Most hiring dashboards report volume. Applications received, profiles sourced, interviews conducted, offers rolled out. Every number goes up when the team works harder, which is why leadership likes them and why they explain nothing.
A funnel is not a set of counts. It is a set of ratios and delays. The useful questions are where candidates stop moving and how long they sit before they do, and neither of those is answered by a bigger number at the top.
The funnel, stage by stage
Before any measurement, agree the stages and where each one starts and ends. Half the arguments about hiring data are actually arguments about definitions.
- Sourced or applied: candidates entering the pipeline, by channel
- Recruiter screen: passed a first review against the actual requirement
- Hiring manager review: profile accepted as worth interviewing
- Interview: first round completed
- Final round: last interview or panel completed
- Offer extended
- Offer accepted
- Joined
- Passed the first six months
Note that the funnel does not end at accepted, and this is where a lot of reporting stops. An offer accepted by someone who never turns up is not a hire, and in Indian hiring the distance between those two events is measured in months.
A funnel measured only to offer acceptance is a funnel that congratulates itself for something that has not happened yet.
Stage conversion, and what a drop actually means
Conversion is the pass-through rate between two adjacent stages. Compute it for every pair, for each role family, and the shape of the funnel tells you where the problem is.
The diagnosis is usually specific to the stage, which is what makes this the most actionable metric on the page.
| Where conversion drops | What it usually means |
|---|---|
| Applied to recruiter screen | The job advertisement is attracting the wrong people, or the channel mix is wrong. High volume with low pass-through is a targeting problem, not a success |
| Recruiter screen to hiring manager accept | Recruiter and hiring manager do not share the same picture of the role. This is an intake problem and it is fixed in a conversation, not with more profiles |
| Hiring manager accept to interview held | A scheduling and availability problem. Candidates are lost to competitors while waiting, and this is the cheapest leak to fix |
| Interview to offer | Either the bar is unclear or the screening is too generous. Consistently low here means interviews are doing the screening that should have happened earlier |
| Offer to acceptance | Compensation, competing offers, or a slow and impersonal process. Decline reasons tell you which, if anyone records them |
| Acceptance to joining | The notice period gap. Counter-offers and second thoughts, and almost entirely a function of what happens during those weeks |
On benchmarks: be careful with the ones you read. Published figures range so widely by role, seniority, channel and market that comparing yourself to a global average is mostly noise. A junior volume role might convert from application to screen in the low single digits, a referred senior candidate might convert at one in three, and both are normal. Your own last two quarters, by role family, is the comparison that means something.
The diagnostic: where candidates are sitting
If you only add one metric to an existing report, add median days in each stage. Not average, because one abandoned candidate at 90 days distorts it, and not total elapsed time, because that hides the location of the delay.
What this almost always surfaces is that the delay is not in recruiting. Profiles sit unreviewed for six days, interview slots take nine days to appear, feedback after the final round takes four. Recruiters get held accountable for a time to fill number that is mostly composed of other people's response times.
Once the data exists, hiring manager response time becomes a metric you can put in front of a leadership team with a straight face. It is the single highest-leverage number in most funnels, and it is the one nobody wants to publish.
Time to fill against time to hire
These get used interchangeably and they answer different questions.
Time to fill runs from the requisition being approved to the offer being accepted. It includes the time spent waiting for approvals, deciding what the role actually is, and finding the first credible candidate. It is a measure of the business as much as of recruiting.
Time to hire runs from a specific candidate entering the pipeline to that candidate accepting. It measures the process a candidate experiences, which is what determines whether you lose them to a faster company.
Report both, and report the median rather than the mean. Then add a third figure if you can: time from requisition approved to first candidate submitted. When time to fill is bad and time to hire is fine, the problem is sourcing or role definition. When both are bad, it is the process.
Offer acceptance, and the gap India has to manage
Offer acceptance rate is offers accepted divided by offers extended. It is a reasonable measure of whether your compensation and your process are competitive, and it is easy to compute.
It is also not the end of the story here. With notice periods of 60 to 90 days common, an accepted offer sits exposed for two or three months while the candidate remains in an office where their employer now has every reason to counter-offer. So track offer-to-join separately: of the offers accepted in a quarter, how many people actually started.
That number is where the real loss usually is, and it responds to process rather than to money. Candidates who hear nothing between accepting and joining are the ones who do not arrive. Structured contact during the notice period, an early document and onboarding start, a conversation with their future manager, and a realistic view of the first month all move it. Recording the reason each time someone drops out is what turns it from an unpleasant surprise into something you can act on.
For roles where drop-out is predictable, the mature response is to plan for it in the requisition rather than to be shocked twice a year. That is uncomfortable to say out loud and it is how experienced teams actually hit their numbers.
Cost per hire, and how it gets calculated wrong
The standard definition is the total of internal and external recruiting costs for a period, divided by the number of hires in that period.
External costs are the ones people remember: job board subscriptions, agency fees, advertising, assessment tools, background checks, referral payouts, travel. Internal costs are the ones that get left out: recruiter salaries and the proportion of hiring manager and interviewer time spent on hiring. Leave the internal side out and cost per hire looks flattering while telling you nothing about where the money goes.
Two cautions. First, a single number across very different roles is meaningless, so segment by role family or seniority at minimum. Second, cost per hire is a cost metric and not a quality metric, and treating it as a target creates the obvious perverse incentive. The cheapest hire and the best hire are not usually the same person.
Source of hire: measure yield, not volume
Channel reporting almost always shows applications by source, which rewards whichever channel is best at producing volume. That is the wrong ranking.
Rank channels by hires produced and by what happens to those hires afterwards. A referral channel that sends 20 candidates and produces four who are still performing after a year beats a job board that sends 2,000 and produces two who leave in month seven. Once you can see conversion, cost and retention by source together, the budget conversation becomes straightforward instead of political.
Attribution needs a rule, decided once. When someone applies through a job board and is then referred, count it consistently rather than letting whoever reports it choose.
Quality of hire, the one everybody wants
This is the metric every leadership team asks for and very few organisations actually maintain, because it requires a definition and patience.
A workable version combines two or three things measured at a fixed point, usually six or twelve months: the performance rating or a manager assessment against the expectations set at hiring, whether the person is still there, and whether the hiring manager would make the same hire again. Some teams add time to productivity where the role has a measurable ramp.
The discipline that matters is holding the definition steady. A composite that gets redefined every year produces a trend line that means nothing, and you will not have enough data points for two or three years anyway. Pick a version you can live with, write it down, and leave it alone.
First-year attrition is the cheap proxy while you build the real thing. Anyone leaving inside twelve months is worth reviewing individually, because the reason is usually visible in the hiring process in hindsight: a role described differently to how it turned out, a bar bent to close a requisition, or a manager relationship that was already visible at interview.
A one-page funnel dashboard
What earns a place on a page that a hiring leader will actually read every week.
- Conversion rate at each stage, this quarter against last, by role family
- Median days in each stage, with the stages owned outside recruiting marked as such
- Open requisitions by age, with anything past 60 days named rather than aggregated
- Time to fill and time to hire, both medians
- Offer acceptance and offer-to-join, with decline and drop-out reasons
- Hires by source, with cost and first-year retention beside the count
- Interviewer load, because the constraint is often panel availability rather than candidate supply
Everything on that list answers a question someone can act on this week. Anything that does not belongs in a quarterly review or nowhere.
Metrics that mislead
A few numbers are popular precisely because they always look like progress.
- Applications received: a volume metric that rises when targeting gets worse
- Interviews conducted: measures activity, and goes up when screening gets weaker
- Total pipeline size: counts candidates who are not moving as though they were assets
- Average time to fill across all roles: mixes a two-week support hire with a five-month specialist search and describes neither
- Offers made: a mid-funnel count that says nothing until you know how many were accepted and how many joined
The test for any hiring metric is whether a specific person could change a specific behaviour next week in response to it moving. Applications received fails that test. Median days waiting for interview feedback passes it easily.
Frequently asked questions
What are the most important recruitment funnel metrics?
Stage conversion rate between each pair of stages, median time in each stage, time to fill and time to hire as separate medians, offer acceptance and offer-to-join rates, cost per hire by role family, source yield, and quality of hire at six to twelve months.
What is the difference between time to fill and time to hire?
Time to fill runs from the requisition being approved to the offer being accepted, so it includes approvals and role definition. Time to hire runs from a candidate entering the pipeline to that candidate accepting, so it measures the experience the candidate has.
How do you calculate recruitment funnel conversion rate?
Divide the number of candidates who reach a stage by the number who reached the previous stage. Compute it for every adjacent pair rather than only from application to hire, since a single overall ratio hides which stage is leaking.
How is cost per hire calculated?
Total internal recruiting costs plus total external recruiting costs for a period, divided by the number of hires in that period. Leaving out internal costs such as recruiter and interviewer time is the most common error.
What is a good offer acceptance rate?
It varies too much by role and market for a single figure to be useful. Compare against your own trend, and track offer-to-join separately, because with 60 to 90 day notice periods acceptance and joining are months apart.
How do you measure quality of hire?
Combine a performance assessment at six or twelve months, retention past the first year, and whether the hiring manager would repeat the hire. Hold the definition steady for several years, and use first-year attrition as the interim proxy.
Why is application volume a poor recruitment metric?
It rises when targeting gets worse, so it can improve while hiring gets harder. Pass-through rate from application to recruiter screen tells you whether the volume is the right volume.
How often should funnel metrics be reviewed?
Conversion and time in stage weekly, since both point at something fixable this week. Cost per hire, source yield and quality of hire quarterly, because they need enough hires behind them to be stable.
Start with two numbers rather than a dashboard. Conversion between each stage, and median days in each stage, for one role family you care about. Those two will show you where the funnel leaks and who is holding it up, and they will probably tell you something uncomfortable about response times rather than about candidate supply. Everything else on this page is worth adding once somebody is acting on those.
