Sometimes a Recruitment Agency Is Expensive. Sometimes Not Having One Costs More.

19 Nov 2026

Imagine you’re a founder and an agency sends you a proposal to recruit a software engineer. The salary is around $136,000, the fee is 20%, and suddenly there’s a $27,196 invoice to consider. I understand why the first reaction might be to ask whether the company could just hire the person itself, especially when there’s already an HR team, a careers page and someone who can start approaching candidates tomorrow.

What I’d want to know before making that decision is how much recruitment is already costing the business. How many positions are we actually going to fill this year? How much work is waiting for the engineer we’re trying to hire? How many hours are our managers spending on interviews and follow-ups? And if the person we eventually hire doesn’t work out, how expensive will it be to start again?

Those numbers are rarely sitting in one spreadsheet. The agency fee appears as a very visible invoice, while the cost of managing a vacancy can be spread across salaries, delayed projects, engineering time and budgets that nobody thinks of as recruitment expenses.

I’ve seen why companies hesitate to pay for outside recruitment, and I think that hesitation is reasonable. There are searches where an agency is more expensive than doing the work internally. There are also situations where avoiding the fee can cost considerably more, so I wanted to bring the calculations together and show how to tell the difference.

For this article, I’ve used published US salary data and recruitment software prices available in November 2026. The examples are built around a US-based software engineering role, because those figures are independently verifiable. The financial assumptions about vacancy losses, interview hours and failed hires are clearly identified as illustrations. A company in Dubai, Cyprus, Armenia or elsewhere should replace the US salary and employment-cost inputs with its own market data.

1. How much does an agency actually charge?

For permanent direct-hire recruitment, fees are commonly expressed as a percentage of the employee’s first-year compensation. Upwork’s 2026 guide places direct-hire fees at approximately 18% to 25%, while SHRM has described 20% to 25% as a common range for contingency recruitment and around one-third for some retained executive searches. The agreed percentage, the definition of compensation, the payment milestones and the replacement terms all matter. [1][2]

Let’s take a software engineer earning $135,980 per year, the median US annual wage for software developers reported by the Bureau of Labor Statistics for May 2025. This is a broad occupation benchmark, not a quote for every senior engineering specialisation. [3]

Agency placement fee at 15%, 20% and 25%
Agency feeCalculationPlacement fee
15%$135,980 × 0.15$20,397
20%$135,980 × 0.20$27,196
25%$135,980 × 0.25$33,995

Placement fee at three common fee levels, applied to the BLS median US software-developer wage ($135,980, May 2025). Not a quote from any particular agency.

At a 20% fee, the agency invoice is $27,196. This is a real calculation based on a documented salary benchmark and a rate within the published market range, though it is not a quotation from a particular agency. Depending on the contract, the percentage might apply to base salary, total cash compensation or another defined amount. Taxes and any additional contracted services could also change the final bill.

Now let’s compare that visible invoice with a less visible cost: employing an internal recruiter.

2. What does an in-house recruiter actually cost per year?

The Bureau of Labor Statistics reports a $75,940 median annual wage for US human resources specialists in May 2025. Recruitment specialists are included in this occupational group, although a dedicated senior technical recruiter may command a different salary. [4]

Base salary is only the beginning. In June 2026, the Bureau of Labor Statistics measured wages and salaries at 70% of total private-industry employer compensation and benefits at 30%. In other words, the aggregate benefits-to-wages ratio was approximately 42.9%. Applying that national ratio to a specific recruiter is an approximation, because actual benefit costs differ by employer, occupation and state. It is still much more transparent than pretending an employee’s salary is the complete cost. [5]

For a company considering its first dedicated recruiter, here is one model using those published figures and entry-level software subscriptions:

Estimated annual cost of a new in-house recruiter
Annual recruitment capacityCalculationAnnual cost
Recruiter base payBLS occupation median$75,940
Estimated employer benefits and additional compensation$75,940 ÷ 0.70 − $75,940$32,546
LinkedIn Recruiter Lite$170/month × 12$2,040
Workable Standard ATS$299/month × 12$3,588
Estimated annual fixed cost$114,114

One model of a new dedicated recruiter’s annual cost, using BLS median pay, the BLS aggregate benefits ratio and entry-level tool list prices.

For reference, the $170 LinkedIn figure is the widely reported US monthly price for a single Recruiter Lite seat (LinkedIn shows the final price at checkout); the $299 Workable Standard figure is the advertised monthly rate for its smallest company-size tier. Recruiter Lite has limitations, while full LinkedIn Recruiter is custom-priced. Larger businesses may pay more for their ATS or need other sourcing tools, assessments and advertising. Some smaller businesses could spend less, and companies with an existing ATS should not charge its full subscription a second time. [6][7][8]

This $114,114 is an estimated yearly capacity cost, not a claim that every American recruiter costs exactly that amount. It excludes extra job advertising, some overhead and hiring managers’ time. It also assumes the employer is deciding whether to add a new dedicated recruiting position, rather than counting a recruiter already employed and paid regardless of this decision.

Here is the question I would ask next: how many successful hires will that person actually make for the company over the next 12 months?

3. At what hiring volume does an in-house recruiter become cheaper?

To keep the first comparison readable, we’ll look only at dedicated recruitment capacity. Hiring-manager interviews, candidate assessments and onboarding still happen whichever route a company chooses, so those common costs are excluded here. Specific advertising expenses and other differences would need to be added for a real decision.

We will use $114,114 in annual internal recruiting capacity and $27,196 per agency placement, assuming the same salary and a 20% success fee for every role.

Annual cost by number of successful hires
Successful hires in one yearNew in-house recruiter: fixed capacityAgency: 20% per placementLower cost in this model
2$114,114$54,392Agency by $59,722
4$114,114$108,784Agency by $5,330
5$114,114$135,980In-house by $21,866
6$114,114$163,176In-house by $49,062
8$114,114$217,568In-house by $103,454

Dedicated recruitment capacity only; costs common to both routes are excluded. The lower-cost option in each row is highlighted.

Here is what that comparison looks like visually. The switch happens between four and five successful hires per year in this particular example, because a new in-house recruiter is a fixed annual commitment while an agency fee rises with every placement.

Figure 1 · Annual recruitment cost

The number of hires changes the answer

A new dedicated in-house recruiter costs the same whether they make two hires or six. Agency fees rise with every placement. Illustrative US example.

Chart summary: at 2 hires the agency costs $54,392 versus $114,114 for a new in-house recruiter, so the agency is $59,722 cheaper. At 4 hires the agency costs $108,784, $5,330 cheaper. At 5 hires the agency costs $135,980, so in-house is $21,866 cheaper. At 6 hires the agency costs $163,176, so in-house is $49,062 cheaper. Break-even is about 4.2 hires.

1–4 hiresThe agency costs less
5+ hiresA new in-house recruiter costs less
Figure 1. Illustrative annual recruitment-capacity comparison: a $114,114 fixed cost for a new dedicated recruiter versus a 20% agency fee on a $135,980 role. Costs shared by both routes are excluded, and the model assumes the internal recruiter can complete the hires. Exact figures are in the table above.

The break-even equation is simple:

$114,114 annual internal capacity ÷ $27,196 agency fee = 4.20 successful hires.

In this model, the agency is less expensive for one to four placements, and the new internal recruiter becomes the lower-cost option from five placements onward. The exact point depends on how much recruitment work the employee can complete successfully, whether they contribute other valuable work, what the agency charges, and whether the company’s hiring demand is stable.

The agency’s percentage changes that break-even point quite substantially:

Break-even volume by agency fee
Agency feeFee on a $135,980 roleApproximate break-even volumeNew in-house recruiter becomes cheaper at
15%$20,3975.59 hires6 hires
20%$27,1964.20 hires5 hires
25%$33,9953.36 hires4 hires

How the agency’s percentage moves the break-even point, using the same $114,114 annual internal capacity.

A company planning to hire ten developers regularly may have an excellent business case for its own talent acquisition team. A company hiring two specialised engineers this year may find it much more expensive to create a full-time recruiting role that it cannot keep productively occupied. It is also possible that the employer will recruit some roles internally and bring an agency into only those searches where specialist networks or extra capacity make a difference.

There is another qualification that matters a great deal: a recruiter who is already employed is a different calculation. If the company pays that person whether it uses an agency or not, the salary is already committed. The agency needs to produce a measurable additional benefit on that individual search, and we should compare the incremental costs accordingly.

4. How much does leaving a position unfilled really cost?

I often see people estimate vacancy costs by taking an employee’s annual salary and dividing it by the number of working days. That method produces a salary figure, but it doesn’t tell us what the company is losing by not having the person there. During a vacancy, the employer may temporarily save the pay and some benefits that would otherwise have gone to the missing employee.

The relevant number is the incremental business impact of the vacancy, less costs the company genuinely avoids. For a sales position, that could be contribution margin lost from sales that will never be recovered. For an engineer, it might be the financial impact of a delayed project, contractor invoices, overtime or a missed customer commitment. Revenue that is simply postponed shouldn’t automatically be counted as a permanent loss, and a contractor who delivered the work must receive credit for that output.

Here’s a worked example of a critical technical vacancy lasting an extra 20 working days. These four inputs are hypothetical figures from a company’s internal project ledger, not industry-average costs:

Net economic cost of a 20-day vacancy
Impact of an additional 20 working daysCalculationFinancial impact
Contribution lost from work that will not be recovered, after crediting temporary cover’s output20 × $1,600+$32,000
Additional overtime or external coverage actually paid20 × $200+$4,000
Employee wages and benefits genuinely avoided during those days20 × $600−$12,000
Net economic cost of the 20-day vacancy$32,000 + $4,000 − $12,000$24,000

Hypothetical inputs from one company’s internal project ledger, not industry averages.

This company is experiencing a $1,200 net daily economic loss, because the vacancy is holding up valuable work. Another company could experience a $200 daily loss, or almost none at all. A team able to absorb the workload with no meaningful delivery impact shouldn’t manufacture a vacancy cost to justify engaging a recruiter.

The $600 daily avoided-pay figure is a specific assumption for this example, rather than the BLS salary mechanically divided into days. Not all benefit costs disappear immediately when a position remains vacant, and some fixed costs remain in place. A finance team should enter the amounts it can actually avoid.

Now suppose an agency can credibly help this company get a productive engineer started 20 working days earlier than the internal route. The value of that difference would be $24,000 in the model above. That is a much more useful number than counting how many CVs the agency sent in the first week.

5. What if the company already employs a recruiter?

This is where I would change the calculation entirely. Imagine the company has an HR team and is considering an agency only for one difficult engineering vacancy.

For an illustrative labour-cost calculation, the recruiter’s $75,940 BLS median and the same employer-benefit ratio work out to roughly $52 per paid hour using a 2,080-hour year. The BLS median for US computer and information systems managers was $175,140 in May 2025, which works out to approximately $120 in loaded compensation per paid hour using the same method. These figures indicate the economic value of people’s time, although they do not automatically represent additional cash that the employer would save. [4][5][9]

Let’s assume the internal team needs 55 hours of recruiter time, 18 hours of hiring-manager time and $750 of paid advertising to complete the search. An agency-supported process requires eight hours of internal coordination, 12 hiring-manager hours, and the 20% agency fee. The hours and advertising budget here are illustrative workload assumptions, and both routes still require technical assessment and a sensible hiring decision.

Cost of one search, internal versus agency-assisted
Cost of one searchInternal searchAgency-assisted search
Internal recruiter or HR work55 × ~$52 = $2,8698 × ~$52 = $417
Engineering hiring-manager work18 × ~$120 = $2,16512 × ~$120 = $1,444
Paid advertising$750$0 assumed
Agency placement fee$0$27,196
Total allocated recruiting resources$5,784$29,057

Allocated cost of one search when the recruiting team is already on payroll. Hours and advertising budget are illustrative workload assumptions.

The agency route costs approximately $23,273 more on the search itself. An agency shouldn’t describe that number as a saving. The employer will generally keep paying its existing HR employees regardless of whether their time is used on this vacancy, so much of the difference in staff hours represents freed capacity rather than cash recovered.

The decision becomes more interesting if we bring in the vacancy-cost calculation. Assume the agency’s process helps the engineer reach a productive start 20 working days earlier. In the previous example, the company loses $1,200 net per extra vacant day:

20 days × $1,200 = $24,000 of avoided vacancy impact.

$24,000 benefit − $23,273 extra recruiting cost = approximately $727 in favour of the agency.

Figure 2 · One critical vacancy, existing recruiter on payroll

When the faster hire changes the result

An agency-assisted search compared with the internal route, assuming the agency helps the engineer reach a productive start 20 working days earlier. Illustrative scenario.

Avoided vacancy impact

$24,000

20 productive days × $1,200 net daily vacancy impact

Extra agency recruiting cost

$23,273

$29,057 agency-assisted search − $5,784 internal search

Net, in favour of the agency

+$727

A margin of about 3% of the benefit. Small enough that one assumption can reverse it.

Change one assumption and the answer moves

20 days earlier · $1,200 a day

+$727

$24,000 benefit. The agency is slightly cheaper.

10 days earlier · $1,200 a day

−$11,273

$12,000 benefit. The internal route is cheaper.

20 days earlier · $300 a day

−$17,273

$6,000 benefit. The internal route is comfortably cheaper.

Figure 2. Illustrative agency-speed scenario. The result depends on the company’s real net vacancy cost and an actual or credible improvement in time to productive start; the model does not assume agencies are always faster.

In this example, the agency is the financially better option, although only by a small margin. The difference is useful precisely because it shows how quickly the answer can change when the business impact or time-to-start changes.

At that level of business impact, the agency would need to bring forward the productive start by approximately 20 working days to justify its incremental cost through speed alone. If it improves the start date by only ten working days, the business receives $12,000 of value and still spends around $11,273 more than the internal route. If the vacancy costs only $300 per day, even a 20-day improvement is worth just $6,000, leaving the internal option comfortably cheaper.

This analysis doesn’t assume an agency always recruits faster. Search speed varies by employer, role, agency and market, and a shortlist date isn’t the same as an employee’s productive start date. The speed advantage needs to come from comparable hiring data or a defensible forecast.

6. The cost of a bad hire is usually spread across several budgets

A poor hiring decision can create expenses long after the initial recruitment cost has been approved. The hiring manager may devote weeks to extra supervision, engineers may need to redo work, a contractor may be brought in to help, and the company may face a second recruitment process followed by another vacancy.

There are plenty of dramatic claims online about every bad hire costing two or three times annual salary. I wouldn’t use a universal multiplier to make this argument. A failed hire who leaves after two weeks and a senior engineering leader who creates months of rework have very different financial consequences. Companies can calculate their own losses using a simple incident-based ledger.

Consider this hypothetical case of an engineer who leaves after a period of underperformance. We will count only additional costs attributable to the unsuccessful hire, separating the period before departure, the vacant period after departure and the new hire’s ramp-up. Ordinary salary already budgeted, the original recruitment spend and benefits already included elsewhere are not added twice.

Additional cost of a failed hire
Extra consequence of a failed hireIllustrationCost
Extra manager supervision beyond normal onboarding40 hours × ~$120$4,812
Rework by other engineers on separate deliverables50 hours × ~$93 loaded hourly cost$4,670
External contractor brought in to remedy specific workIllustrative invoice$6,000
Additional replacement-search advertising and assessment expensesIllustrative spending$1,000
Vacancy after departure, using the demonstrated $1,200 daily net impact15 working days × $1,200$18,000
Reduced contribution during replacement’s initial ramp-up, after their start10 days × $400$4,000
Total additional impact in this scenario$38,482

Hypothetical failed-hire ledger that counts only additional costs attributable to the unsuccessful hire.

The $93 loaded engineering hour uses the same software-developer median and BLS aggregate benefits method as before. As with management time, diverting a salaried colleague’s hours represents an opportunity cost unless the company also incurred overtime or replacement labour. The contractor’s contribution must be accounted for before valuing any remaining lost work.

And the $38,482 doesn’t necessarily include a second agency invoice. If an agency made the initial placement, its replacement guarantee could provide another search without a new placement fee or a partial refund, depending on the agreement. If no guarantee applies and the company needs to pay a second 20% fee on the same salary, that would add $27,196, taking this particular scenario to roughly $65,678 in incremental impact. The original hiring fee remains a cost of the original hiring decision and should be included only once in an end-to-end financial analysis.

Replacement guarantees can be valuable, but they rarely reimburse lost productivity, damaged deliverables or the manager’s time. I would always check their duration, conditions, exclusions, refund provisions and whether they cover one replacement or something broader.

7. Can recruitment quality be included in the ROI calculation?

Yes, provided that the company has comparable evidence. I would not claim that an agency automatically reduces bad hires. An excellent in-house recruiter can outperform a weak agency, and an agency may have an advantage on a specialist search where the company’s existing recruiters have little market experience.

For instance, suppose a business has enough hiring history to establish that a comparable group of internally recruited employees has a 12% unsuccessful-hire rate, while a genuinely comparable group of agency-supported hires has a 7% rate. These rates are hypothetical, not Meettal outcomes or published statistics.

If the company has established an average incremental failed-hire impact of $38,482, the expected difference would be:

(12% − 7%) × $38,482 = approximately $1,924 per hire.

That expected improvement could be added to the agency’s financial value, but only after accounting for differences in role difficulty, candidate seniority, hiring volume and sample size. If the employer doesn’t have reliable comparative data, the prudent approach is to leave the estimated quality benefit at zero until evidence becomes available.

8. The calculation I would actually ask a founder or CFO to make

For the decision to create an in-house recruiting position, start with expected annual hiring volume, the fully loaded cost of the additional recruiter, their tools and any role-specific expenses. Compare that with the agency fee multiplied by the likely number of successful placements, adding the internal tasks that remain with an external agency. Don’t assign the entire annual cost of an existing HR department to only two vacancies, and don’t assume every agency fee would have been avoided if the company couldn’t complete those hires itself.

For a decision about one vacancy when an internal recruiting team already exists, I would calculate a different number:

Additional agency cost = full agency-assisted search cost − incremental internal-search cost.

Then compare it with the advantages that can reasonably be measured:

Estimated agency value = avoided net vacancy impact + verifiable freed capacity + evidence-based change in failed-hire risk.

If the expected advantages exceed the extra agency cost, engaging the agency may be the financially better choice. If the agency has no meaningful effect on speed, capacity or quality, the fee may simply add expense. Most importantly, finance teams should distinguish actual cash savings from the economic value of existing employees’ time, and avoid counting the same vacancy or productivity loss twice.

I also wouldn’t evaluate a recruitment partner solely on whether it promises to send candidates quickly. I would ask about comparable placements, access to the relevant specialist market, assessment quality, expected time to a productive start, the internal time required, and what happens if the hire leaves. Those answers give a much stronger basis for estimating financial value than a presentation promising access to a large database.

Calculator · Use your own numbers

Is an agency cheaper for your hiring plan?

The two comparisons from this article, with every input editable. The defaults reproduce the US examples above. Replace them with your own salaries, fees and vacancy costs. The arithmetic works the same in any currency.

Your hiring plan
A new in-house recruiter

In this plan

–

  • Fee per placement–
  • Recruiter’s employer benefits–
  • Break-even volume–

Counts dedicated recruitment capacity only. Interviews, assessments and onboarding happen on both routes and are left out. It assumes the new recruiter can actually complete the hires.

An illustrative model, not a quote or a forecast. It does not include the quality of hire, replacement guarantees or the cost of a failed hire, which are discussed separately above. Neither route is guaranteed to be cheaper or faster.

When the cheaper-looking choice costs more

When I look at the cost of hiring, I want to see the recruitment fee next to the cost of maintaining an internal team, the financial effect of leaving the role vacant and the possibility that a hiring decision might need to be repeated. The useful comparison is the money and capacity the company would genuinely spend, save or recover under each option.

A business with eight predictable openings a year might have every reason to invest in its own recruiting team. A business hiring two specialist engineers may find that paying for two successful agency placements costs considerably less than maintaining a new dedicated recruiter for a year. And a business with an existing recruiter may still find value in outside help for one critical vacancy, provided that the agency can make a meaningful, measurable difference.

At Meettal, this is the conversation I’d like more companies to have before deciding how to recruit. If they have a strong internal team with enough capacity, the numbers may favour keeping the search in-house. If a critical specialist role is sitting vacant, or the company would have to build recruitment capacity for just a few hires, the calculation can turn out very differently.

Sometimes a recruitment agency is expensive. Sometimes not having one costs more. I’d rather show a company how to calculate that difference than ask them to take it on trust.


Sources and calculation notes

  1. Upwork, How Much Do Staffing Agencies Charge for Fees in 2026?. Published in 2026; reports approximate 18%-25% direct-hire fees. General pricing guidance rather than a statutory or universal rate.
  2. SHRM, How to Work Effectively with an Executive Search Firm. Discusses approximately 20%-25% contingency fees and higher retained executive-search fees; older qualitative evidence, used for context.
  3. US Bureau of Labor Statistics, Software Developers, Quality Assurance Analysts, and Testers. Median software-developer salary $135,980, May 2025, published/updated in 2026.
  4. US Bureau of Labor Statistics, Human Resources Specialists. Median HR-specialist salary $75,940, May 2025; includes recruitment specialists.
  5. US Bureau of Labor Statistics, Employer Costs for Employee Compensation, June 2026. Private-industry wages/salaries were 70% of total compensation and benefits were 30%. This is a population-average ratio, not an employer-specific benefit rate.
  6. LinkedIn, Recruiter Lite and Recruiter comparison. Plan comparison. When checked in November 2026, this page no longer displayed a price; the $170/month US single-seat Recruiter Lite figure is as reported by independent pricing trackers, for example 100Hires, LinkedIn Recruiter Pricing 2026. Regional and plan terms can vary. Full Recruiter is quote-based.
  7. Workable, Pricing. Standard advertised at $299/month for its 1-20 employee tier when researched in November 2026; price and eligibility depend on company size and plan selection.
  8. LinkedIn, Recruiter pricing. Confirms professional Recruiter pricing depends on plan, hiring volume, geography and contract.
  9. US Bureau of Labor Statistics, Computer and Information Systems Managers. Median wage $175,140, May 2025.
  10. SHRM, 2026 Recruiting Executives Benchmarking. Median time-to-fill of 39 calendar days for nonexecutive positions; this is background context, not a prediction for senior engineering searches or for agency-versus-internal performance.

Methodology: Values in the fee table and salary-based annual capacity calculation derive from published benchmarks and vendor prices. All illustrative recruiting hours, job-ad budget, contractor invoices, vacancy impact, failed-hire scenarios and relative hiring outcomes are assumptions, not market averages. BLS wage and benefits measures come from different data series, so applying the overall benefits ratio to individual wage medians is an approximation. Total cash expense and allocated employee capacity are not identical. In a business-specific calculation, avoid double-counting overlapping manager time, payroll savings, rework, contractor output, lost contribution or consecutive vacancy periods. Neither route is guaranteed to be cheaper or faster.

MEET(TAL) . TALENT PARTNER

Hiring for a real role and want it done properly?

Meettal recruits senior engineers, HR leaders and CTOs across Armenia, the UAE, the US, Europe, Singapore, Australia and Taiwan. Twenty minutes with our team and you will know whether we can help. If we cannot, we will tell you who can.

Book a 20-minute call →