When to run a retained search instead of posting the role

Picture the five people who'd be the strongest hires for this role today. Are they checking job boards? For most senior and specialist roles, the honest answer is no — they're employed, not unhappy enough to be looking, and would only move for a conversation, not a listing.

This test is deliberately uncomfortable, because most hiring managers, asked directly, already know the answer. The value of asking it explicitly is that it forces the decision to be made before the search opens, rather than three weeks into a contingent process that's quietly producing an applicant pool of people who were already looking — which, for a senior or specialist role, is very often a pool of people other companies have already passed on.

Posting works well for roles where supply is genuinely broad — early-career hires, high-volume functions, anything where a large pool of qualified, actively-looking candidates already exists. It breaks down fastest for roles with a narrow, specific scope: a founding engineer for a particular stack, a first data hire who needs to have done it before, an executive who has to be poached out of a role they're not unhappy in.

The tell we look for is how the role would read as a job ad. If the honest version of the ad is “we need someone who has already solved this exact problem at this exact stage of company, and there are perhaps thirty such people globally,” posting isn't going to find them — not because the ad is bad, but because none of those thirty people are checking job boards that week.

Across the searches we run, we track how each hire was actually sourced — inbound application versus direct outreach — against role seniority. The split is stark, and it gets starker the more senior or specialized the role.

Time spent identifying and reaching people who aren't applying anywhere, and a recruiter accountable to one search instead of splitting attention across a contingent pipeline. It costs more upfront. It's also the only approach that reaches candidates who were never going to see the ad.

The accountability difference matters more than it sounds. A contingent recruiter is paid only on placement and is rationally incentivized to work the searches most likely to close fastest — which tends to mean the easiest ones, not yours. A retained search commits a recruiter's time to one mandate, which is the only structure that makes sense when the whole premise is that the candidates worth reaching aren't going to come to you on their own.

A common middle-ground move is to post the role and run some outreach alongside it, treating the two as complementary rather than choosing. For broad roles this can work fine. For genuinely narrow, senior searches, it tends to produce the worst of both — a posted-role inbound pipeline that consumes real screening time without producing strong candidates, running alongside outreach that gets under-resourced because attention is split.

The clearer approach is deciding upfront which motion is actually going to produce the hire, and putting the bulk of the effort there. A small amount of posting for optionality is fine; treating it as a genuine parallel channel for a role where the target candidates aren't job-seeking usually isn't worth the screening overhead it creates.

The clearest mid-search signal that a posted role should have been a retained search is a pipeline that's active but not converting — plenty of applicants, several rounds of interviews run, nobody who's actually a strong fit for the scope. That pattern usually means the ad is reaching people, just not the right people, and no amount of additional posting time fixes a targeting problem.

The reverse signal also shows up: a retained search that's producing outreach responses but slow ones, on a role that turns out to have broader supply than expected. That's a sign the role could have been posted from the start, and it's worth revisiting the channel mix rather than assuming the original call was right just because it was made deliberately.

Most hiring teams don't need a rule for every role — they need a quick, honest gut-check applied consistently before a search opens, rather than after a posted role has already stalled.