The Recruitment Funnel and Candidate Drop-Off Rate

A recruitment funnel is the sequence of stages candidates pass through - reached, applied, screened, interviewed, offered, hired - with fewer people at each step. Candidate drop-off rate is the share who leave a stage without completing it, calculated as candidates who exit a stage divided by candidates who entered it. Measuring both tells you which stage is losing the people you wanted. The Cognitive targets the 2 stages that usually leak worst: waiting to be screened, and waiting for an interview slot.

The stages, and where candidates drop off

Stage	What it measures	Usual cause of drop-off	What fixes it
Reached (outbound)	People contacted about the role	Generic messages, or contact details that never resolved	Fewer, better-chosen approaches and verified contact data
Applied	Candidates who completed an application	Long forms and account creation before any human contact	Cut the form; ask for what you will actually read
Screened	Candidates who got a real first evaluation	The wait - a queue that nobody has hours to clear	Screen with an interview everyone can take, not a call slot
Interviewed	Candidates who completed a substantive interview	Scheduling friction and timezone or shift conflicts	Candidate self-scheduling inside a wide window, including evenings
Offered	Candidates who received an offer	Slow decisions after the last round	A same-day debrief against the rubric
Hired	Offers accepted and started	Compensation, counter-offers, or a process that felt impersonal	Speed, and a human owning the close

The stage before the funnel: sourcing decides how wide the top is

Sourcing channel	Who it reaches	Typical cost	Why the funnel narrows here
Job advert and job boards	Active jobseekers who find the advert	Advertising spend per role	Reaches only people already looking, which on a scarce skill is a small share of the market
Employee referrals	Networks adjacent to your current team	Referral bonus per hire	High conversion, low volume, and it reproduces the team shape you already have
Past applicants in your ATS	People who applied to earlier roles	Almost nothing — usually a team's cheapest pipeline	Nobody re-engages it, and contact details decay faster than the profiles do
Manual boolean search	Passive candidates, one query at a time	Recruiter hours per search	Capped by hours; the search is never recorded and stops when the tab closes
AI candidate sourcing	Passive candidates matched against a whole role brief, with contacts enriched	Credits per candidate returned and per successful reveal	Widens the top, but a sourced candidate is still unassessed until they interview

Every drop-off rate below the first stage is a percentage of whatever entered it, which makes the width of the top the quiet determinant of the whole funnel. A team measuring a 60% screening drop-off on 100 applicants and a team measuring the same rate on 1,000 are not in the same situation, and only one of them has a screening problem.

Candidate sourcing is what widens the top deliberately rather than hoping the job advert does it: searching profiles for people who have not applied, checking them against the role brief, revealing contact details and making the first approach. It is the only funnel stage where the number entering is a decision rather than a result — and on a scarce skill, where the advert reaches almost nobody who is employed and content, it is the only stage worth working on first.

Two failure modes are specific to this stage. Approaching too many people with a generic message produces a reply rate low enough that the whole channel looks broken; and contact details that do not resolve produce silent loss that never appears as drop-off at all, because the candidate was never actually reached. Both are measurable: track candidates surfaced per search, reply rate per sequence, and cost per contact successfully revealed, and treat 'reached' as an outcome you verified rather than an action you took.

How to calculate candidate drop-off rate

Drop-off rate for a stage is candidates who left that stage without completing it, divided by candidates who entered it, times 100. Compute it per stage rather than for the funnel as a whole: a single top-to-bottom number tells you that you are losing people and nothing about where. Track it per source too - a channel with high volume and high drop-off is usually cheaper to fix than to replace.

Why the screening stage leaks hardest

The screening stage is where a candidate's interest is highest and your response time is slowest, which is the exact combination that produces drop-off. It is a capacity problem rather than an interest problem: a team that can run 40 screens a week cannot clear 400 applicants, so the queue ages until the good ones accept elsewhere. Interviewing every applicant with a live AI interview removes the queue - The Cognitive's interviews run in parallel at any hour, so the stage stops being a waiting room.

Frequently Asked Questions

How does sourcing change the recruitment funnel?

Sourcing widens the top of the funnel on purpose rather than waiting for a job advert to do it. Because every rate below the first stage is a percentage of whatever entered it, a narrow top caps the whole funnel no matter how good the later stages are — a team with a 60% screening drop-off on 100 applicants and one with the same rate on 1,000 do not have the same problem. Sourced candidates also enter warmer, because they were approached about this role rather than answering a general advert.

What is a good candidate drop-off rate?

There is no universal figure worth quoting, because it moves with role type, market and application length. Use your own baseline: measure per stage for a quarter, then treat any stage above your funnel's average as the one to fix. The diagnostic that matters more than the absolute number is whether drop-off clusters around waiting - that is a capacity problem, and it is solvable.

What are the stages of a recruitment funnel?

Most funnels run: reached or attracted, applied, screened, interviewed, final round, offer, hire - with a start date at the end. Some teams add a sourcing stage before 'reached' and a 90-day retention checkpoint after the start date, which is what connects the funnel to quality of hire rather than only to speed.

How do you reduce candidate drop-off?

Shorten the wait and shorten the form. Drop-off concentrates wherever candidates are asked to spend effort before receiving anything back, or to wait while a queue clears. Practical levers: cut the application to what you will read, let candidates self-schedule instead of negotiating slots, run first-round evaluation in parallel rather than serially, and decide within days of the last round.