The real 8.33%–25% range, what the fee actually buys, the unit economics that make 4% impossible, and the eight questions that tell a real search firm from a CV forwarder.
Nobody in my industry wants to write this post.
Ask a recruitment agency what they charge and you'll get a number. Ask them why that number, and you'll get a pause, then something about "market standard," then a pivot to a case study. The pricing conversation in Indian recruitment is conducted almost entirely in the dark, which suits agencies fine — it's much easier to defend a number nobody understands than a number somebody has audited.
I've run Savanna HR for ten years and overseen more than 3,500 placements. I'm going to show you the actual economics: what agencies cost to run, what the fee buys, why the market splits into tiers, and why the 4–5% quote sitting in your inbox is not a bargain but a different product entirely.
Savanna HR charges at the upper end of the Indian market. This post explains exactly why — including the sections where you should push back on us, and the roles where you shouldn't hire us at all. If you want the short version: you are not paying for CVs. If you were, 4% would be too expensive.
What the Indian market actually charges
Let's start with the honest numbers. Indian recruitment fees are quoted as a percentage of the candidate's first-year annual CTC, invoiced after the candidate joins.
| Role level | Typical CTC | Standard rate | Premium / specialist |
|---|---|---|---|
| Entry level & bulk | Up to ₹8 L | 8.33% (1 month) | 10% |
| Mid level (2–5 yrs) | ₹8–20 L | 8.33–10% | 10–12.5% |
| Senior IC / Managerial | ₹20–40 L | 10–12.5% | 12.5–16.67% |
| Leadership / Niche | ₹40 L–1 Cr | 12.5–16.67% | 16.67–20% |
| CXO / Board (retained) | ₹1 Cr+ | 20–25% | 25–33% |
The reference points behind these numbers:
- 8.33% is one month's salary. That's where the number comes from — it isn't a market insight, it's 1/12. It became "standard" because it's easy to explain, not because it reflects the cost of the work.
- Industry surveys of the Indian market consistently put permanent recruitment between 8.33% and 16.67% for the bulk of roles, with executive search at 25–33%.
- Globally, the same work prices higher: contingency search in the US and UK runs 15–25%, retained 25–35%. India's rates are roughly half the global benchmark for identical work.
- Contract/flexi staffing is priced differently — typically 10–25% markup on monthly CTC, billed monthly, not a one-time fee. Don't compare the two.
You hire a Plant Head at ₹45 L CTC. At 12.5%, the fee is ₹5,62,500, invoiced once, 30–45 days after joining. At 8.33%, it's ₹3,74,850. The delta is ₹1,87,650 — roughly 0.4% of that person's first-year cost to your business, and a rounding error against what they'll influence in output. Hold that number.
The five categories of recruitment agency
"Recruitment agency" is a category error. These are five different businesses that happen to share a name. Confusing them is the single most expensive mistake in vendor selection.
01 The CV Aggregator
Buys a Naukri or Monster database licence. Runs keyword searches against your JD. Forwards whatever matches. No screening beyond a phone call confirming the candidate is alive and interested.
The recruiter handling your mandate is likely working 15–25 roles simultaneously and has never spoken to your hiring manager. They are not evaluating candidates; they are routing them.
This is a legitimate business. For high-volume, low-differentiation roles with a deep candidate pool — telecallers, entry-level ops, bulk field sales — it is genuinely the correct choice. Paying 12% for a role where the database has 4,000 qualified matches is you being overcharged, not them being cheap.
02 The Generalist Contingency Firm
The Indian market's mainstream. Some real screening, some client relationship, a recruiter who's read the JD properly. Works across functions and industries without deep specialisation anywhere.
Competent for mid-level roles with reasonable talent supply. Struggles the moment a role needs domain judgement — because the recruiter can evaluate whether a CV matches the JD, but not whether the person can actually do the job.
03 The Specialist Search Firm Where we sit
Picks a domain — manufacturing, fintech, academic faculty, GCC leadership — and refuses to leave it. The recruiter has placed this exact role thirty times. They know the twelve companies worth poaching from, the four who overpay, the two whose "Senior Manager" title is actually a Team Lead.
The economics only work if you say no to most inbound. We turn down mandates outside our practice areas, and we turn down mandates inside them where the brief is broken and the client won't fix it. That refusal is the product.
04 Retained Executive Search
Paid in thirds — on engagement, on shortlist, on joining — regardless of outcome. You're buying committed capacity and a genuine market map, not a lottery ticket.
Correct for CXO hires, board roles, confidential searches, and any mandate where the wrong hire is an existential problem rather than an operational one.
05 RPO / Embedded
Recruiters embedded in your team, working only your roles. The percentage looks low because you're paying for time, not outcomes — the risk has moved to you.
Makes sense above roughly 40–50 hires a year. Below that, you're paying for idle capacity.
Categories 1 and 5 both quote you 5–8%, and they are opposites. One transfers all risk to you cheaply. The other transfers all risk to you expensively. Neither is "an agency that charges less."
What actually happens between brief and joining
Here's the part clients rarely see, and it's the reason the fee isn't a finder's fee. Take a real mandate shape: Head of Quality at a mid-size auto component manufacturer, Pune, ₹38 L.
-
The brief 3–5 hrs
The JD says "10+ years in quality, IATF 16949 experience, team handling." Useless. Every quality head in Pune matches it. The real brief only surfaces in conversation: they've failed two customer audits, the last incumbent was a documentation person not a shop-floor person, and the actual need is someone who'll walk the line at 6 AM.
That's not on any JD. Extracting it takes a real conversation with the person who owns the outcome — and a recruiter working 20 mandates never gets that conversation.
-
Market mapping 8–15 hrs
Not searching a database. Building the universe: every auto component manufacturer within a 60 km radius, their quality org structures, who reports to whom, which plants have IATF-certified processes, which have had recent audit failures. Typically 150–400 named individuals, most of whom are not looking and will never appear on a job board.
The 4–5% agency does not do this step. There is no version of the economics where they can.
-
Outreach 15–25 hrs
The best candidate for this role is employed, performing, and not on Naukri. Reaching them takes a credible approach from someone who understands their work well enough to be worth replying to. Response rates on cold senior outreach run 10–20% at best. That's 300 approaches for 40 conversations.
-
Screening 10–20 hrs
The expensive part, and the part that gets skipped first. Not "do you have IATF experience" but "walk me through the last audit non-conformance you closed, what the root cause turned out to be, and what you changed in the process." Twenty minutes of that separates the person who ran quality from the person who was present while quality happened.
Savanna HR runs at a 98% CV qualification rate — 98% of profiles we submit are accepted as legitimately screenable. The industry norm is closer to 30–50%. That gap is entirely this stage.
-
Calibration 5–10 hrs
First shortlist is partly a diagnostic instrument. The client's reactions reveal the requirements they couldn't articulate. Recalibrate, re-map, go again.
-
Process management 10–20 hrs
Interview coordination, panel wrangling, feedback chasing, keeping candidates warm across a 6-week process where your hiring manager goes quiet for 12 days. Every day of silence costs offer acceptance probability.
-
Offer and close 5–15 hrs
Where mandates die. Counter-offers, notice period negotiation, the spouse who doesn't want to relocate, the competing offer that arrives at hour eleven. India's average offer acceptance rate sits around 75% — one in four offers is declined. Managing that down is a skill, and it's exercised at the moment the agency has already done 100% of the work and been paid 0%.
-
Joining and guarantee 3–8 hrs
The 60–90 day window between offer acceptance and actual joining is when Indian hires evaporate. Staying in contact through notice period is unglamorous, uncompensated, and the difference between a placement and a restart.
60–120 hours of skilled work per mandate, against roughly 45–70 days elapsed for a mid-senior Indian search. Remember this number for the next section.
The economics nobody shows you
Now the arithmetic. This is the section my peers will dislike.
The cost base of one specialist recruiter
| Line item | Annual cost |
|---|---|
| Recruiter CTC (specialist, 5+ yrs, metro) | ₹8,00,000 – ₹14,00,000 |
| LinkedIn Recruiter licence | ₹4,00,000 – ₹8,00,000 |
| Job board / database licences | ₹1,50,000 – ₹3,00,000 |
| ATS, tooling, assessment platforms | ₹75,000 – ₹2,00,000 |
| Allocated overhead (office, admin, legal, finance) | ₹2,50,000 – ₹4,00,000 |
| Fully loaded cost of one recruiter | ₹16,75,000 – ₹31,00,000 |
Now capacity. At 60–120 hours per mandate, a specialist recruiter running proper search can hold 8–12 live mandates at a time and close realistically 18–30 placements a year at a 60–85% fill rate.
Take the conservative middle: fully loaded cost ₹22 L, 22 placements a year. Break-even cost per placement: ₹1,00,000.
Don't take my word for it. Move the sliders.
The recruiter break-even calculator
Set the agency's real cost base and the rate you want to pay. Watch what's left to actually work your mandate.
Here's the same thing as a fixed table, at an average mandate CTC of ₹25 L:
| Rate | Fee per placement | Gross margin | Viable? |
|---|---|---|---|
| 4% | ₹1,00,000 | ₹0 | Break-even at best |
| 5% | ₹1,25,000 | ₹25,000 | 20% — no reinvestment |
| 8.33% | ₹2,08,250 | ₹1,08,250 | Works at high volume |
| 12.5% | ₹3,12,500 | ₹2,12,500 | Sustains specialisation |
| 16.67% | ₹4,16,750 | ₹3,16,750 | Sustains retained-grade rigour |
At 4%, a firm doing genuine search loses money on every placement. Not "makes less" — loses.
So how do 4–5% agencies exist? They change the product. They cannot afford 22 placements a year per recruiter, so they need 60+. That means 60–120 hours per mandate becomes 8–15. Market mapping goes first. Deep screening goes second. Offer management goes third. What's left is database keyword matching and CV forwarding — which is exactly the service you receive.
And critically: they need a low fill rate to be survivable, so they run huge volume. If your fill rate is 15%, every successful placement must fund the six that failed. Your mandate isn't being worked. It's being held as a lottery ticket alongside forty others.
You are not funding your search. You are funding the 85% of searches that failed.
Why 4–5% cannot work
A recruitment fee is not a margin on a product. It's the price of a process. Lower the price and the process must shrink to fit — there is no third option. The work does not compress. It gets deleted.
Here's what specifically gets deleted at 4–5%, in the order it goes:
- The brief conversation. Replaced by reading the JD. Result: they're searching for the role you wrote down, not the role you have.
- Market mapping. Replaced by a database keyword query. Result: you see only candidates who are actively looking — roughly 15–20% of the qualified universe, and systematically the weakest slice of it. The best person for your role is employed and not applying anywhere. At 4%, they will never be contacted.
- Deep screening. Replaced by a 4-minute availability call. Result: your hiring managers become the screening layer. You are now doing the work you outsourced, at senior salaries, and paying 4% for the privilege.
- Calibration. Replaced by nothing. Same wrong profiles, resubmitted.
- Offer management. Replaced by hope. Your acceptance rate collapses toward the 75% baseline and below.
- Joining support. Replaced by an invoice.
Say the 4% agency sends 40 CVs for that ₹38 L quality role. Your hiring manager screens them at, conservatively, 15 minutes each. That's 10 hours. A senior manager's loaded cost runs ₹4,000–8,000 an hour. You just spent ₹40,000–80,000 of internal time doing the screening you paid an agency to do — against a "saving" of ₹1,87,650 versus a 12.5% firm that would have sent you four.
Add the two extra interview rounds because the shortlist was weak. Add the six weeks of vacancy because the first three shortlists failed. The 4% agency stopped being cheaper somewhere around week two, and nobody put it in a spreadsheet.
And watch for the fees that reappear elsewhere. The low headline rate is frequently recovered through database access charges (₹5,000–25,000/month), job posting fees (₹3,000–15,000 per listing), and — the expensive one — a replacement guarantee shortened to 15–30 days when the Indian notice period alone is 60–90 days. A 30-day guarantee on a hire who joins after a 90-day notice is not a guarantee. It is a document.
The fluke hire problem
Here's the objection I get, and it's a fair one: "We've hired great people at 5%. It works."
Yes. Sometimes. And that's precisely the trap.
At a 15–20% fill rate with high volume, low-rate agencies do produce good hires. Push 40 CVs at a role and occasionally one is excellent. This is not a swindle — it's mathematics. Volume produces outliers.
But understand what you're observing. You are not seeing a repeatable process. You are seeing survivorship bias with an invoice attached.
The fluke hire has three properties worth knowing:
- It cannot be commissioned. It happened. You can't order another one. Every subsequent mandate is a fresh roll of the same dice, and you have no idea what the odds are because nobody publishes their fill rate.
- It is invisible in your cost accounting. You remember the one that worked. You don't have a line item for the eleven weeks of vacancy on the two mandates that failed first, or the ₹60,000 of hiring manager time spent screening noise.
- It carries unpriced downside. This is the one that matters.
Because the mechanism producing the fluke hire — no screening, no calibration, volume over judgement — is the identical mechanism that produces the bad hire. Same process, different outcome. And in India, the bad hire is catastrophically expensive.
The commonly cited "30% of salary" figure is imported from US and European markets where notice periods run 2–4 weeks. In India, that number is wrong. With 60–90 day senior notice periods, a re-recruitment cycle of 8–12 weeks, and a 60–90 day ramp for the replacement, credible Indian estimates for mid-senior bad hires land at 50–100% of annual CTC, with 60–80% a sensible planning assumption. Total elapsed time from "this isn't working" to "fully productive replacement" is rarely under 9 months.
Run it on the ₹38 L quality head:
| 12.5% specialist | 4% aggregator | |
|---|---|---|
| Fee | ₹4,75,000 | ₹1,52,000 |
| Apparent saving | — | ₹3,23,000 |
| Bad hire probability | ~10% | ~35% |
| Cost if bad (at 65% of CTC) | ₹24,70,000 | ₹24,70,000 |
| Probability-weighted cost of failure | ₹2,47,000 | ₹8,64,500 |
| True total expected cost | ₹7,22,000 | ₹10,16,500 |
The ₹3.23 L saving is worth negative ₹2.94 L.
And this model is generous to the aggregator — it ignores vacancy cost entirely. On a role where an empty seat costs even ₹3 L a month in lost output, six extra weeks of vacancy adds ₹4.5 L and the comparison stops being close.
The 4% fee isn't a discount. It's an uninsured deductible.
Then there's the second-order cost nobody models. When a senior hire fails in an Indian manufacturing or BFSI context, people follow them out. One production head's exit at a Manesar unit triggered five more departures within three months — one ₹24 L departure that became ₹1.34 Cr of total damage. Your 4% saving is a rounding error against that.
What premium actually buys
So what are you buying at 12.5–16.67% that you don't get at 8.33%? Not effort — everyone claims effort. Here's what's actually different, stated as things you can verify:
Refusal
We say no to mandates. Outside our practice areas, or inside them where the brief is broken and the client won't engage on fixing it. An agency that accepts every mandate has a fill rate that reflects it — and you're funding the ones they never had a chance at. Ask any agency what percentage of inbound mandates they decline. If the answer is "we're always happy to help," you now know their model.
Fill rate, disclosed
The number that matters and nobody volunteers. Specialist search should run 60–85%. Contingency generalists run 20–40%. Aggregators run 10–20%. When we take a mandate, we intend to close it, because we've already declined the ones we couldn't.
A map, not a search
For any mandate we take, we can name the universe — the companies, the org structures, the people. Including the ones who aren't looking, which is where the actual talent is. This is the single largest difference between what 4% buys and what 12.5% buys, and it's invisible in the CV you receive.
Screening that has an opinion
A shortlist of four with a written argument for each, and — more usefully — a written argument for the three we rejected who looked good on paper. That's what a 98% CV qualification rate means in practice: your hiring managers stop being the filter.
Domain judgement
Our manufacturing practice knows which plant's "Senior Manager" is your "Team Lead." Our academic practice knows an FPM is not an MBA and that a PhD "pursuing" is not a PhD. Our fintech practice knows an RBI-designated Compliance Officer is a legally distinct thing from a Compliance Manager. This knowledge is only accumulable by refusing to work outside a domain — which is exactly what the 4% model cannot afford.
Confidentiality that actually holds
Stealth searches, sensitive replacements, mandates where naming the client before offer stage would be damaging. Nearly impossible at volume, because volume requires broadcasting.
Retention, measured
Our reference point is 96% first-quarter retention. Placement is not the deliverable. Someone still performing at month twelve is the deliverable.
For a bulk hiring mandate with 4,000 qualified candidates on Naukri, none of the above is worth 12.5%. You should not pay us for that role. We'd tell you so, and we'd be right.
When you should NOT pay premium
I'd rather you push back on us in the right places than nod along and resent the invoice.
- The talent pool is deep and undifferentiated. Telecallers, entry-level ops, bulk field sales. The database is genuinely the right tool. Pay 8.33% and don't apologise.
- You're hiring 50+ of the same profile. That's an RPO engagement. Stacking percentage fees is you funding someone's business model.
- You have a strong internal TA function with real market access. You need overflow capacity, not judgement.
- The role is genuinely easy. Some are. Any agency that says every role is hard is selling.
- You want one vendor to also be your ATS, employer brand, and comp consultant. Buy those separately, from people who do them properly.
- The best candidate is employed and not looking — which is most senior roles
- The role requires domain judgement your interviewers don't have
- The search is confidential
- The talent pool is under ~500 people nationally
- A bad hire costs more than 50% of CTC — which in India, at mid-senior levels, is nearly always
- Time-to-fill has direct revenue consequence
- You've already failed on this role once
The test isn't "is this expensive." It's "what does failure cost." If failure is cheap, buy cheap. If failure is expensive, buying cheap is the risk.
How to audit any agency in eight questions
Use these on us. Use them on everyone. The answers are more diagnostic than any pitch deck.
- "What's your fill rate on mandates like this, last 12 months?"Anything under 40% means you're a lottery ticket. A firm that doesn't track this doesn't manage it.
- "How many mandates is the recruiter on my role handling right now?"Above 12, real search is arithmetically impossible.
- "What percentage of inbound mandates did you decline last year?"Zero means no specialisation and a fill rate to match.
- "Show me a market map for a role you closed recently."They'll have one or they won't. This question ends most conversations.
- "How many CVs will you send?"Fewer is better. Forty is a confession. Four with written rationale is a search.
- "What's the guarantee period, and does it start at offer or at joining?"Must exceed the notice period. 90 days from joining is the floor in India. 30 days is theatre.
- "What's your first-year retention on placements from the last two years?"Placement rate flatters. Retention tells the truth.
- "Walk me through a mandate you failed and why."Anyone who's never failed hasn't taken hard mandates — or isn't being straight with you.
"What does this seat sitting empty cost per month?" Most companies have never calculated it. Until you do, every fee conversation is happening without the denominator.
The uncomfortable summary
The Indian recruitment market has been trained to negotiate the one number that's visible and ignore the four that aren't.
The fee is visible. Vacancy cost, hiring manager time, offer acceptance rate, and bad-hire probability are not. So the fee gets negotiated to the floor, the process shrinks to match, the failures get absorbed into the general noise of "hiring is hard," and everyone concludes the agency market is low-value — which becomes self-fulfilling, because at 4% it is.
The rate isn't the price. The rate is a description of the process you're going to get.
At 4%, someone will run a database query and forward the results. That is genuinely what 4% buys, and for some roles it's the right purchase. At 12.5%, someone maps the market, approaches people who aren't looking, screens with an opinion, and manages the close. Also correctly priced — for a different job.
The mistake isn't paying too much or too little. It's buying the first product and expecting the second.
If you're negotiating a fee right now
Take the eight questions above to whoever's quoting you — including us. If we can't answer them straight, don't hire us.
And if you want to talk through what your specific role actually needs, including whether it needs a premium firm at all, that conversation is free and comes with no obligation. Sometimes the honest answer is "this is an 8.33% role, here are three firms who'll do it well." We'd rather tell you that than take a mandate we shouldn't.
Talk to our team →Frequently asked questions
What is the standard recruitment agency fee in India in 2026?
Standard permanent recruitment fees in India range from 8.33% to 16.67% of the candidate's first-year annual CTC. Entry-level and bulk roles typically attract 8.33% (one month's salary), mid-level roles 8.33–12.5%, senior and leadership roles 12.5–16.67%, and CXO/executive search 20–33% on a retained basis. The fee is a one-time charge invoiced after the candidate joins, usually payable within 30–45 days.
Why do recruitment agencies charge 8.33%?
Because 8.33% is exactly one month's salary — 1/12 of annual CTC. It became the Indian market default because it's easy to explain and easy to benchmark, not because it reflects the actual cost of conducting a search. It's a convention, not an economic calculation.
Is the fee calculated on total CTC or fixed CTC?
Almost always total first-year CTC, including base, variable, and allowances. ESOPs and equity are typically excluded, though this varies. For roles with a large variable component, confirm this upfront in writing — on a ₹30 L offer with 20% variable, the difference is material.
Can I negotiate recruitment agency fees in India?
Yes, and there are legitimate levers: exclusivity, multiple mandates, guaranteed volume, or faster payment terms all justify a lower rate because they change the agency's economics. What doesn't work is negotiating rate alone with no change to the deal — the agency simply reduces the work to match. Negotiate the deal structure, not just the number.
Why can't recruitment agencies deliver quality at 4–5%?
Because the arithmetic forbids it. A fully loaded specialist recruiter in India costs ₹17–31 lakh a year and can properly close 18–30 placements annually, putting break-even around ₹1 lakh per placement. At 4% on a ₹25 L role, the fee is exactly ₹1 lakh — zero margin. To survive, low-rate agencies must run 3–5x the mandate volume per recruiter, which cuts per-mandate effort from 60–120 hours to 8–15. That deletes market mapping, deep screening, and offer management. What remains is database keyword matching.
Do cheap recruitment agencies ever produce good hires?
Yes — through volume, not process. At a 15–20% fill rate with high throughput, occasional excellent hires occur. But this is survivorship bias, not a repeatable capability. The same absence of screening that produces the occasional fluke hire produces bad hires at a much higher rate, and in India a bad mid-senior hire costs 50–100% of annual CTC with a 9–12 month recovery window.
What is the real cost of a bad hire in India?
Between 50% and 100% of annual CTC for mid and senior roles, with 60–80% a reasonable planning assumption. The widely quoted "30% of salary" figure comes from US and European data where notice periods run 2–4 weeks. India's 60–90 day notice periods, 8–12 week re-recruitment cycles, and 60–90 day ramp times make the real number substantially higher. Total elapsed time from identifying a bad hire to a fully productive replacement is rarely under nine months.
What replacement guarantee should I expect?
90 days from the candidate's joining date is the Indian market standard, and it should be your floor. Critically, check whether the clock starts at offer acceptance or at joining — with 60–90 day notice periods, a guarantee that starts at offer can expire before the candidate walks in. A 15–30 day guarantee offers effectively no protection.
What hidden fees should I watch for?
Database access charges (₹5,000–25,000/month), job posting fees (₹3,000–15,000 per listing), short or offer-dated replacement guarantees, and ownership clauses letting the agency claim a fee if you hire any candidate they ever sent, sometimes for 180 days. A low headline rate is frequently recovered through these. Ask for the full commercial terms in writing before comparing rates.
How many CVs should a good agency send per role?
Four to six for a properly scoped senior mandate, each with written rationale. Forty CVs is not thoroughness — it's an admission that no screening happened and that your hiring managers are now the filter. Fewer, better-argued profiles is the single clearest signal of a real search process.
What's the difference between contingency and retained search?
Contingency means the agency is paid only on successful joining — you carry no financial risk, they carry all of it, and they manage that risk by working many mandates simultaneously. Retained means staged payment (typically thirds: engagement, shortlist, joining) regardless of outcome — you carry risk, and in exchange you get committed capacity, genuine market mapping, and exclusivity. Retained is standard for CXO, confidential, and business-critical searches.
How do I know if a recruitment agency is actually specialist or just claiming to be?
Ask what percentage of inbound mandates they declined last year. Genuine specialists refuse most work outside their domain — that refusal is what allows the domain knowledge to accumulate. An agency that accepts everything is a generalist regardless of what the website says. Then ask them to name the candidate universe for your role. Specialists can. Generalists search.
- Staffing Industry Analysts (SIA) — global contingency and retained fee benchmarks
- SHRM — staffing agency fee guidance and cost-of-bad-hire framework
- Keka × SHRM India — HR Pay Pulse Report 2026 (talent acquisition compensation)
- Gallup — employee replacement cost research
- Indian market fee surveys, 2025–26 (TIGI HR, PlacementIndia, Wisemonk)
- Savanna HR internal placement data, 2015–2026
