Recruitment agencies: how the model works, and where it stops working

Most buyers compare recruitment agencies on reach and on speed. Reach is easy to assume and hard to verify, so ask any agency where its last five placements were first found before you decide that reach is what you are buying. The thing that reliably predicts what arrives in your inbox is not reach. It is how the agency gets paid. This page explains both dominant fee models honestly, including the searches where an agency is the right purchase, and then what the model structurally cannot do.

What a recruitment agency actually sells

An agency sells introductions. It finds and forwards candidates for a role you have already decided to fill, and it is paid in one of three shapes. The shape is written into the contract, and it decides what happens on Monday morning.

Contingency

Paid only if one of its candidates joins, as a percentage of that person's first-year package. Nothing is owed if nobody joins. The role is usually worked alongside other agencies and alongside your own pipeline, so speed of submission is the winning strategy.

Retained

Paid in stages across the search whether or not a hire happens, and usually reserved for senior seats. You are buying committed effort rather than a race, and you carry the risk if the search does not land.

Temp and contract staffing

The person stays on the agency payroll and you pay a marked-up rate for as long as they work. This is a payroll and flexibility product, and comparing it with a permanent search is a category error that costs buyers real money.

A recruitment agency next to a talent acquisition function

Both find people. What differs is what each is paid for, what accumulates between searches, and what neither one can do.

Dimension	Recruitment agency	A talent acquisition function
What you buy	Introductions to candidates for one open role.	The hiring process itself, run end to end, across whatever roles are open.
Cost structure	A percentage of the hire's package, invoiced when they join. No hire, no invoice.	A flat engagement that does not move when the offer moves.
Incentive alignment	Revenue rises with the size of the offer and with the number of placements. Neither of those is your interest, and the good agencies are the ones that argue against their own contract.	A flat fee earns nothing from an inflated offer, so the number that closes is the number you set.
Capacity	Elastic in theory. In practice your role competes for one recruiter's attention against every other live role on their desk, and the ones closest to closing win.	Committed to your open roles, and the reading load moves off your calendar rather than onto it.
Continuity	The engagement ends at the placement. The next search starts from the job advert again, often with a different recruiter.	Every decision you make tightens the next shortlist, and the bar lives in the system rather than in one person's head.
What it structurally cannot do	Own anything after the person joins. The replacement clause covers the fee, not the empty seat or the quarter it costs you.	Replace your judgement. You still make the calls, and the function is only ever as calibrated as the feedback you give it.

The row that costs the most is continuity, because a re-search after a wrong hire is paid for twice: once under the replacement clause and once in the weeks the seat is empty again. Work out what an open seat costs you per week before deciding how much the fee model matters.

When a recruitment agency is genuinely the right call

This is not a rhetorical section. There are searches where an agency is the correct purchase and a function is not.

One role, once, in a market you have no reach into and no reason to build reach into. Contingency puts the search risk on the agency, which is exactly what you want when the need will not repeat.
A confidential replacement, where running any internal process would tell the incumbent before you are ready to.
A specialist pool where the agency holds real relationships rather than a database. Licensed, certified and regulated roles, where the people are known to each other and to whoever recruits them.
Genuinely lumpy demand. Two hires this year and none next does not justify building a function, and the honest advice is to keep the agency relationship warm.

What the agency model structurally cannot do

None of this requires an agency to behave badly. Each one follows from the contract, and the best agencies are the ones that spend their energy fighting their own incentives.

The fee moves with the offer: A contingency fee is a percentage of the package, so a higher offer is a larger invoice. When the number is being argued at the close, the agency is arguing for the candidate with its own revenue attached to the argument. A flat fee has nothing riding on it in either direction.
The unit of work is a submission, not a shortlist: When only a join is paid, sending one more profile is close to free for the sender and expensive for the reader. Volume is the rational strategy, and the filtering lands on your calendar rather than theirs.
Running three agencies makes this worse, not better: Every agency reads the same signal, that the role is contested and their odds are poor, so each submits fast rather than deep. The same candidates arrive from several desks with competing ownership claims, and you spend management time arbitrating a fee dispute over somebody you had already seen.
Nothing accumulates: The rejections you explained in March are not available in September. A new recruiter, or the same recruiter on a new desk, starts from the job advert again. You pay for the calibration repeatedly and never own it.
The replacement clause protects the fee: If a hire leaves inside the replacement period you get another search. You do not get the quarter back, the onboarding back, or the team's confidence back. The clause is written around the invoice because the invoice is the part the agency controls.

The number that separates the two fee structures is how many offers close inside the band the client set. That figure, the sample it comes from, and the caveat on it are published with the workings rather than asserted.

What a talent acquisition function does differently

The work itself is different, and so is what survives the engagement.

The brief becomes a success profile before anyone is sourced, calibrated with the people who will manage the role, so the shortlist matches the job rather than the job advert.
Screening happens before the calendar. Candidates are screened on 40+ signals, called and verified, and scored against the role, so a review is a decision rather than an investigation.
Your feedback is an input rather than a complaint. Every rejection changes how the next batch is screened, which is why an engagement gets sharper over months instead of flatter.
The engagement is flat, so the size of an offer changes nothing about what it earns.

Continuity1 is not a recruitment agency. There is no per-placement fee, no candidate ownership claim, and no invoice tied to what somebody is paid. If what you need is one introduction in a market you will never hire in again, an agency is the better buy and this is the wrong page.

How Continuity1 runs this funnel

The screening above is the job. These are the numbers it produces when a function owns it end to end, set against the published benchmarks for the same market.

Read the offer-band figure against the fee structures above. It is the number a percentage-of-package model has a reason to move and a flat engagement does not.

1 in 3: Shortlisted candidates you meet who become the hire; Aligned engagements run nearer 1 in 2, distant ones nearer 1 in 10. The market takes about 180 applicants to make one hire, and that sifting lands on your team rather than ours.; Continuity1 tracked engagements
95%: Offers that close inside your stated band; 20 of the last 21. A flat fee earns nothing from an inflated offer; a percentage of CTC earns more.; Continuity1 tracked engagements
20 to 49%: Shortlist-to-interview rate over one engagement; The industry curve points the other way; interview load keeps rising.; Continuity1 tracked engagement
~3: Interviews your team sits in, per hire; Ashby puts technical roles at 17.6 interviews per hire across the whole process, up 52% since 2021. The rest of that load sits with the function, not with you.; Ashby talent-trends report

Every brief becomes a success profile before sourcing starts, calibrated with the people who will manage the role. That calibration is the step most hiring skips, and it is why a shortlist either matches the job or matches the job advert.

You review a scored shortlist and make the calls. The filtering never lands on your calendar.

Share a role See the workings and the citations

Questions buyers ask

What is the difference between a contingency and a retained agency?

Contingency is paid only when someone joins, so the agency carries the search risk and works your role alongside others. Retained is paid in stages across the search whether or not it lands, so the firm commits time up front and you carry the risk. Neither is better in the abstract. Retained suits a small named pool, contingency suits a market where reach is the actual problem.

Should we give the same role to several agencies?

It feels like more coverage and behaves like less. Each agency now believes its odds are low, so each submits quickly rather than carefully. You receive duplicates, ownership disputes, and a shortlist optimised for speed of submission. One agency at a time with a real briefing beats three in a race.

How do we tell whether our problem is reach or screening?

Count the applications you already have against the hours anyone actually spends reading them. If a role draws hundreds of applicants and nobody has capacity to read past the first page, buying more reach adds to a pile nobody is clearing. If a posting draws twenty applicants in a fortnight and none are close, that is a reach problem and an agency may well solve it.

Does a replacement clause protect us if the hire does not work out?

It protects the fee. Read what triggers it, how long it runs, whether it survives a redundancy or a change of manager, and whether it refunds or only re-searches. Then price the part it does not cover, which is the seat sitting empty for a second time and the team absorbing the work again.