Recruiting Metrics: 12 KPIs to Track in 2026 (and How AI Changes Them)

Recruiting metrics tell you where hiring is slow, expensive or leaking good candidates. Most teams track a handful, such as time to fill and cost per hire, and then add AI tools without measuring whether those tools help. This guide covers twelve recruiting KPIs with formulas and recent benchmarks, and shows which ones to watch when automation enters your process. Use it to build a one-page dashboard you can review every month.

Table of Contents

• Speed metrics
• Efficiency and cost metrics
• Quality and candidate metrics
• Metrics for AI-assisted recruiting
• How to build the dashboard
• FAQ
• Conclusion

Speed metrics

1. Time to fill

Time to fill is the number of days from opening a requisition to an accepted offer. It reflects how long the business waits for the role. JobScore's 2026 recruiting benchmarks cite SHRM's 2025 report at 44 days for non-executive roles and 45 for executive roles, and 61 and 60 days for extra-large organizations. Technical roles often run longer, which our guide to reducing time-to-hire for technical roles addresses.

2. Time to hire

Time to hire counts days from a candidate's first contact or application to their accepted offer. It reflects the candidate's experience and is usually shorter than time to fill. Track both, because a long gap between them means requisitions sit idle before sourcing starts.

3. Time in stage

Measure how long candidates wait between steps: application to screen, screen to interview, interview to decision. The slowest stage is where to automate or add capacity. Delays here also feed ghosting, covered in our guide to reducing candidate ghosting.

Efficiency and cost metrics

4. Cost per hire

Cost per hire equals total internal and external recruiting costs divided by hires in the period. JobScore cites a median of $1,340 from Appcast's 2026 report and SHRM's 2025 figures of $1,200 for non-executive and $10,625 for executive roles. Engineering roles often cost more, so compare against our cost per hire benchmark for engineering roles.

5. Source of hire

Count which channels produce actual hires, not just applicants. Job boards often produce most applicants, while referrals and outbound sourcing may produce a higher share of hires. Review this quarterly and move budget toward channels that convert.

6. Recruiter workload

Track open requisitions per recruiter and hires per recruiter per month. Overload shows up as slow responses and weak sourcing. Our article on recruiter productivity and placements per recruiter explains where agency recruiters lose their week.

7. Pipeline conversion by stage

Divide the number moving to the next stage by the number entering the current one: applicants to screens, screens to interviews, interviews to offers. A drop at one stage points to a targeting, screening or interview problem. For flood-of-applicants roles, see how to screen high-volume applicants.

Quality and candidate metrics

8. Offer acceptance rate

Accepted offers divided by offers extended. JobScore cites Appcast's 2026 average at 75%. A low rate usually signals pay, speed or a mismatch between the pitch and the job.

9. Candidate response rate

Replies divided by outreach messages sent. Reliable benchmarks are scarce, so track your own by channel and by template, then test one change at a time.

10. Candidate satisfaction

Send a one-question survey after screens and interviews. Candidate feedback is an early warning for process issues, and it matters more when interviews are automated.

11. Quality of hire

Quality of hire has no single formula. Common proxies are 90-day retention, a manager rating at 90 days and ramp time to productivity. Pick one or two, apply them consistently and compare by source and by recruiter.

Metrics for AI-assisted recruiting

If you use AI for sourcing, screening or interviews, measure whether it saves effort without lowering quality or fairness. Three checks do most of the work.

12. AI impact: three paired measures

• Recruiter hours per hire. Log time on sourcing, screening and scheduling before and after the tool. If hours do not fall within two or three hiring cycles, the tool is not paying for itself.
• Shortlist acceptance rate. The share of candidates a hiring manager accepts from an AI-assisted shortlist. A low rate means the brief or the tool is off.
• Adverse impact by group. Compare pass rates across groups at each AI-influenced stage. US agencies use a four-fifths rule of thumb in the Uniform Guidelines on Employee Selection Procedures, and some jurisdictions, such as New York City under Local Law 144, require bias audits for covered tools. Check with counsel.

Also sample rejected candidates by hand each month. A tool can look efficient while screening out good people, and only a human review will catch it. Our AI candidate screening software guide lists evaluation criteria.

Tools that report metrics natively make this easier. ConnectDevs keeps search, enrichment, outreach and AI interview reports in one workspace, so stage data and interview summaries live together. Its credit-based pricing also makes cost per search or per interview easy to calculate: a search costs 1 credit and an interview with report costs 10.

How to build the dashboard

1. Choose one metric per question. Speed: time to fill. Cost: cost per hire. Quality: 90-day retention. Candidate experience: satisfaction. AI: recruiter hours per hire.
2. Define each metric in writing. Agree on start and end dates, so the numbers mean the same thing every month.
3. Segment the data. Split by role family, level and source. A single average hides most problems.
4. Review monthly, act quarterly. Monthly numbers are noisy, so look for trends before changing process.
5. Tie each metric to an owner and an action. If nobody owns a number, it will not move.

FAQ

What are the most important recruiting metrics?

Time to fill, cost per hire, offer acceptance rate, source of hire, pipeline conversion and quality of hire cover most needs. Add recruiter workload and candidate satisfaction as you scale.

What is the difference between time to fill and time to hire?

Time to fill starts when the requisition opens, and time to hire starts when the candidate enters your process. Time to fill measures the business's wait, while time to hire measures the candidate's experience.

What is a good time to fill?

JobScore cites SHRM's 2025 benchmark of 44 days for non-executive roles and 45 for executive roles. Technical and senior roles often take longer, so compare with similar roles.

How do you measure quality of hire?

Use proxies such as 90-day retention, manager ratings and ramp time. Choose one or two, define them in writing and apply them to every hire.

How do you measure whether AI recruiting tools work?

Compare recruiter hours per hire, shortlist acceptance by hiring managers and pass rates by candidate group before and after adoption. Review a sample of rejections by hand.

How often should you review recruiting metrics?

Review monthly for trends and make process changes quarterly. Short-term numbers swing with the mix of roles you are hiring.

Conclusion

Good recruiting metrics are few, clearly defined and tied to action. Start with speed, cost and offer acceptance, add quality and candidate satisfaction, and measure any AI tool against hours saved and fairness. Compare yourself with your own trend before you worry about industry averages.

Want search, outreach and interview data in one place? Explore ConnectDevs AI candidate sourcing and run your next role through it.