How to Reduce Recruitment Agency Spend: The CFO's Cost Model

Most CFOs do not have an agency problem. They have a fee-structure problem. When the recruiting line item arrives each quarter, the question is rarely "are we hiring too many people?" but "why does this number scale exactly in line with hiring volume, with no efficiency curve and no sign that paying more this quarter reduces what we pay next?" That is what this article answers, in numbers.

If you are the head of talent, this is the article to share when you make the internal cost case: it hands you the fee math and model comparison in the language finance responds to, so you walk into the budget conversation with the numbers already on your side.

The UK recruitment sector contributed £40.6 billion to the economy in 2024, according to REC's industry data, down from £44.4 billion the year prior in a cooling labour market. Even in a softer year, recruiters remain an enormous category of corporate spend, and CFOs have begun to model it accordingly. If you have already concluded the agency model is the wrong shape for your hiring volume, the broader strategic frame sits in our case for moving away from agency-led hiring. What follows here is narrower: how agency fees behave at scale, why preferred supplier lists and renegotiation do not change that shape, and what the cost curve looks like when companies stop buying placements and start buying recruiting capacity.

Key Takeaways

• Agency fees are a tax on hiring volume, not a payment for difficulty. Cost scales with salary times number of hires, regardless of role complexity.
• Renegotiation gives you a smaller version of the same problem. A 5 to 10 percent discount on a structurally inefficient model is still a structurally inefficient model.
• Preferred supplier lists do not collapse the cost; they consolidate it. Volume rebates require deeper dependency, not less.
• Failed placements double the effective cost on a meaningful share of hires. Rebate periods often expire before the real attrition signal arrives.
• The only way to flatten the curve is to stop paying success fees per hire. Infrastructure-based recruiting decouples capacity from compensation.

The Agency Fee Math: Why Spend Scales Linearly With Hiring

Recruitment agency fees in Europe typically fall between 15 and 30 percent of the candidate's first-year base salary, with most permanent placements priced in an 18 to 25 percent band depending on role, sector, and country. UK roles cluster at the lower end, Germany spans the full range, and senior or hard-to-fill positions move toward 30 percent, per country-level benchmarks published by Hiring Hub.

The structural point is not the percentage. It is the base. Fees are anchored to compensation, not to the work the agency performs. Sourcing a €55,000 customer success manager and sourcing a €110,000 backend engineer can take a similar number of recruiter hours; the fee for the second is double. The model rewards salary inflation and penalises specification clarity. Tighter markets compound the pressure: Eurostat reports that 57.5 percent of EU enterprises had difficulty filling ICT specialist vacancies in 2023, rising to 72 percent in Germany.

Per-Hire Cost at Typical Fee Ranges

For a mid-level European role at the Eurostat 2024 salary benchmarks, where the EU average full-time adjusted salary reached €39,800 in 2024 (with DACH, Benelux, and Ireland clustering well above the average), the per-placement cost at €55,000 (an upper-quartile midpoint for professional and specialist roles) breaks down as:

How the Fee Compounds Across Annual Hiring Volume

Using a blended 22 percent fee at the €55,000 benchmark and at €80,000 for higher-end roles:

At 24 hires per year, agency spend at typical fees is roughly equivalent to three to four fully loaded internal recruiters. At 60 hires, it funds an entire embedded recruiting function with budget left over. This is the compounding pattern CFOs see on the P&L: a function whose marginal cost per hire never drops, even as the company gets more efficient at every other operating expense.

What the Headline Fee Hides

The placement fee is not the full bill. Two secondary costs sit underneath it.

First, the replacement multiplier. Most agency contracts include a rebate window of 8 to 12 weeks; outside that window, the full fee is owed again on the backfill. The CIPD Resourcing and Talent Planning Report, based on a UK sample of over 1,000 HR professionals, reports that 41 percent of organisations had new recruits who always, mostly, or sometimes resigned within the first 12 weeks.

The exposure is asymmetric: the agency is paid for placement; the company is exposed to the consequences of fit. If even 10 percent of annual hires need replacement past the rebate window, effective per-hire cost rises by ~10 percent without any change in headline fee. On a 24-hire plan at €290,400, that is another €29,000 in unbudgeted backfill fees per year.

Second, the internal cost that survives outsourcing. Hiring managers still write specs, run interviews, and make offers. Internal TA still vets agency candidates. For DACH-headquartered companies, the Kienbaum and DGFP HR-Kostenstudie puts the HR function cost at approximately €2,600 per employee per year across the entire workforce, with 72 percent going to internal personnel cost. (This is total HR function cost per head, not a cost-per-hire metric.) Read alongside agency spend, the architecture is visible: HR cost is largely fixed, agency spend is variable on top, and neither line decreases when the other rises.

Cost-Per-Role Across Models: The Comparison CFOs Need

The gap between the first row and the rest is the whole argument. Moving a mid-level role from a contingency placement (€12,000 to €14,000) to a marketplace or on-demand model (€4,000 to €8,000) cuts the per-role cost by roughly 40 to 70 percent on a like-for-like comparison. How much of that reaches the P&L depends on how much volume you move: shift the bulk of your repeatable roles and total external spend falls in the same band; keep strategic hires on retained search and the blended figure is lower.

Two patterns produce that gap. First, every alternative breaks the salary-anchored fee structure. Second, only the last three break the linear-with-hires scaling: capacity flexes up or down without renegotiating a contract or adding an FTE. This is where recruiter marketplaces enter the math as infrastructure rather than as a vendor.

Why Renegotiation and Preferred Supplier Lists Do Not Fix the Structure

The standard playbook is to negotiate lower fees, consolidate to a preferred supplier list (PSL), and add volume rebates. Each move looks like savings on paper. None changes the cost shape.

A successful renegotiation typically reduces fees by 3 to 5 points (for example, from 25 to 20 percent). On a €290,400 annual spend, that is €58,000 saved. Real money, but the curve is unchanged: next year's spend still scales linearly with hiring volume, at a shallower slope.

Bottom line: Negotiating fees lowers the slope. It does not flatten the line.

PSLs standardise fees across a small panel in exchange for guaranteed work volume. Unit economics are unchanged: each placement is still priced as a percentage of salary, paid on success, with the same rebate windows. The savings are largely administrative (fewer onboardings, cleaner reporting) rather than structural. PSLs that reward volume also tend to deprioritise specialised roles where panel partners are weaker, pushing hard-to-fill positions to off-PSL agencies at full rack rate.

Volume rebates produce a more subtle problem. A clause returning 2 percent at €500,000 spend and 5 percent at €1 million is net-positive only if hiring volume rises. Hitting the threshold becomes a budgeting target rather than a side effect.

The blunt framing: you can negotiate harder and save 5 to 10 percent on fees, or you can ask why you are paying success fees at all when the model incentivises speed over quality and rewards salary inflation. One is optimisation. The other is transformation.

Signs Your Agency Spend Has Crossed the Sustainability Line

Use this as a five-minute diagnostic with the head of TA.

• Agency fees represent 8 percent or more of total compensation for new hires across a full fiscal year.
• More than 20 percent of agency placements require backfill within the rebate window or within six months.
• Hiring managers can name three or more agencies by recruiter, suggesting the same shortlist arrives from multiple sources.
• Time-to-fill has not decreased despite year-over-year increases in agency spend.
• Internal recruiters spend more time vetting agency candidates than sourcing their own.
• Procurement has renegotiated fee rates twice in the past 24 months and the line item has continued to grow.

If three or more apply, the issue is not the agency partner. It is the operating model. The CFO question shifts from "how do we get a better rate?" to "what do we replace this with?"

How to Avoid Recruitment Agency Fees Entirely: Redesigning Capacity Access

The path to eliminating agency fee dependency is not to find cheaper agencies. It is to stop buying placements and start buying recruiting capacity. Three shifts make this possible.

Decouple Cost From Placement

In the agency model, you pay nothing until a hire is made, then pay a percentage of compensation. In an infrastructure model, you pay for recruiter time, project scope, or platform access, regardless of whether a specific role closes. Cost becomes a function of recruiting work performed, not of compensation paid to new hires. Budget becomes a planning input.

For roles where execution can be delegated cleanly, the cleanest path is on-demand recruiting, priced per role or per hour with no long-term contract.

Match Capacity to Role Type, Not to Hiring Volume

Not every role belongs in the same model:

• High-volume repeatable roles (SDRs, customer success, junior engineering): handled internally or through dedicated recruiters hired without agency contracts, at amortised per-hire costs of €3,500 to €5,500.
• Specialised but predictable roles (mid-level engineering, finance, product): routed through a vetted marketplace or fractional recruiter at €4,000 to €8,000 per role, with no rebate-window exposure.
• Genuinely scarce or strategic hires (senior leadership, regulated specialists): retained search remains defensible at 25 to 30 percent fees, on a small absolute number of roles per year.

Most companies overpay because they route every role through the most expensive channel, then negotiate the rate.

Replace the Fee With Subscription or Per-Role Pricing

A marketplace approach to agency replacement prices recruiting differently. Per-role flat fees, monthly subscription access to recruiter capacity, or Jobslot-style infrastructure pricing share one property: the marginal cost of one more hire does not scale with the salary of that hire. The line on the P&L starts to look like every other piece of operating infrastructure: fixed cost, measurable utilisation rate.

For multi-country hiring, the cost differential widens. Marketplace and on-demand models access local specialists without separate vendor contracts per country, eliminating both a procurement layer per geography and a fee layer per hire.

What a 90-Day Transition Looks Like

Moving the cost curve does not happen in a single quarter.

• Days 0 to 30: audit 12 months of agency invoices, map every hire to a fee, role type, and backfill outcome. This typically surfaces 15 to 25 percent of spend on roles internal capacity could have closed.
• Days 30 to 60: route two or three high-volume role categories away from agencies into a marketplace or on-demand engagement; hold fee-based relationships in place for strategic roles only.
• Days 60 to 90: expand the share routed through infrastructure-based models based on pilot data.

The cost shape changes in week four. The savings show on the P&L in quarter three. The structural change holds through every subsequent cycle.

The Question Behind the Line Item

Reducing recruitment agency spend is not a procurement project. It is a model question. Negotiating fees and consolidating panels lower the spend slightly while preserving the structure that produces it. Redesigning how recruiting capacity is accessed removes the structure entirely.

You do not fix agency dependency by negotiating better rates. You fix it by eliminating the dependency.

Frequently Asked Questions

How much can a company realistically reduce agency spend by redesigning the model rather than renegotiating?

It depends less on the rate you negotiate than on how many roles were sent to an agency that never needed one. Negotiation lowers the price of each placement but keeps a success fee on every hire, so the saving resets at the next renewal. Redesign changes what you are buying, so the saving holds and compounds. That structural difference, not a headline percentage, is what shows up on next year's P&L.

Is a recruiter marketplace just a cheaper agency?

No. The pricing difference is real, but the structural difference matters more. An agency owns the candidate relationship and gets paid on placement, which incentivises speed and salary inflation. A recruiter marketplace connects companies directly with independent specialised recruiters on a per-role or project basis, with no long-term commitment and no success-fee anchoring to compensation. The company retains process control and pays for recruiting work performed, not for placements made.

Is on-demand recruiting a cheaper alternative to RPO?

Often, yes, particularly for companies with variable hiring volume. Traditional RPO is built around fixed scope and multi-year contracts, which produces stable per-hire costs when volume matches the contract but expensive idle capacity when it does not. On-demand recruiting prices capacity by the hour or the role, which flexes with actual hiring need. For organisations that historically used RPO mostly to absorb agency spend, the marketplace and on-demand models typically come in lower at the same quality bar.

What about preferred supplier lists with rate cards? Do not large enterprises run these effectively?

PSLs work as a procurement governance tool, not as a cost-reduction strategy. They standardise contracts, reduce vendor sprawl, and produce cleaner reporting. They do not change the fee-per-placement structure. Large enterprises that have run PSLs for years still find that recruiting cost scales linearly with hiring volume, because the underlying unit pricing has not changed.

How should a CFO model agency spend in next year's budget?

Run two parallel forecasts. The first assumes status-quo agency routing and projects spend as (hiring plan × blended salary × blended fee percentage), then adds a replacement factor of 8 to 12 percent for rebate-window failures. The second assumes a phased reallocation toward infrastructure-based recruiting, with agency spend held only for strategic roles. The delta is the size of the prize. For companies hiring 24+ roles per year, it is typically a six-figure annual saving and a flatter cost curve in subsequent years.

Does redesigning the model work in multi-country hiring, or is it just for single-market companies?

It works particularly well in multi-country hiring. The agency model produces the highest costs across borders because it adds a per-geography procurement layer on top of a per-hire fee layer. Marketplace and on-demand models access local specialists without separate vendor contracts per country, which removes both layers simultaneously. Companies hiring across DACH, Benelux, and Iberia typically see the largest absolute savings in this configuration.