Leadership & Strategy

How Big Should Your HR Team Be?

29 July 202611 min read

The HR-to-employee ratio is the most requested and least useful benchmark in people operations. It is requested because it appears to answer a real budget question in one number. It is useless in that form because four credible sources give four different answers to the same question, and none of them is measuring quite the same thing.

This piece does three things: sets out what the published figures actually say and why they conflict, explains what genuinely drives a ratio up or down, and offers a capacity-based method that is more defensible than any benchmark when you have to argue for a headcount.

Four sources, four answers
ADP Research2.6 per 100
973-company startup analysis, sub-scale2.2 per 100
SHRM Human Capital Benchmark Report1.7 per 100
CIPD survey of UK HR leaders1 per 60 (≈1.7)
Same startup analysis, scale-ups at 200+1.4 per 100
HR staff per 100 employees, by source
A spread of nearly 2x between the highest and lowest published figure. All sources are US or UK derived.
ADP Research 2.6 Startups, sub-scale 2.2 SHRM 1.7 CIPD (UK) 1.67 Startups, scale-up 1.4

Why the figures disagree

The spread is not measurement error. It comes from three definitional choices that each source makes differently and rarely states clearly.

What counts as HR. Does the recruiting team count? Payroll? Learning and development? HR systems administration? A company that runs a six-person talent acquisition team inside HR will look twice as heavy as an identical company that reports recruiting under a separate function. This single choice explains most of the gap between the ADP figure and the SHRM one.

What counts in the denominator. Permanent FTE only, or contractors and contingent workers too? In a business with a large contract workforce, the two produce very different ratios — and the contract workforce usually generates real HR work regardless of which side of the line it sits on.

Who answered. Survey panels skew towards organisations with mature HR functions, because those are the organisations that participate in HR benchmarking surveys. The direction of that bias is upward.

There is no usable India benchmark
We looked. Every figure in circulation for India is either an unsourced restatement of the SHRM number or a vendor blog citing another vendor blog. India has no equivalent of the SHRM Human Capital Benchmarking Report with a published methodology and sample. Anyone presenting you with an authoritative "India HR ratio" should be asked for the sample size and the definition of HR used. We would rather tell you the number does not exist than invent one.

The one finding that is directionally solid

Across every source, the ratio falls as companies grow. An analysis of 973 UK-based startups found a ratio of 2.2 per 100 for sub-scale companies against 1.4 for those at 200 or more employees. The same study found that most companies make their first dedicated HR hire between 40 and 50 employees, and that nearly all have at least one HR FTE by 100.

The mechanism is straightforward. Fixed obligations — a handbook, a compliance framework, a payroll process — cost roughly the same to build whether you have 60 people or 600, so they weigh far more heavily on a small denominator. Above that, specialisation and automation take over.

Do not, however, read the declining ratio as an efficiency target. A company that hits 1.4 by understaffing rather than by systematising has the same number with a very different meaning.

What actually moves your ratio

FactorDirectionWhy
Operating in multiple states or countriesUp, substantiallyCompliance work multiplies with jurisdictions, not with headcount
High-volume or continuous hiringUpRecruiting load is driven by requisitions, not by employee count
Regulatory density of the sectorUpHealthcare, financial services and manufacturing carry heavier obligations
Distributed sites or decentralised structureUpLocal presence is often required and cannot be pooled
Employee self-service and automation maturityDownRemoves transactional volume rather than judgement work
Outsourced payroll or a PEO arrangementDown, apparentlyMoves cost off the HR headcount line without removing it from the business
Shared services at scaleDownThe main reason large enterprises run leaner ratios

Note the fourth-from-last row carefully. Outsourcing improves the ratio without improving anything else, which is precisely why the ratio is a poor stand-alone metric for a board.

A better method: size by capacity, not by benchmark

If you are arguing for an HR headcount, a benchmark is a weak argument — someone in the room will produce a different benchmark. Capacity arithmetic is much harder to dismiss.

One — list the work, in categories. Recruiting, onboarding, payroll and benefits administration, employee relations, compliance and statutory filing, performance and development, systems and reporting.

Two — attach a volume driver to each. Recruiting scales with requisitions. Onboarding with joiners. Payroll with pay runs and states. Employee relations with headcount and manager quality. Compliance with jurisdictions.

Three — time the work honestly for one cycle. Not an estimate from above. Ask the people doing it to log a month.

Four — convert to FTE and compare to what you have. The gap is your case.

Worked illustration

A 250-person company hiring 60 people a year, operating in three states, running one monthly pay cycle:

Work categoryDriverIllustrative annual hours
Recruiting60 hires at 18 hours each1,080
Onboarding and exits60 in, ~42 out at 5 hours each510
Payroll and benefits admin12 cycles at 40 hours480
Compliance and filing3 states, monthly and annual420
Employee relations and manager support250 people600
Performance, L&D, reportingTwo cycles plus BAU500
Total3,590 hours

At roughly 1,700 productive hours per FTE a year, that is about 2.1 FTE — a ratio of 0.84 per 100, well below every published benchmark. Which tells you something useful: either this illustrative company is genuinely lean, or the hour estimates are too low, or the published benchmarks include work this list does not. All three are worth investigating, and the investigation is the point. Every number in that table is an assumption we chose for illustration; the method matters, the numbers do not.

How to present this to a board

Never lead with the ratio. Lead with the work and the volume drivers. Use the ratio as a sanity check at the end.

State your definition. Say explicitly whether recruiting and payroll are inside your number. Half the disagreement in the room will disappear.

Show the trend, not the point. Your own ratio over eight quarters against headcount growth is more informative than any external comparison.

Pair it with an outcome metric. A ratio on its own can only be argued down. Paired with time-to-hire, payroll accuracy or case resolution time, it becomes a capacity argument rather than a cost argument.

Frequently asked questions

What is the ideal HR to employee ratio?
There isn't one. Published figures range from 1.4 to 2.6 per 100 depending on what the source counts as HR. Use the range to sanity-check a capacity calculation, not to set a target.

When should we hire our second HR person?
When the first is consistently unable to complete the compliance and payroll cycle within the month, or when recruiting volume alone exceeds roughly half an FTE. Both are observable; a headcount trigger is not.

Does HR software change the right ratio?
It removes transactional volume — data entry, reconciliation, chasing forms — and leaves judgement work untouched. So it changes the composition of the team more than the size, and it does so mainly at the junior end.

Should recruiting be counted in the ratio?
Count it or don't, but say which. Recruiting is the single largest source of variance between published benchmarks.

The benchmark figures cited are from the SHRM Human Capital Benchmarking Report, CIPD survey data, ADP Research, and an analysis of 973 UK-based venture-backed startups, all reported via secondary compilations. All are US or UK derived; no equivalent India benchmark with published methodology was found. Every figure in the capacity worked example is an assumption chosen for illustration and is not a benchmark. Current as at July 2026. General information, not advice.