The 4 Recruiting Metrics an Indian SMB Should Track

Track four numbers: time-to-fill, cost-per-hire (including your own screening hours), offer-drop rate, and quality-of-hire at six months. Everything else — applications per opening, resumes screened, careers-page views — is noise for a company hiring under 50 people a year. All four fit in one Google Sheet. No ATS required.

Most recruiting-metrics articles hand you a list of 23. That list is written for a talent-acquisition team of twelve at a company with a data analyst. If you are the HR generalist at a 60-person firm in Pune, or the founder doing your own hiring in Bengaluru, a list of 23 metrics is not a plan — it is a reason to track nothing at all.

So here are four, with the exact spreadsheet columns, the exact formulas, and an honest note on which one actually matters.

The dashboard you are building

SOURCE-OF-HIRE — bar length = hires, shaded = still here at 6 months Referral Naukri LinkedIn Agency

Notice what the source-of-hire strip is doing. It is not ranking channels by volume of hires. It is ranking them by how many of those hires were still there at six months. That single change is the difference between a metric that flatters your job-board spend and one that tells you where to put next quarter's budget.

1. Time-to-fill and time-to-hire are not the same thing

They get used interchangeably and they measure different failures.

• Time-to-fill = days from the requisition being approved to the offer being accepted. This measures your organisation. A long time-to-fill usually means the hiring manager is slow to give feedback, or approvals are stuck, or nobody wrote the JD for three weeks.
• Time-to-hire = days from a specific candidate entering your pipeline to that same candidate accepting. This measures the candidate's experience. A long time-to-hire is why good people take the other offer.

You can have a perfectly healthy time-to-hire of 12 days sitting inside a horrible time-to-fill of 70 days — because the req sat idle for eight weeks before anyone sourced for it. Two very different problems, two very different fixes. Collapsing them into one number hides both.

There are several competing definitions in circulation (some people end the clock at joining, not at offer acceptance). It does not matter which you pick. It matters enormously that you write your definition down once and never change it, because the only benchmark worth anything is your own number last quarter. External "average time-to-hire" figures are mostly not comparable, since the sources rarely agree on where the clock starts.

2. Cost-per-hire: the part everyone leaves out is your own time

The standard formula, as defined in the ANSI/SHRM-06001.2012 Cost-per-Hire Standard (approved February 2012), is:

Cost-per-hire = (total external costs + total internal costs) / total hires in the period

External costs are the easy ones: job-board subscriptions, agency fees, background verification, assessment tools, referral bonuses, any relocation.

Internal costs are where small companies systematically lie to themselves. The standard includes recruiter and hiring-manager time — and for an SMB, that is the majority of the real cost, because you have no agency fees to speak of and your Naukri subscription is a rounding error next to the human hours.

So price your hours. It is not hard:

Your loaded hourly cost = annual CTC / 2000      (≈ working hours in a year)
Screening cost per role = screening minutes / 60 × loaded hourly cost

If you are on ₹12L CTC, your hour is roughly ₹600. Spend three hours reading 120 resumes for one role and you have burned ₹1,800 of company money on that role's shortlist alone — before a single interview. Do that across a hiring manager, a panel of two, and a coordinator, and the "free" internal process is comfortably the largest line item on the sheet.

Track a column called screening minutes. It is the single most actionable number on the dashboard, because it is the only cost you can cut without cutting quality. Reading 120 CVs to find 8 worth interviewing is not judgement; it is data entry. The lever is either a sharper job description that stops junk applications arriving in the first place — our free keyword scanner pulls the terms out of a JD so you can see what you are actually asking for, and the ATS checker shows you how a parser reads a CV — or automating the first-pass shortlist. We build ShortlistAI for exactly that, and plenty of other tools do the same job. The point is not which tool. The point is that until you write down the minutes, you cannot prove the tool paid for itself.

3. Offer-drop rate: the India-specific number nobody tracks

Global recruiting-metrics articles have "offer acceptance rate" and stop there. That is fine for a market where a candidate starts in two weeks. It is close to useless here, because in India the notice period creates a second, later, more expensive failure: the candidate who accepts, serves 30 to 90 days at their current employer, gets counter-offered or picks up a better offer in the meantime, and never turns up on day one.

Track both stages separately. They have different causes and different fixes.

Offer-decline rate = offers declined / offers made
Renege rate        = accepted-but-never-joined / offers accepted
Offer-drop rate    = 1 − (candidates who actually joined / offers made)

A high decline rate is usually a compensation or role-clarity problem — you found out too late what they wanted. A high renege rate is almost always a notice-period engagement problem: 60 days of silence between offer letter and joining date, during which your competitors are actively calling them.

If renege is your leak, no amount of top-of-funnel sourcing fixes it. You are not short of candidates; you are losing the ones you already won. That distinction is invisible if you only track "offer acceptance rate", which cheerfully counts a renege as a success.

4. Quality-of-hire: the only one that matters, and the only one nobody computes

Everything above is a proxy. A fast, cheap hire who leaves in five months is a catastrophe that shows up as two green numbers on your dashboard. Quality-of-hire is the metric that closes the loop — and it is the one every small company skips, because it needs you to go back and look at a decision you made six months ago.

You do not need a performance-management system. You need two data points and a calendar reminder:

QoH score (0–100) = 50 × (still employed at 6 months? 1 : 0)
                  + 50 × (hiring manager's rating at 6 months / 5)

That is it. Ask the manager one question — "knowing what you know now, would you hire this person again? Rate 1–5" — and record the answer. Yes, it is subjective. A subjective number you actually collect beats an objective number you never collect. And because you are averaging it across every hire from a given source, the individual noise cancels out and the signal about the source survives.

Then join it back to source-of-hire. That is the whole game: not "which channel gives me the most CVs" but "which channel gives me people who are still here and whose manager would hire them again." Those two rankings are frequently not the same, and the gap between them is where your recruiting budget is being wasted.

The metrics to ignore

Cross these off and feel nothing:

• Applications per opening. More applications is not better. More applications is more of your ₹600 hours spent screening. A JD that attracts 300 applicants for a role that needs 8 good ones is a badly written JD, not a successful campaign.
• Number of resumes screened. A workload measure dressed up as an achievement.
• Careers-page views / social followers. Not hiring. Marketing.
• Time saved by automation. Unmeasurable and self-congratulatory. If a tool works, it will show up as a fall in screening minutes and therefore in cost-per-hire. Let it prove itself there.
• Cost per applicant. You do not want cheap applicants. You want the right one.

Build the sheet in ten minutes

One row per candidate who reached at least the offer stage — plus every hire. Columns:

The formulas, in Google Sheets:

• Time-to-fill (per hire): =F2-D2
• Time-to-hire (per hire): =F2-E2
• Median time-to-fill: =MEDIAN(FILTER(F:F-D:D, H:H="Y")) — use median, not average; one nightmare req will wreck a mean.
• Internal cost per hire: =(J2/60)*600 — swap 600 for your own loaded hourly rate.
• Cost-per-hire: =(SUM(K:K) + SUMPRODUCT(J2:J100/60, hourly_rate)) / COUNTIF(H:H,"Y")
• Offer-drop rate: =1 - COUNTIF(H:H,"Y") / COUNTA(F2:F100)
• Renege rate: =COUNTIFS(G:G,"Y",H:H,"N") / COUNTIF(G:G,"Y")
• Quality-of-hire: =50*IF(L2="Y",1,0) + 50*(M2/5)
• Source quality: pivot on column C, average the QoH score. Not the count. The average.

Set a recurring calendar reminder for the first Monday of every month: fill in column L and M for anyone who crossed six months, and read the four numbers. That is the entire operating procedure.

What to do with the numbers

Four readings, four different moves:

• Time-to-fill is high, time-to-hire is low. Your process is stalling before candidates ever appear. Fix approvals and JD turnaround, not sourcing.
• Cost-per-hire is high and mostly internal hours. Your screening is manual and your JD is too vague to filter. Tighten the JD; automate the first pass.
• Renege rate is high. Stop sourcing. Build a notice-period engagement routine — a call every fortnight, an early buddy introduction, anything that makes disappearing feel costly.
• Quality-of-hire is low from your highest-volume source. Move the budget. This is the finding that pays for the whole spreadsheet.

One honest caveat: four months of data on a handful of hires is a small sample, and you should treat any single quarter's movement with suspicion. The value is not in this quarter's number. It is in having the same four numbers, defined the same way, a year from now.

FAQ

How many hires do I need before these numbers mean anything? For time-to-fill and cost-per-hire, you get useful signal from about six to eight hires — enough for a median that isn't hostage to one outlier. Quality-of-hire and source-of-hire need more, realistically a year of hiring, because you are slicing by source. Start recording now anyway; the sheet is worthless retroactively and you cannot reconstruct screening minutes from memory.

Isn't quality-of-hire too subjective to be a real metric? Yes, it is subjective. It is also the only metric that measures whether recruiting worked. A subjective score you actually collect every month beats a rigorous one you never build. Ask the manager the same question in the same words each time — consistency is what makes the comparison across sources valid, not precision.

What is a good time-to-hire in India? Nobody can honestly tell you, and be sceptical of anyone who does. Published averages start and stop the clock in different places, mix industries and seniority levels, and are usually drawn from companies nothing like yours. Your benchmark is your own median from last quarter. Beat that.

Do I need an ATS to track this? No. Thirteen columns in a spreadsheet is genuinely enough for a company hiring under about 50 people a year, and the discipline of typing the row in yourself is what makes you notice the screening minutes. Buy an ATS when the sheet becomes the bottleneck — when multiple people need to write to it at once — not before.