Retained vs contingency: the math that actually matters
The debate usually gets framed as "cost vs quality". That is the wrong frame. The real question is which model produces the right economic outcome for the specific hire you are running. For most senior specialist searches the answer is not close.
Every quarter we get a client who wants to talk about the choice between retained and contingency. Most of the conversations follow the same shape - “we usually run contingency, is this specific role different, and if so why”. The honest answer is that the choice is not a matter of preference. It is a matter of what actually produces the outcome for the specific role you are hiring.
Below is how we think about the math when we are advising a client on which model to use. The advice is not the same for every hire.
When contingency wins on economics
Contingency works well when three things are true.
The role is in a market where the realistic candidate pool is large enough that speed and volume actually help. Mid-level software engineers in Sydney is a good example. There are several hundred candidates who could plausibly fit a well-run mid-level search. Multiple agencies running in parallel produce a wider funnel and a faster time-to-fill.
The role is well-understood by the client and the client has good internal signal on what “good” looks like. When the client can quickly recognise a strong candidate on paper, the volume-plus-triage model works. The client absorbs the triage cost and the market signal is efficient.
The cost of a slightly-suboptimal hire is manageable. For a mid-level role at a large firm, a candidate who turns out to be a B+ rather than an A is a productivity loss the firm can absorb. For a senior role at a small firm, that same B+ can be the difference between a product shipping and not shipping.
For roles that fit those three criteria, contingency is the right economic model. There is no reason to pay retained-search fees for a role where the volume-plus-triage model works.
When contingency loses on economics
Contingency loses when three things are true.
The pool is small and named. Once you are running a search into a candidate pool of eighty or fewer people, the volume-plus-triage model stops working. Multiple contingency agencies calling the same eighty candidates produces the same result each time: candidates hear from three or four recruiters in a fortnight about the same role. They discount the role. They discount every recruiter contacting them. The client’s brief starts to look like a low-conviction search. Serious candidates step away.
The engagement is sensitive to sequence and signalling. In small specialist markets, the order in which candidates hear about a search matters. A first conversation with the top candidate, run properly, is worth much more than a fifth conversation with the same candidate. Contingency, which structurally does not control sequence, will regularly send the wrong candidate to the top first. That is not a criticism of the agencies. It is a structural feature of the incentive model.
The candidate you eventually hire is expected to add compounding value. If your senior hire is worth two or three million dollars of shipped revenue over three years (which most senior specialist hires are), then a fifteen or twenty thousand dollar difference in search fee is not a variable worth optimising on. The variable worth optimising on is the probability that the search closes with the right person.
The compounding cost of a bad senior hire
The uncomfortable truth about most senior specialist searches is that the cost of a bad hire is not the salary paid to the bad hire. It is the opportunity cost of the eighteen months of leadership drift that follows. For a scaleup, a bad Head of Engineering hire can push a Series-B round by six months. A bad Sales Director hire can lose a quarter. A bad first regulatory-affairs hire in medtech can push a TGA submission by an entire regulatory cycle. Those costs dwarf the search fee for the good hire.
Retained search does not eliminate the possibility of a bad hire. It does two things that materially change the probability distribution. It sequences the conversations properly, so the top-of-shortlist candidates hear about the role in the right order and with the right conviction. And it forces the client to write a brief that is specific enough to be actionable - the process of taking a retained brief regularly reveals that the client had not actually decided what they were hiring for, and that reveal alone saves months.
The framing we advise clients to use
The question we advise clients to hold is not “which is cheaper”. It is: given the specific role, given the specific pool, given the specific cost of getting this wrong, which model produces the highest probability of hiring the right person on the right timeline?
For mid-market volume roles, that answer is contingency. For senior specialist roles in small markets, that answer is retained. The mid-cases - senior generalist roles in mid-market pools - are genuinely close and can go either way based on the specifics of the firm and the role.
The trap most firms fall into is running senior specialist searches on the contingency model out of habit and then wondering why the searches keep failing. The answer is that the model was the wrong one for the specific search from day one.