A recruiter opens her inbox and finds 200 CVs, two critical vacancies and a meeting with the hiring manager in an hour. She remembers that one candidate “sounded good” on Tuesday, but she can't say which channel is working best, how long each person has been sitting in the process, or why one search is slipping.
That scene sums up the problem for a lot of recruiters, headhunters and agencies. The data is there — in the ATS, in spreadsheets, in calendars, in sourcing tools, in email — but the final call still runs on memory. People analytics turns that scattered set of signals into concrete hiring decisions, from which vacancy to prioritise to which source deserves more budget.
Why recruiters need data, not just instinct
Instinct counts for something. A senior recruiter recognises patterns, picks up nuance in an interview and knows when a profile fits even if the CV doesn't spell it out. The trouble starts when instinct replaces evidence and the team has to choose between a dozen open searches on incomplete information.
An agency can have several urgent roles running at once. Look only at the number of applicants and you might prioritise the highest-volume vacancy, even if that search converts terribly. Look instead at time-to-fill, response rate by channel, time spent in each stage and offer acceptance rate, and you can see which process is the real threat to the business.
From individual recall to a view of the funnel
People analytics doesn't require kicking off a complex business-intelligence project. The first step is simply to organise the questions the team is already asking:
- Which vacancies are stalling? Look at accumulated time per stage and the supply of qualified candidates.
- Which channel produces real conversations? Compare replies, interviews and hires — not just profiles found.
- Where do candidates drop out? Watch the conversions between sourcing, screening, interview, offer and signature.
- Which recruiter needs backup? Spot uneven workloads, repetitive tasks or stages that pile up on one person.
Modern recruiting needs this reading because a search can look active and still be stuck. The opposite happens too: a vacancy with few visible candidates may have a healthier pipeline if the people contacted actually reply and move forward.
Rule of thumb: don't start by asking how many CVs you have. Ask what decision the team has to make, and which data point can back it up.
Analytics acts as a bridge between the ATS, the candidate CRM, the spreadsheets and the outreach tools. It doesn't remove the recruiter's judgement. It frees them from remembering every detail so they can spend more attention on assessment, on the candidate relationship and on the conversation with the client.
For an agency, that difference goes straight to the ability to prioritise mandates. For an in-house team, it helps explain to the hiring manager why a role needs redefined requirements, more channels or a different outreach message. Without comparable data, every search looks equally urgent.
What people analytics is, and how a hiring team thinks with it
What people analytics is can be explained with a simple analogy: it's an X-ray of the hiring funnel. A spreadsheet can tell you how many candidates you have. The X-ray shows where the process jams, which signals keep repeating, and how the way you hire relates to what happens afterwards.
The methodology analyses people and process data to make better HR decisions. Wolters Kluwer describes people analytics as the analysis of a company's people data to improve decision-making, with metrics covering turnover, absenteeism, retention, headcount cost and workplace climate. In hiring, that approach narrows down to candidates, stages, sources, timings and hiring outcomes.
Three questions for reading the data
A team can move from a basic description all the way to a practical recommendation:
- Descriptive — what happened. How many profiles came from each source? How long did it take to fill a role? At which stage did most candidates drop out?
- Predictive — what might happen. Which vacancies are most at risk of slipping? Which kind of profile would respond better to a campaign? Which pipeline is unlikely to produce an offer?
- Prescriptive — what to do about it. Should you change the channel, adjust the filter, redistribute the workload or rewrite the outreach message?
An example helps tell them apart. The descriptive report shows that a software role is taking longer than the others. The predictive analysis detects that vacancies with a certain mix of seniority, tech stack and location tend to accumulate time in screening. The prescriptive analysis recommends opening an extra source, simplifying the initial filter or changing the outreach sequence.
The European Central Bank's privacy statement on people analytics describes the discipline in similar terms: statistical methodologies and machine-learning algorithms applied to employee data to extract patterns and build predictive models. For a recruiter, the translation is direct — use the history to decide where to invest effort, not to replace the interview.
The data suggests. The recruiter validates. The hiring manager decides with context. If a model recommends a profile, someone still has to check skills, motivation, availability and fit with the role. People analytics improves the quality of the question and of the prioritisation, but it doesn't turn a hire into an automatic operation.
Key people analytics metrics applied to recruitment
Your ATS already holds most of the information you need. The challenge isn't piling up more fields — it's connecting each metric to a decision. A useful dashboard answers what the team should do after looking at the number.
| Metric | What it measures | How it's calculated | Action it enables |
|---|---|---|---|
| Time-to-fill | Time from opening the vacancy to signature | Days between opening and acceptance or signature | Spot slow searches, adjust priorities and locate bottlenecks |
| Source of hire | The channel that delivers the hires | Hires attributed to each source, compared against the total | Shift budget and effort towards the channels that actually produce hires |
| Quality of hire | The outcome of the hire once they've joined | A combined assessment of performance, tenure and adaptation over a defined period | Review which sources and criteria produce better joiners |
| Cost-per-hire | The cost attached to each hire | Tooling, campaign, internal hours and vendor costs divided by hires | Compare vendors, channels and operational load |
| Pipeline conversion rate | Progression between stages | People moving from one stage to the next, divided by those who entered the previous one | Change filters, interviews or messaging wherever the drop happens |
How to read each indicator
Time-to-fill isn't just a speed marker. If most of the time sits between the technical interview and the client's feedback, the problem isn't sourcing. If it piles up before first contact, you may be short of suitable profiles or your message isn't landing.
Source of hire needs care too. One channel can produce lots of candidates and few hires. Another can bring fewer profiles but move them further. The comparison has to include quality and cost, not just volume.
Quality of hire closes the loop that many teams leave open. It connects hiring to later performance and tenure. Without that connection, a recruiter is only optimising movement through the funnel, not the outcome of the hire.
To go deeper on how to organise these indicators, have a look at this guide to recruitment KPIs. Stage conversion rate deserves its own reading: a drop between screening and interview can mean the criteria are too broad, while a drop between offer and acceptance can point to compensation, communication or a lack of alignment.
Cost-per-hire should account for the team's real work. If a tool removes manual tasks, the saving doesn't show up only on the vendor invoice. It also shows up as hours won back for contacting, interviewing and closing roles. The goal is for every metric to end in an operational conversation, not in a deck nobody uses.
Practical use cases for TA teams and agencies
People analytics makes most sense when the data changes a decision. A headhunter doesn't need to know everything about every candidate. They need to know which action is most likely to unblock the search in front of them.
Four common situations
Prioritising vacancies. An agency with many open roles can rank searches by commercial urgency, profile difficulty, accumulated time and probability of closing. That way it avoids spending the same effort on a vacancy with a healthy pipeline as on one that hasn't generated a single meaningful conversation.
Locating a bottleneck. If a search goes from a 60% to a 12% conversion between technical interview and offer, the team should examine that transition rather than automatically pushing more sourcing. The data can lead you to check whether the briefing is aligned, whether the technical test is too demanding, or whether the client is slow to decide. The percentages in this example are illustrative, not results attributed to any specific company.
Segmenting passive talent. A recruiter can split profiles by seniority, technical specialism, sector, location or previous company size. That segmentation makes it possible to adapt the message and compare each group's response, instead of sending the same pitch to the whole database.
Redesigning outreach. If one campaign gets an 8% reply rate and another reaches 22%, the team can compare subject line, value proposition, length, channel and timing. Those figures are a working assumption, not a published case study. The right move is to identify which variable changed before replicating the sequence.
| Case | Metric analysed | Action taken | Outcome |
|---|---|---|---|
| Search prioritisation | Time-to-fill, difficulty and pipeline progression | Assign resources first to the vacancy with the highest operational risk | More focus on the roles that could delay the service |
| Technical bottleneck | Conversion between interview and offer | Review briefing, test and feedback turnaround | Identified the stage that needed intervention |
| Passive talent segmentation | Response by seniority and specialism | Build differentiated filters and messages | More relevant outreach for each group |
| Campaign redesign | Response rate by message and channel | Compare variants and keep the most effective one | Better basis for deciding which sequence to scale |
A staffing agency can apply the same logic to shift coverage. An RPO consultancy can measure performance by client, job family and recruiter. An independent headhunter may discover they're spending far too many hours on profiles that never reply, however qualified the initial list looked.
Analytics alone doesn't guarantee an outcome. It makes the relationship between effort and progress visible, and that's the basis for fixing the process.
How to connect people analytics to your ATS and your sourcing
The architecture can stay simple as long as each tool has a clear job:
ATS ← HeyTalent ← People Analytics
The ATS keeps the transactional history. That's where the vacancy, the stages, the dates, the rejection reasons, the interviews and the status of each application live. It doesn't replace analytics, but it provides the base that makes the process measurable. If you're still working out what a tracking system should store, this explanation of what an ATS is helps separate repository, automation and analysis.
Which data to extract
Start with consistent fields, not with every field imaginable:
- Pipeline by stage: sourcing, contact, screening, interview, offer and close.
- Time per stage: entry and exit date for each phase.
- Rejection reasons: technical requirements, compensation, availability, location or client decision.
- Source: own database, referral, inbound, direct outreach or vendor.
- Downstream outcome: performance and tenure, where the organisation can legitimately connect that data.
The sourcing layer adds context the ATS may not have. HeyTalent lets you work with Boolean searches, apply customisable AI variables that return whatever profile information you define — seniority or fit with the role, for instance — and enrich profiles with contact data such as emails and phone numbers. Inside a data architecture, that information can help you segment candidates and measure response by profile type or channel.
