When EBITDA Lies: Why People Businesses Break the Metric

Any recruitment business valuation that starts from EBITDA starts from the wrong number. Revenue includes other people’s wages, EBITDA excludes the rent, and neither sees the fee-earner payroll that walks out of the door every evening. This paper shows how to value a recruitment agency – and a consultancy – on the gauges that survive: net fee income, the conversion ratio, cash, and the lease you still owe.

The professional services, recruitment and consulting paper in the Bloor Finance series: a CFO-perspective reading of Robert Walters plc (LSE: RWA) against the UK-listed recruitment peer set – Hays, PageGroup and SThree – with Accenture as the consulting contrast, on each company’s latest audited year and latest published half-year or trading update to 1 October 2026.

Prepared by Darren Sack · Research Director · Bloor Research International · October 2026

The opening paper in this series made one claim. Wherever an organisation embeds AI, cost leaves the wage bill – where EBITDA (earnings before interest, tax, depreciation and amortisation) sees it – and reappears as depreciation, interest, leases and provisions, where it does not. The technology paper followed that cost into Oracle’s data centres. The banking paper found it in Lloyds’ severance line. Professional services is the sector where the wage bill is the business – and where the opening paper put it most bluntly.

Its section on the sector named the exhibit and the signal. Robert Walters: ‘NFI down 15% and the group into an operating loss, with the IFRS 16 estate quietly holding EBITDA up as the business broke beneath it.’ Accenture: a business-optimisation programme and roles removed where reskilling was judged not to be a viable path. Both claims are re-tested below on the primary documents. Both survive. One gets larger.

Since the opening paper went to press, the rest of the sector has reported. Hays’ full year to June, published on 20 August, showed net fees down 7%, a £41.0m statutory operating loss after £89.6m of exceptional charges, and 74 offices closed or consolidated. PageGroup’s half-year showed a conversion ratio of 2.5%. SThree raised its profit guidance in September on one-offs it says will not recur. Accenture’s full year, on 1 October, announced no new optimisation programme. And the market itself turned, faintly: the KPMG and REC Report on Jobs recorded permanent placements rising slightly in August for the first time since late 2022, with UK vacancies still 10.9% below their pre-pandemic level.

EBITDA misleads in a people business because it takes out the one fixed cost that behaves like capital – the office lease – and leaves in the one that behaves like debt: the fee-earner payroll. It flatters most precisely as the business shrinks.

In a people-business, revenue is a vanity number; net fee income is the honest top line and the conversion ratio – operating profit ÷ NFI – is the honest margin.

A recruiter or consultancy owns almost nothing that depreciates. Its assets are people, client relationships and a lease. Under IFRS 16, the rent on that lease no longer runs through operating costs; it is split into depreciation of a right-of-use asset and interest on a lease liability – both of which EBITDA adds back. The payroll, meanwhile, stays above the line, fixed in the short run, and paid whether the placements come or not. The result is a metric that improves when the business does nothing differently except sign a lease, and that keeps printing positive numbers after the operating line has gone negative. The honest gauges are older and plainer: net fee income, the conversion ratio, debtor days, cash. The rest of this paper works through where EBITDA lies in the sector and what each lie looks like in a filing. 

A recruiter that supplies a temporary worker usually books the worker’s whole pay as its own revenue. The fee it actually earns is a fraction of that. That is a principal-versus-agent judgement, and it decides whether a recruiter’s turnover is seven times its net fee income or twice.

IFRS 15 settles the question by control. An entity is a principal if it ‘controls the specified good or service before that good or service is transferred to a customer’, and an agent if it does not; the indicators include primary responsibility for fulfilment, inventory risk and discretion over price. A recruitment firm that employs or contracts the worker and sets the charge-out rate will usually conclude it is principal and report gross. The UK’s Financial Reporting Council has said plainly that accounting policies ‘should clarify whether revenue is presented gross or net’.

The spread across the UK-listed recruiters shows what the judgement does to the headline. Hays reported turnover of £6,421.2m for the year to June 2026 and net fees of £905.5m – 14% of turnover. Robert Walters’ 2025 revenue of £781.1m produced net fee income of £274.2m, 35%. SThree’s £1,302.2m produced £322.7m, 25%; PageGroup’s £1,596.6m produced £769.5m of gross profit, 48%. Those ratios describe business models – how much temporary and contract work each firm runs through its own books – not the quality of the business. Compare revenue multiples across them and you are comparing payroll pass-through.

Robert Walters’ own first half shows the vanity at work in a single set of numbers. Revenue rose 8% to £433.7m. Net fee income fell 4% to £134.6m. The firm’s revenue line includes, in its own words, ‘the remuneration costs of temporary candidates’. Hays defines net fees as ‘turnover less remuneration of temporary workers and other recruitment agencies’. That definition is the honest top line. Net fee income is the number to value, to benchmark and to pay on.

Fees that have been earned but not yet invoiced sit on the balance sheet as accrued income – what IFRS 15 calls a contract asset, a right to consideration that is still conditional on something other than the passage of time. EBITDA counts it as earned the day it is recognised. Cash counts it when the client pays.

In recruitment the item is usually modest and seasonal, which is exactly why it repays attention. Robert Walters carried accrued income of £39.3m at the end of 2025, down from £45.6m – about 14% of the year’s net fee income. PageGroup’s accrued income was £98.1m at 30 June 2026 against £90.3m a year earlier, up 9% in a half in which gross profit fell 1% and net trade receivables fell from £205.9m to £201.6m. That may be timing; it may be mix. It is the kind of movement a board should be able to explain in one sentence.

Debtor days complete the picture. Hays reports 36 days (37 a year earlier); SThree, with a contract-heavy book, 53 days (55). In consulting the same lie is larger and more formal: Accenture’s days services outstanding rose to 50 at its 2026 year-end from 47 a year earlier, and its non-current contract assets rose from $120.3m in 2024 to $327.2m in 2026. Under IFRS 9 contract assets carry expected-credit-loss provisions; a buyer will test them anyway.

The cash line is where Robert Walters’ story becomes unambiguous. Cash generated from operations was £14.4m in 2025 against £26.2m in 2024; free cash flow was negative £14.6m, and negative £9.0m again in the first half of 2026. Net cash fell from £52.5m at the end of 2024 to £17.2m eighteen months later. No measure of earnings before anything tells you that.

Robert Walters reported an operating loss of £14.9m in 2025. Add back the £22.5m of depreciation and amortisation in its cash-flow statement and EBITDA was about £7.6m – positive. Of that add-back, £14.3m was depreciation of right-of-use assets: offices and cars.

Robert Walters does not publish an EBITDA figure; the arithmetic is Bloor’s, on the company’s own numbers. Take the lease back out – £14.3m of right-of-use depreciation and £3.2m of lease interest, roughly what the rent would have been before IFRS 16 – and the lease-adjusted figure is a loss of about £9.9m. A year earlier the same arithmetic gave EBITDA of £28.2m and a lease-adjusted £10.4m. EBITDA fell 73% and stayed positive. The lease-adjusted measure went from a £10.4m profit to a £9.9m loss. One of those two readings describes a business that broke in 2025. The other does not.

The estate was not idle. The company paid £17.6m of lease principal in 2025 and still carried £68.0m of lease liabilities at the year-end, against net cash of £26.2m – lease-adjusted net debt of about £41.8m. At 30 June 2026 it was £59.1m of leases against £17.2m of net cash: still about £41.9m. The leases have shrunk. The cash has shrunk faster.

The peers show the same mechanism from other angles. SThree is unusual in publishing its own EBITDA: £13.9m for the half to May 2026, against operating profit of £3.4m – four times as large. Hays’ £89.6m of exceptional charges for the year to June includes £14.0m of accelerated right-of-use depreciation on offices it is leaving. Strip exceptionals and add back depreciation, and the cost of exiting 74 offices disappears twice. And the conversion ratio, which sees all of it, tells the long story the opening paper asked for: at Robert Walters, 13.6% in 2022, 6.8% in 2023, 1.6% in 2024 and minus 5.4% in 2025.

In a recruiter the fee earner is the production line, and the payroll is fixed long before the fees arrive. In a consultancy the equivalent gauge is the utilisation rate – billed hours as a share of hours worked – and the industry’s best-known firm has stopped publishing it.

Robert Walters has cut hard. Total headcount fell from 4,356 at the end of 2022 to 2,782 at 30 June 2026, down 36%; fee earners from 1,964 at the end of 2024 to 1,648. Net fee income fell from £428.2m in 2022 to £274.2m in 2025 – also down 36%. Headcount has followed fees down almost exactly, and the conversion ratio collapsed anyway: the monthly cost base was cut from £25.5m at the exit of 2024 to below £24m at the exit of 2025, with about £22m targeted for the exit of 2026, and £4.4m of redundancy cost was booked in 2025 alone. The fixed cost is not the fee earner. It is everything the fee earner sits inside.

Hays’ consultant headcount was 5,194 at 30 June 2026, down 12% in the year and 41% below its 2023 peak; the firm closed 52 locations and consolidated 22. PageGroup’s gross profit per fee earner was £77.4k for the half, the highest since 2022. The peers all speak the language of AI as an aid to the consultant – ‘human skills become more, not less, important in an AI-enabled future of professional work’, in Robert Walters’ half-year statement – and SThree goes furthest: ‘the winners will be AI enabled businesses, including agentic AI.’

Accenture is the contrast, and the signal. In September 2025 it began a six-month business-optimisation programme that its CFO guided at about $865m; the amount actually recognised was $923m, of which $628m was severance and $295m impairments, mainly on two divested acquisitions. The chief executive described exiting, ‘on a compressed timeline, people where reskilling, based on our experience, is not a viable path for the skills we need’. Headcount at the end of the programme was about 814,000 – up from about 779,000 a year earlier.

…this is labour debt being written off, not a market being served.

Accenture reports its operating margin both ways: 15.4% under US GAAP for the year to August 2026, 15.8% excluding the $307.5m of optimisation cost recognised in the year. That 0.4-point gap is the consulting version of the EBITDA reflex – severance treated as outside the business, in a business whose only product is people. Utilisation, the gauge that would show whether the rotated workforce is billing, is not disclosed and is shown here as n/d.

One more payroll lie sits outside the accounts altogether. Since April 2021 medium and large UK clients, not contractors, decide whether off-payroll (IR35) rules apply, and the fee-payer – the party paying the worker’s company, often the agency – accounts for the tax; from 6 April 2026 agencies also became responsible for PAYE where workers are supplied through umbrella companies. SThree runs 84% of its net fees through contract work. No recruiter in the peer set discloses a provision for employment-status or umbrella-PAYE exposure; the cell is n/d, not zero.

On the honest gauges the four UK recruiters sit in a narrow, uncomfortable band: conversion ratios between minus 5.4% and 8.1%, net fee income falling at all four, and lease liabilities larger than net cash at three of them. The table sets the gauges side by side on each firm’s latest audited or reported full year.

Honest gauges – Robert Walters against the UK-listed recruitment peer set (latest full year; £m unless stated)

Company (year-end)	Revenue	NFI / gross profit (% of revenue)	NFI change	Operating profit: statutory (underlying)	Conversion ratio	Lease liabilities vs net cash	RoU depreciation	Fee-earner / consultant change
Robert Walters (Dec 2025)	781.1	274.2 (35%)	−15%	(14.9) (no underlying measure)	(5.4%); 2024 1.6%	68.0 vs 26.2	14.3	1,964 → 1,679 (−15%)
Hays (Jun 2026)	6,421.2	905.5 (14%)	−7%	(41.0) (48.6 pre-exceptional)	5.4% pre-exceptional; 2025 4.7%	153.2 vs 20.1	39.5	6,070 → 5,194 (−12% as reported)
PageGroup (Dec 2025)	1,596.6	769.5 (48%)	−9%	20.9 (n/d)	2.7%; 2024 6.2%	132.3 vs 31.4	n/d	5,370 → 4,968 (−7%)
SThree (Nov 2025)	1,302.2	322.7 (25%)	−13%	26.1 (n/d)	8.1%; 2024 17.9%	47.5 vs 68.0	13.1	Average −10%; period-end n/d

Sources: each company’s latest full-year results announcement (Robert Walters 11 March 2026; Hays 20 August 2026; PageGroup FY2025 results; SThree 27 January 2026). PageGroup reports gross profit, the others net fee income or net fees; definitions differ in detail. Conversion ratio is operating profit ÷ NFI as each company reports it (Hays on profit before exceptional items). Hays’ consultant change is as the company reports it. Debtor days are 36 at Hays and 53 at SThree; Robert Walters and PageGroup do not disclose a figure (n/d). Lease liabilities and net cash at the year-end; net cash excludes lease liabilities. n/d – not disclosed in the cited document; never treated as zero. No company is ranked; the comparison is indicative and the year-ends differ.

Three readings, none of them a ranking. First, the conversion ratio fell in the latest year at three of the four; at the fourth, Hays, it rose on the pre-exceptional measure while the statutory line went to a £41.0m loss – the gap between the two is the £89.6m of exceptional charges. Second, three of the four carry more lease liability than net cash; only SThree, with the smallest estate relative to its fees, does not. Third, the one firm in the set that publishes fee-earner productivity, PageGroup, reports it at its highest since 2022 – and a conversion ratio of 2.5% for the half. Productivity per head is necessary. It is not sufficient while the estate stays fixed.

A buyer will start from your EBITDA and then take it apart: normalise the founder’s pay, discount the key person, put the lease back, and pay part of the price later, contingent on fees that have not yet been earned. Knowing the sequence is the whole of exit readiness.

The published UK guidance on recruitment business valuation is consistent on the method, if not on the number. Multiples of 4x to 8x EBITDA are cited for well-run agencies, attributed to BDO; one corporate-finance boutique quotes net fee income multiples of 1.5x to 2x; another argues that recurring contract income ‘commands a higher valuation multiple’ than permanent fees. These are directional ranges from advisers, not market evidence, and they are quoted here as such. What the guidance agrees on is the normalisation.

Founder pay comes first. Advisers list ‘market-rate salaries for business owners’ among the basic adjustments: a founder who pays themselves below market has inflated EBITDA by the difference, and a buyer who must hire a replacement managing director will take it back out. At the 4x to 8x range quoted, every £1 of under-paid founder salary is between £4 and £8 off the price. The same guidance warns against the owner being the ‘top biller’ – key-person dependency that no multiple prices cleanly – and flags concentration where the top three clients exceed 40% of revenue.

Earn-outs are how buyers price what they cannot normalise. ‘Most trade sales include an earn-out period of one to three years’, on one adviser’s account; Deloitte describes the same one-to-three-year range and the purpose – ‘to bridge valuation gaps between Buyers and Sellers’. The share of consideration typically deferred is not published in the guidance reviewed and is shown as n/d. For a seller the implication is direct: an earn-out measured on EBITDA rewards the lease and the under-paid founder; one measured on net fee income or cash rewards the business.

Two further adjustments complete the buyer’s arithmetic. The lease goes back in – EBITDAR or lease-adjusted net debt, so that £68.0m of Robert Walters lease liability is read against £26.2m of net cash rather than ignored. And contractor status is diligenced: IR35 and umbrella-PAYE exposure is a liability EBITDA never sees. For consulting firm valuation the same logic holds with one substitution – utilisation and severance-inclusive margin replace net fee income and conversion.

The professional-services edition of the depreciation-aware dashboard is short, because the business is simple. Applied to Robert Walters’ 2025 accounts, it reads as follows.

– Net fee income, not revenue – and its split between permanent, temporary and outsourcing. Robert Walters: £274.2m, down 15%; permanent −14%, temporary −17%, recruitment outsourcing −15%.

– The conversion ratio over three years, not one. Robert Walters: 6.8%, 1.6%, (5.4%).

– Lease-adjusted EBITDA and lease-adjusted net debt. Robert Walters: EBITDA about £7.6m; lease-adjusted about £(9.9)m; lease liabilities £68.0m against net cash £26.2m.

– Cash conversion and free cash flow. Robert Walters: cash generated from operations £14.4m; free cash flow £(14.6)m.

– Accrued income and debtor days, together. Robert Walters: accrued income £39.3m (14% of net fee income); debtor days n/d.

– Fee-earner productivity and the labour-debt ratio – NFI per fee earner against the non-fee-earning cost base. Robert Walters: 1,679 fee earners and 1,209 non-fee earners at the year-end; NFI per average fee earner n/d.

– Restructuring as a share of the saving it buys. Robert Walters: £4.4m of redundancy cost; the saving is stated as a run-rate, not a year-on-year figure, so the ratio is n/d.

– Contractor-status and umbrella-PAYE exposure. Peer set: n/d.

For a professional-services CFO the work is specific. Report net fee income above revenue. Publish the conversion ratio over three years, and the lease-adjusted EBITDA alongside the EBITDA. Disclose debtor days and accrued income in the same sentence. Put a number – any number, with a method – against contractor-status exposure. And measure any earn-out you agree on fees or cash, never on EBITDA.

Robert Walters is not the weak case for this paper; it is the honest one. It reports no adjusted operating profit, publishes fee-earner headcount every quarter, and its accounts show the lie in its plainest form: an operating loss on the face of the income statement and a positive EBITDA one subtraction away. That is why it is the right exhibit. If a recruitment business valuation would have read that year as profitable, the valuation method – not the business – is what needs reform.

Watch for the remaining When EBITDA Lies industry papers as they publish. For boards and owners who want an early diagnostic read of their honest gauges before a sale process or a results season, the Bloor Finance Desk is available for a confidential conversation.

Analysis only; not investment advice.

Every figure in this paper is drawn from a named, dated document; the principal ones are listed below with links, and a full Source Note listing the origin, pinpoint and status of every figure accompanies this paper. EBITDA, lease-adjusted EBITDA, lease-adjusted net debt and the percentage ratios are Bloor arithmetic on reported figures; no company in the peer set other than SThree publishes an EBITDA figure. Valuation multiples and earn-out terms are directional adviser guidance, attributed. Items not disclosed in the cited documents are shown as n/d, never as zero. No company is ranked. This is analysis for a CFO audience and is not investment advice.

Robert Walters plc

• 2025 Full-Year Results (11 March 2026): https://www.robertwalters.com/content/dam/robert-walters-redesign/global/files/investors/2026/RWA-FY25-results-statement_FINAL%20v2.pdf
• 2026 Half-Year Results (30 July 2026): https://www.investegate.co.uk/announcement/rns/robert-walters–rwa/2026-half-year-results/9694660
• FY23 Results (7 March 2024): https://www.lse.co.uk/rns/RWA/fy23-results-xrq2ax8jgkt8gyk.html

Hays plc

• Preliminary results for the year ended 30 June 2026 (20 August 2026): https://www.rns-pdf.londonstockexchange.com/rns/4022R_1-2026-8-19.pdf

PageGroup plc

• Half-year results to 30 June 2026 (6 August 2026): https://www.investegate.co.uk/announcement/rns/pagegroup–page/half-year-results-for-the-period-ended-30-june-26/9707293

SThree plc

• FY26 Half-Year Results and Q3 Trading Update (21 & 22 July/September 2026): https://www.investegate.co.uk/announcement/rns/sthree–stem/fy26-q3-trading-update/9783226

Accenture plc

• Q4 and full-year fiscal 2026 results (Form 8-K, 1 October 2026): https://www.sec.gov/Archives/edgar/data/0001467373/000146737326000037/q4fy26earnings8-kexhibit.htm
• Form 10-Q for the quarter ended 30 November 2025: https://www.sec.gov/Archives/edgar/data/1467373/000146737325000222/acn-20251130.htm

Regulatory & Standards

• IFRS Foundation – IFRIC Update, April 2022: https://www.ifrs.org/news-and-events/updates/ifric/2022/ifric-update-april-2022/
• HMRC – April 2021 off-payroll working changes: https://www.gov.uk/guidance/april-2021-changes-to-off-payroll-working-for-clients
• HMRC – Umbrella company market changes (26 November 2025): https://www.gov.uk/government/publications/paye-changes-for-the-umbrella-company-market/umbrella-company-market-changes-to-income-tax-rules-to-tackle-non-compliance

Adviser & Market Guidance

• Deloitte UK – ‘Earn-out or burn-out’ (24 November 2020): https://www.deloitte.com/uk/en/services/financial-advisory/blogs/2020/earn-out-or-burn-out-a-pricing-mechanism-for-future-performance.html
• Recruitment Leadership – Exit-readiness guide (3 March 2026): https://recruitmentleadership.co.uk/how-to-prepare-your-recruitment-business-for-sale-the-exit-readiness-guide/
• UHY Hacker Young (14 May 2024): https://www.uhy-uk.com/insights/third-party-sales-securing-best-deal-your-recruitment-business
• PEM Corporate Finance: https://www.pemcf.com/valuing-a-recruitment-company/
• KPMG and REC – UK Report on Jobs (7 September 2026): https://kpmg.com/uk/en/media/press-releases/2026/09/kpmg-and-rec-uk-report-on-jobs-september-2026.html
• ONS – Vacancies and jobs in the UK (15 September 2026): https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemployeetypes/bulletins/jobsandvacanciesintheuk/september2026

Print Article for this blog Download here