Recruitment Metrics and Recruitment KPIs: 14 Talent Acquisition Metrics to Track in 2026

By Sparsh Goyal, Founder at The Cognitive. Published 2026-06-05 · Last updated 2026-11-04

• Recruitment metrics, or talent acquisition metrics, measure how fast, how well and at what cost you hire, and a recruitment KPI is one you set a target for.
• The 14 recruitment KPIs fall into four stages (sourcing, interviewing, offer and quality of hire), and the stage numbers such as time to interview and drop-off tell you what to fix.
• SHRM's 2025 median time to fill is 44 days for nonexecutive roles, and Gem's 2026 benchmarks show 82% of offers accepted.

Recruitment metrics, also called recruiting metrics or talent acquisition metrics, are the numbers that show how fast, how well and at what cost you hire. A recruitment KPI is one of those metrics you have set a target for. The median US employer takes 44 days to fill a nonexecutive role, according to SHRM's 2025 Recruiting Executives Benchmarking report. Most teams track that number and still cannot say which stage is eating those six weeks.

There are 14 recruitment KPIs worth tracking, in four stages: sourcing, interviewing, offer and quality of hire. Leaders usually report five of them upward: time to fill, cost per hire, offer acceptance rate, quality of hire and first-year attrition. The stage numbers underneath, such as time to interview, scheduling time and drop-off by stage, are the ones that tell you what to fix. Each KPI below comes with its formula, what it tells you, what it does not, and a sourced benchmark where one exists.

What are recruitment metrics and recruitment KPIs?

Recruitment metrics are any measurement of your hiring process: applications per role, days in each stage, cost per hire, how long new hires stay. Talent acquisition metrics are the same numbers seen from the leadership seat, rolled up across teams and quarters. Each one is defined in our recruiting metrics glossary.

Metrics vs KPIs: what is the difference?

A metric is anything you can count. A KPI is a metric you picked because it signals progress on a hiring goal, and you gave it a target and an owner. "Time to interview" is a metric. "Time to interview under five business days for engineering roles, owned by the recruiting lead" is a KPI.

Track many metrics; hold yourself to a handful of KPIs.

Why most teams track the wrong recruitment KPIs

Most teams track the outcome but never the obstruction. They see the delay, but they miss the bottleneck causing it:

• The resume black hole: high-volume applications sit unreviewed for days because the screening criteria are too vague.
• The interview loop: candidates get stuck in a feedback vacuum while three managers try to line up their calendars.
• The "perfect fit" fallacy: roles stay open because the must-have list is really a wish list for someone who does not exist.

Tracking time to hire alone shows that something is slow, not where. The fix for a sourcing problem looks nothing like the fix for an interview problem, so the 14 KPIs below are grouped by stage. Treating them as one issue is how a team spends six months rewriting job descriptions when the real bottleneck is panel availability.

What is the formula for each recruitment KPI?

Here is every metric below with its formula and the question it answers. The first three rows are the outcome numbers leaders report; the numbered rows are the 14 stage KPIs (1 to 4 sourcing, 5 to 10 interviewing, 11 and 12 offer, 13 and 14 quality).

Cost per hire adds what you spend inside the company (recruiter pay, referral bonuses) to what you spend outside it (job ads, agency fees) and divides by hires. SHRM puts the 2025 median at $1,200 for a nonexecutive hire and $10,625 for an executive one. If your number is far off, check what you counted before you change anything.

Which sourcing KPIs should you track?

Track four sourcing KPIs: application conversion rate, source of hire, pipeline coverage and sourcing channel efficiency. Together they show whether enough of the right people enter the funnel, and at what cost.

KPI 1: Application conversion rate

What it is: the percentage of job page visitors who submit an application.

• Formula: (Applications received ÷ job page visits) × 100

What good looks like: there is no reliable public benchmark. Compare postings against each other. A posting that converts far below your others usually has a title, job description or pay transparency problem.

What it actually tells you: this is a copywriting and positioning metric more than a recruiting one. A low rate usually means your job title does not match how candidates search. It can also mean your pay range is hidden or uncompetitive, or your requirements filter out people who could do the job.

What it does not tell you: whether the people applying are qualified. A high conversion rate with low downstream offer rates means your targeting is off. You are attracting volume, not signal, and the goal is more qualified applicants, not more applications.

KPI 2: Source of hire

What it is: the split of where your hires came from: job boards, LinkedIn, referrals, agencies, direct sourcing, organic search and so on.

• Formula: (Hires from source X ÷ total hires) × 100, per source

What good looks like: no universal benchmark, as it varies by role and industry. The cross-company data does show how unequal sources are. In Gem's 2026 Recruiting Benchmarks (1 December 2025), direct sourcing delivered 11% of hires from just 2.6% of applications, and referrals converted at 11 times the rate of inbound applicants.

What it actually tells you: where to put your sourcing budget. Say 40% of your hires come from employee referrals and you spend $8,000 a month on job boards. That is a budget allocation problem.

What it does not tell you: quality of hire from each source. Source of hire only becomes a useful KPI when paired with quality of hire data (see KPI 13). A source that produces volume but poor 90-day retention is a liability, so build a cost per hire × quality of hire grid by source and defund the low-quality, high-cost sources first.

KPI 3: Candidate pipeline coverage ratio

What it is: the number of qualified candidates in your pipeline relative to open roles.

• Formula: Qualified candidates in active pipeline ÷ number of open roles

What good looks like: a common rule of thumb is 3 to 5 qualified candidates per open role. At 2 to 1, one candidate dropping out creates a crisis.

What it actually tells you: how much sourcing buffer you have. A thin pipeline is the leading indicator of a time to hire blowup, not the lagging one.

Common mistake: counting volume, not quality. Gem's 2026 data shows only 8% of applicants advance past initial screening. A pipeline of 40 unscreened applicants is not a 40 to 1 ratio.

KPI 4: Sourcing channel efficiency

What it is: the cost and time needed to generate one qualified candidate from each channel.

• Formula: (Channel spend + recruiter time cost) ÷ qualified candidates generated from that channel

What good looks like: it varies by role level and industry. The goal is comparison across your own channels, not an absolute number. For outbound sourcing, Ashby's candidate sourcing report (4 March 2024, 500,000+ sequences) found an average response rate of 19.6%, a useful reference for your own outreach.

What it actually tells you: which channels produce a qualified pipeline efficiently, not just cheaply. For example, a channel that costs $200 per qualified candidate with a 70% offer acceptance rate beats one that costs $80 per candidate with a 20% acceptance rate.

Why it matters in 2026: with 291 applications per hire (Ashby, April 2026), efficiency per qualified candidate matters more than raw volume.

Which interview KPIs show where hiring slows down?

Six interview KPIs show where hiring slows down: time to interview, interview to offer rate, scheduling time, completion rate, interviewer consistency and drop-off by stage. If sourcing looks healthy but time to hire is long, start here.

KPI 5: Time to interview

What it is: the number of days from a candidate's application to their first interview.

• Formula: Date of first interview minus date of application, averaged across candidates

What good looks like: a common rule of thumb is under five business days for competitive roles.

What it actually tells you: how quickly your team moves from "we have a candidate" to "we are evaluating the candidate." A long time to interview is almost always a capacity problem (not enough interviewer time) or a process problem (too many approvals before scheduling).

What it does not tell you: anything about interview quality or candidate fit. Moving fast to a bad interview is still bad.

The hidden cost: waiting hurts how candidates see you. In Talent Board's 2024 Candidate Experience Benchmark (230,000+ candidate responses, reported by ERE on 23 January 2025), candidate resentment was the highest Talent Board had ever recorded. The companies that won its candidate experience awards had 52% fewer candidates waiting to hear back.

KPI 6: Interview to offer rate

What it is: the percentage of interviewed candidates who receive an offer.

• Formula: (Offers extended ÷ candidates interviewed) × 100

What good looks like: set it against your own baseline per role family. A very low rate suggests the interview stage is doing filtering that should have happened upstream, so you are interviewing too many people who were never a match. A very high rate suggests the bar is low, or your sourcing is so precise that the interview adds little.

What most teams miss: the ratio varies a lot by interview type. Structured, competency-based interviews tend to give cleaner offer decisions than unstructured conversations because they filter better, not because they lower the bar.

KPI 7: Interview scheduling time

What it is: the average number of days between "interview requested" and "interview completed."

• Formula: Date interview held minus date scheduling started, averaged across candidates in a period

What good looks like: a common rule of thumb is one to three business days. Past five, scheduling has become a process failure.

What it actually tells you: panel availability drives this number more than anything else. Most companies underestimate the back-and-forth of finding a time for three interviewers, a recruiter and a candidate running another process in parallel. Interviews per hire are up by a third or more since 2021 in both Ashby's and Gem's data, so that coordination load keeps growing.

What it does not tell you: why scheduling is slow. To use it as a diagnostic, break it into parts: time to propose slots, time for the candidate to confirm, and time between confirmation and the interview. Each part points to a different fix, and interview scheduling software mainly helps with the first two.

KPI 8: Interview completion rate

What it is: the percentage of scheduled interviews that actually happen.

• Formula: (Interviews completed ÷ interviews scheduled) × 100

What good looks like: as a rule of thumb, most scheduled interviews should happen. Watch the trend more than the level.

What it actually tells you: it depends on who drops. If candidates cancel, your time to interview is too long or the process feels like a burden. If interviewers cancel, your internal process does not protect candidate-facing commitments, which is also a candidate experience problem.

What it signals with other metrics: a low completion rate plus long scheduling time usually means candidates are accepting other offers while they wait for you. The process was too slow.

KPI 9: Interviewer consistency score

What it is: how consistently different interviewers score the same or equivalent candidates against your defined criteria.

• Formula: varies by setup. It is usually measured as inter-rater reliability: the correlation between different interviewers' scores for the same candidates on the same rubric.

What good looks like: a widely used rule of thumb treats a correlation above 0.70 as reasonable consistency. Below 0.50, your evaluation is adding a lot of noise.

What it actually tells you: whether your interviews measure candidates or interviewers. High variance between interviewers for the same role means the outcome depends on who ran the interview, not on how the candidate performed. It is one of the clearest gaps in AI hiring vs traditional hiring.

Why most teams have never measured it: it needs a structured interview with written competency rubrics and more than one interviewer rating comparable candidates. When interviewers disagree, decisions get escalated, and the extra meetings stretch time to hire.

KPI 10: Candidate drop-off rate by stage

What it is: the percentage of candidates who leave the process at each stage, broken out by where the exit happens rather than as one funnel number.

• Formula: (Candidates who exited at stage X ÷ candidates who entered stage X) × 100, per stage

What good looks like: expect more drop-off at early screening and very little at later stages. A late stage that loses noticeably more people than it did last quarter warrants a look.

What it actually tells you: exactly where your candidate experience fails. A spike at scheduling means your scheduling loses people. A spike after the first interview means candidates got a negative signal, from the process itself or from a competing offer that moved faster. To hear it from the candidate's side, run a short candidate experience survey after each process ends.

Which offer and quality of hire KPIs matter?

Four KPIs cover the end of the funnel: offer acceptance rate and time to offer at the offer stage, then quality of hire and first-year attrition once people start. The quality pair is how you check whether everything before it worked.

KPI 11: Offer acceptance rate

What it is: the percentage of extended offers that are accepted.

• Formula: (Offers accepted ÷ offers extended) × 100

What good looks like: Gem's 2026 benchmarks put offer acceptance at 82%, the highest since 2021. If you sit well below that, look at pay, the candidate experience and speed, in that order.

What it actually tells you: a low acceptance rate means one of three things: your pay is uncompetitive, the hiring process created doubt, or a competitor moved faster.

The diagnostic question: log decline reasons, and ask every decline whether they accepted another offer. If most say yes, the problem is speed, not pay.

KPI 12: Time to offer

What it is: the number of days from a candidate's first interview to a formal offer.

• Formula: Date offer extended minus date of first interview, averaged across hires

What good looks like: a common rule of thumb is under seven business days for most roles. Past 14 days, decision-making has stalled.

What it actually tells you: how quickly your team moves from "we like this candidate" to "we are committing to this candidate." It captures debrief time, approval chains, compensation sign-off and offer drafting.

The pattern: a long time to offer usually means too many stakeholders, or no structured output from the interviews to decide on. Narrative feedback instead of scored evidence means more meetings before anyone decides.

KPI 13: Quality of hire

What it is: a composite score of how well new hires perform against expectations, usually measured at 90 days and one year.

• Formula: no universal formula. A common approach averages: (performance rating at 90 days + hiring manager satisfaction + retention at one year) ÷ 3. Convert each part to the same scale, for example a percentage, before averaging. Hiring manager satisfaction is a short rating the manager gives the new hire, often at 90 days.

What good looks like: benchmarks vary by role. The goal is steady improvement within your own hiring cohorts.

What it actually tells you: whether your interview process predicts job performance or just predicts who interviews well. It is the accountability metric for your whole evaluation method.

Why most teams never measure it seriously: it means connecting pre-hire evaluation data to post-hire performance data, and most companies keep those in separate systems. SHRM's 2025 benchmarking found only 20% of organizations use quality of hire metrics, down from 23% in 2017 and 27% in 2022. The most common measures among those that do are interviews to evaluate performance, performance appraisal scores and retention rates. Quality of hire is the KPI that checks every other metric.

KPI 14: First-year attrition rate

What it is: the percentage of new hires who leave, voluntarily or not, within 12 months.

• Formula: (New hires who left within 12 months ÷ total new hires in that cohort) × 100

What good looks like: it varies widely by role and industry, so set a baseline per cohort and watch the direction.

What it actually tells you: split voluntary from involuntary exits and you learn two different things. High voluntary attrition usually means the hiring process promised a different job from the one that exists. High involuntary attrition (performance exits) usually means your evaluation is not assessing competency accurately.

The feedback loop most teams miss: when a hire leaves for performance reasons, pull up their interview evaluation. If scores were high and performance was low, your criteria do not match the job. If scores were low and the person was hired anyway, the process worked but the decision did not.

What are good recruitment metrics benchmarks in 2026?

Benchmarks are useful for a sense of scale, not as targets. Check the definitions first. SHRM counts time to fill from the day the requisition opens. Ashby counts time to hire from the day the candidate applies, which is one reason its numbers are shorter.

SHRM's survey drew 2,371 responses from SHRM members between January and March 2025. Ashby's covers 109 million applications across 247,000 jobs; Gem's covers more than 165 million applicants and 1.2 million hires. For time to hire split by stage, see our guide to time to hire benchmarks.

Which talent acquisition metrics should leaders review?

A head of talent acquisition does not need all 14 KPIs every week. They need the right numbers at the right rhythm, so funnel problems get fixed fast and quality questions get enough data to answer. Here is a review rhythm that works for most teams.

Weekly numbers move fast and point to a fix you can make this week. Quarterly numbers move slowly because a cohort needs months on the job before quality of hire means anything. Mixing the two in one meeting is how quality gets ignored.

Recruiter performance metrics

Recruiter load is the context for every number above. Gem finds recruiters managing 13.4 open roles at a time. SHRM puts the median at 20 requisitions per recruiter a year in organizations with dedicated recruiters. Ashby counts 7.3 hires per recruiter per quarter by early 2026.

Compare recruiters only within the same role family, since an engineering desk and a sales desk are different jobs.

Put them on a dashboard

Chart six numbers and stop: pipeline coverage, time to interview, drop-off by stage, offer acceptance, time to fill and cost per hire. Add quality of hire once a quarter. You can build one from a prompt with the recruiting dashboard resource. If you need software that tracks these metrics for you, compare options in our guide to recruitment analytics software.

Where should you start with recruitment KPIs?

Start by finding the stage where the red clusters, then fix that stage before you add tools anywhere else. The worked example and audit below show how.

A worked example: one slow funnel

The table below is an illustration, not survey data. It shows how the KPIs might read for a team whose time to hire has drifted well past its target.

In this example, the worst KPIs cluster at the interview stage. Scheduling time, completion, drop-off and time to offer all share one root cause: interviewer time is scarce and hard to coordinate.

What fixes interview-stage KPIs, and what does not

Most companies respond to poor interview-stage numbers with one of these:

• Adding a scheduling tool: eases the back-and-forth slightly, does not create availability.
• Cutting the number of interviewers: risks under-evaluating candidates.
• Adding structured scorecards: reduces inconsistency, does not fix speed.
• Hiring more recruiters: expensive, and it does not add panel time.

None of these touches the core constraint: human interviewer time is finite, inconsistent and expensive to coordinate at scale. The change that does is replacing one or more human rounds with an AI interview round. That means a live AI video interview that asks role questions, follows up on weak answers and returns scored evidence with a transcript and recording. To compare options, see our guides to AI recruiting software and AI recruiting platforms.

The KPI audit: a practical starting point

Before you buy a tool or restructure a process, run a one-week KPI audit.

Step 1: Pull your data for the last 90 days:

• [ ] Average time to hire (total)
• [ ] Average time to interview
• [ ] Average interview scheduling time
• [ ] Interview completion rate
• [ ] Candidate drop-off rate by stage
• [ ] Offer acceptance rate
• [ ] Cost per hire
• [ ] First-year attrition (last cohort)

Step 2: Map where the red is:

• Red in sourcing KPIs (1 to 4): invest in sourcing tools or posting optimization.
• Red in interview KPIs (5 to 10): your bottleneck is the interview layer.
• Red in offer KPIs (11 to 12): look at pay and decision speed.
• Red in quality KPIs (13 to 14): your evaluation criteria are not predicting performance.

Step 3: Fix the stage with the most red before adding tools anywhere else. Most teams add tools to the stages they already manage well, because those are easier to improve. The interview layer gets skipped because fixing it means changing how evaluation happens, not just how it gets scheduled.

Where The Cognitive fits

The Cognitive is our product. It works on the front half of the funnel, where the sourcing and interview-stage KPIs above live.

• Pipeline coverage and source of hire: AI Sourcing finds people from about 900M public profiles from a plain description of the role. You can reveal a personal email, or a phone number on Sourcing Pro and above.
• Response rate: on Sourcing Pro and above, outreach emails go from your own Gmail or Outlook mailbox, signed by you.
• Time to interview and scheduling time: AI Interviews runs a live two-way AI video interview of 10 or 20 minutes on a laptop or desktop. There is no panel calendar to line up.
• Interview completion: the candidate books a time that suits them, can reschedule without emailing anyone, and there is no interviewer whose calendar can cancel it.
• Consistent criteria: every candidate is scored against the criteria you set, and you get a scored report with the recording and transcript to check it against.

AI Sourcing and AI Interviews are separate plans, each from $99 a month, and ATS sync covers 100+ applicant tracking systems on higher plans. A team that ran three human interview rounds can make the first one an AI round, cut interview scheduling for that round and keep its people for the final conversations. Offer acceptance, quality of hire and attrition still depend on your pay, your managers and your decisions, so keep measuring those with KPIs 11 to 14.

If the red in your audit sits in sourcing or the interview layer, you can start a free 7-day trial of The Cognitive, no card needed.

Sources

• SHRM, 2025 Recruiting Executives Benchmarking, 2025 (survey January 9 to March 3, 2025; 2,371 responses).
• Ashby, Recruiter Productivity, 2026 Talent Trends Report, 28 April 2026.
• Ashby, Talent Trends: Candidate Sourcing, 4 March 2024.
• Gem, Key takeaways from the 2026 Recruiting Benchmarks Report, 1 December 2025.
• ERE, 12 key takeaways from the 2024 Candidate Experience Benchmark Research (Talent Board), 23 January 2025.

All posts · thecognitive.io