UK Recruitment Agencies

Recruitment is the only sector in this series in which booked revenue can be taken back. A permanent placement fee is invoiced on start and stays reversible for eight to twelve weeks. Revenue is not earned when it is billed; it is earned when the candidate stays.

What the sector publishes

• UK recruitment industry value, APSCo£43bnObserved
• Placements indicated for 2025, from a 25.7m peak in 202221.8mObserved
• UK recruitment agencies counted in 202230,035Observed · dated
• Contract margin as a share of charge rate12–18%Observed
• Net fee income per total head at one listed pure-play£103kObserved
• Modelled retention from invoiced permanent fee to cash82%Modelled
• Net fee income per consultant, SME agencies—Not established

The choice of top line is the most consequential modelling decision in this model. A contract agency placing someone at £500 a day on a 15% margin books £500 of revenue and keeps £75; a permanent agency placing the same person on a £90,000 salary at 20% books £18,000 and keeps all of it. Averaging the two produces a number that describes neither, which is why published margins of 5 to 15% of total revenue cannot be read across agency types. Every corridor here is stated against net fee income, as the listed pure-plays state theirs.

The umbrella exclusion is the one most often blurred: an umbrella company processes contractor payroll for a fixed weekly fee and carries no placement risk, no rebate exposure, no sourcing cost and no consultant desk, so none of the economics transfer. The hybrid archetype is the other thing worth stating plainly. Permanent and contract books fail in opposite directions, and run together the contract book funds payroll while the permanent book funds growth — modelled rebate exposure of 2 to 5% of net fee income against 4 to 8% for the permanent agency.

Recruitment demand is cyclical, unusually well measured by trade bodies, and currently coming off a hard contraction. The sector is one of the few in this series where the market genuinely is the constraint, at least in the near term.

Four separately managed international businesses reporting net fee income down 14%, 13%, 12.8% and 9% is a market signal, not four execution failures — but the contraction is concentrated in permanent hiring, which is the first line a client cuts, while contract demand frequently strengthens, because clients who will not commit to headcount still need the work done. The other distinguishing figure is client switching cost, which is close to zero. Briefs go to several agencies at once, so the contract book is the sector’s only durable relationship asset.

Revenue begins with a consultant desk and improves in quality as more of it recurs without being re-won — through contract margin that accrues weekly, retained work paid in stages, and extensions that require no new placement at all.

Fee percentages and rebate terms are market conventions published by agencies, not surveyed data. Net fee income per consultant for SME agencies is modelled: no body publishes it, and listed pure-play figures describe a different kind of business.

A permanent fee is invoiced on the candidate’s start date, commission is usually calculated on billing and often paid before the window closes, and the whole transaction stays reversible for another two to three months. Thirteen of the eighteen points lost land after the invoice: seven to candidates who accept and never start, six to rebate claims within the window. Both respond to the same intervention — structured contact through the offer-to-start gap and the rebate window, owned and scheduled. The extension rung is the cheapest revenue in the sector and the most commonly neglected.

All figures on this page are stated against net fee income, not total revenue. For a contract agency, total revenue includes the worker’s pay passed straight through, so margins expressed against it are not comparable between agency types.

Base central model: 10 consultant desks · £128k modelled net fee income per desk · 82% retention from invoiced fee to collected cash

The contract specialist earns the lowest net fee income per placement and carries the highest modelled conversion to normalised EBITDA, 23% against the generalist permanent agency’s 16%. Three causes compound: the margin recurs, so a desk’s output is cumulative rather than reset monthly; reversal is almost nil, at a modelled 0 to 2% of net fee income against 4 to 8%; and consultant compensation is a lower share, because commission rewards the placement event while the contract desk earns most of its income after the event has passed.

Consultant desks bind before candidate supply or client demand in every released archetype, and a desk is not capacity on the day it is hired. Ramp time, attrition and the proportion of the desk spent on briefs that will never fill determine whether purchased capacity produces net fee income.

Recruitment is the most activity-managed sector in this series, and the modelled economics do not support the assumption underneath it. A brief that will never fill — salary below market, unrealistic specification, four agencies and an internal process running against it — consumes exactly the same desk hours as one that fills, and produces nothing. Fill rate by brief, not placements per consultant, is what separates a productive desk from a busy one. It also explains the sector’s downturn response: qualifying briefs harder reduces measured activity, and looks like the wrong direction on every report.

Recruitment does not usually have a candidate problem. It has a conversion and retention problem, and the largest single loss occurs after the fee has been invoiced and the commission has in most cases already been paid.

Ask a prospect for these twelve figures. Most agencies can produce net fee income, placements and activity counts. Fill rate by brief, acceptance-to-start rate and rebate claims as a share of fees are the ones that are missing, and they are the three that bracket every loss on this page.

Seven points go to candidates who accept and never start and six to rebate claims, and both occur in a gap the agency has stopped managing. The offer-to-start window and the rebate window are the two periods when the placement is done, the fee is booked and everyone has moved on to the next brief — and the two when the candidate is most likely to be counter-offered or to arrive badly prepared and leave quickly. The intervention is unglamorous and cheap: scheduled contact with candidate and client, owned, recorded. It requires only that the gap stops being nobody’s job.

Recruitment is easy to enter and hard to make transferable. Around 30,000 UK agencies were counted in 2022, the great majority small and founder-billing. The ladder below is what separates an agency that can be sold from a job that pays commission.

If a consultant can leave and take their clients and candidates with them, the agency’s revenue is a collection of individual books that happen to share an office and a brand — which is why rung five is the only rung that creates transferable value. The listed downturn response does not transfer either. Cutting headcount by the percentage net fee income fell is rational for a listed business managing to a margin, but the desks a small agency cuts are usually the newest, and a new desk is not an unproductive desk. It is an unfinished one.

Modelled effect on normalised EBITDA of a one-standard-step move in each driver, hybrid archetype, widest first. All figures are stated against net fee income rather than total revenue.

Attrition appears fifth, which understates it. Every departing consultant is replaced by a ramping desk that costs full compensation and delivers partial contribution for six to nine modelled months, so a high-attrition agency never clears its ramping load — as one desk matures, another is replaced. Average net fee income per consultant, the widest single driver on the page, is therefore structurally depressed independently of how good the consultants are: the agency measures a productivity problem and has a retention problem. The permanent desk can wait for the recovery; the contract desk should not.

Technology creates value in recruitment only when it moves one of the drivers on the previous page. This is not a software catalogue, and the sector is unusually well supplied with tools that increase activity without increasing placements. Each intervention below names the driver it moves and the test that decides whether it worked.

The sector is the most heavily tooled in this series, and the binding constraint is not candidate supply; it is desk hours spent on briefs that will fill. A tool that doubles candidate throughput against an unfillable brief has doubled the cost of not filling it. Brief qualification is therefore first in the sequence and the intervention agencies resist most, because fewer briefs worked looks like a worse month on every report they currently produce. None of the six selects or scores candidates: automated assessment carries a discrimination risk no efficiency gain justifies.

The sector model becomes commercially useful at the moment a named agency is scored against it. The eight dimensions below convert everything in this model into a diagnostic that can be completed in a single working session with the agency’s own data.

The eight dimensions are scored against the agency’s own figures rather than against an interview, and in recruitment that distinction is unusually sharp: agencies are data-rich on activity and data-poor on outcome, and the two are easily confused in conversation. An agency that can produce calls made and submissions sent but not fill rate by brief has told you precisely where the problem is. Sequencing follows the sensitivity ordering with one exception — where relationship ownership scores at level one it is raised early, because every other improvement made to a desk is lost when the consultant leaves.