spend half their reading time on resumes that never had a chance, and the good ones wait in the same pile. The fix is to score every resume against the specific job, with the reasons written down, and let a recruiter review only the uncertain middle. We wrote about the right resume-to-interview ratio for this stage.

2. The no-show leak (invite to completed screen). One in five booked interviews never starting is a logistics problem: the slot did not fit the candidate’s life, or the reminder never reached them. Letting candidates pick their own time, including evenings, and sending calendar invites and reminders reduces it. It does not remove it, so plan capacity for it. More in our guide to reducing interview no-shows.

3. The waiting leak (screened to hiring-team interview). A candidate rated Go on Monday who hears nothing until the following Wednesday is often already in another company’s process. Measure the days between stages, not just the percentages. The two clocks recruiters confuse are explained in time to hire vs time to fill.

4. The wrong-rejection leak (hidden). This is the one no funnel report shows by default: good candidates rejected early, who would have been hired. You only see it if you record outcomes and look at who was hired despite a low screen, and who was rejected early but turned up elsewhere. We covered how to check this in screening verdicts vs hiring outcomes.

How to measure your own recruitment funnel this week

You do not need a new tool for a first version. Pick one high-volume role and one recent month.

1. Write down what each stage means. “Screened” means a completed first conversation, not “resume opened.” Without definitions the numbers shift every month.
2. Count people at each stage for that role and month: applied, shortlisted on resume, invited, completed the screen, interviewed by the team, offered, accepted, joined.
3. Divide each stage by the one before. That is your conversion rate.
4. Record the median days between stages. One slow handover can cost more good candidates than any filter.
5. Find the one stage that is most out of line, by conversion or by wait, and fix only that this month.

Repeat next month with the same definitions. The trend is what you manage; a single month is a snapshot. If you want to put a cost on each stage too, our cost per hire guide shows how.

Seeing the funnel per job in HireQwik

In HireQwik, the screening stage of this funnel is measured per job, because that is where the volume sits. Each resume is scored against its own job description, with reasons, and a job can be set to auto-reject below one score and move candidates forward above another, so a recruiter reviews only the “Needs review” band in the middle. Candidates book their own interview slot, with a calendar invite, and take a 10 to 20 minute voice interview that ends in a verdict with the reasons and the recording.

The part that closes the loop is outcomes. A recruiter can mark each screened candidate as hired, offered, not selected or withdrew, and the job’s funnel then shows a “Screener vs reality” card: verdicts against what actually happened, with the costly miss, a candidate the screen rejected whom the team hired anyway, shown as its own cell rather than averaged away. Until a job has ten outcomes, the card shows counts, not percentages, because rates on tiny samples mislead. It measures; it never changes thresholds on its own. Moving a threshold stays a human decision.

The funnel you draw on a whiteboard is a hypothesis. The one you measure, stage by stage with the same definitions every month, is the one you can fix. If you hire at volume and want to see the screening stage working, book a short demo or read how we approach high-volume hiring.