Delivering the placement · Lesson 15 of 18 · 7 min
Winning the client is half the business. Now you have a job order and a deadline, and no brand for anyone to apply to. Almost every guide to starting an agency stops before this part.
An established agency has inbound: people apply to it, its adverts get responses, its name means something in the market. A new agency has none of that, and pretending otherwise is how first job orders get missed.
So the operating assumption for your first year is that every candidate arrives because you went and got them, or because somebody referred them. That is a constraint, and it also happens to be how the best recruiters work permanently.
The same asset that produced your first clients produces your first candidates. People you placed before, people you interviewed and could not place, colleagues from your own career. Fastest route to a credible shortlist, and the one that costs nothing.
Searching for people who are doing the job now and contacting them directly. This is the core skill of sourcing, and it is where most placements in a specialist desk come from. It works in proportion to how specific your search is and how credible your first message sounds.
Ask every person you speak to — placed, rejected or unavailable — who else does this work well. People in a specialism know each other, and a referred candidate arrives with a reason to trust you. This is the highest-quality source there is and the most consistently neglected.
Posting the role and receiving applications. Cheap and fast, and it reaches only people who are actively looking — a minority of any market, and rarely the strongest. Useful for volume roles, weak for scarce skills. Treat it as one input rather than as the process.
Trade groups, professional associations, specialist forums, local meetups. Slow, and it produces people who are not on any list your competitors are working from. This is how a specialist desk builds an advantage that does not decay.
If you run contract placements, people whose assignments are ending are candidates you already know are reliable. Cheaper and faster than any external search, and covered in the contract desk lesson.
The same five steps every time, which is what makes a desk repeatable rather than heroic.
Get the actual requirement, not the job description
Job descriptions are written for approval processes and list everything anyone could want. Ask the manager which two things genuinely cannot be missing and what they would trade away. A search run against the real requirement finds people a search run against the document never will.
Work out where these people actually are
Which companies do this work, which titles it hides behind, which certifications or tools appear on the right profiles. Titles are inconsistent across employers, so searching by title alone misses a large part of any market.
Build the list before you contact anybody
Twenty or thirty named people, gathered in one pass. Contacting as you find is how you end up with an unrepeatable search and no idea of the size of the market. The list is also reusable — the same role will come round again.
Write to them as a person about a specific role
The same standard as client outreach: a reason it is them, one line about the opportunity that is actually distinguishing, and a small ask. Candidates in scarce skills get approached constantly and can tell within a line whether you know the work.
Follow up, then keep the ones who said no
Most will not reply first time, and most who reply will not be interested now. Both groups belong in your records with a note. The candidate who is not moving this month is frequently the placement six months from now.
Check yourself
You post a job advert for a scarce technical role and get eleven applications, none suitable. What does that tell you?
A strong candidate tells you they are not looking to move right now. What is the most valuable next question?
Every person you speak to is a record: what they do, what they are paid roughly, what would move them, when to come back. Placed, rejected, unavailable, uninterested — all of them.
This is the asset that separates a specialist desk in year two from the same desk in month one. When a familiar role comes in, you shortlist from people you already know instead of starting from nothing, and that speed is frequently what wins the placement against a larger competitor.
It does not need to be an applicant tracking system at the start — a structured spreadsheet is genuinely enough until volume makes it unreliable, which is the buying trigger the tool stack lesson describes.
The same conflict as business development, in the other direction. When you are filling a role, sourcing feels urgent and prospecting does not. When you have no roles, the reverse. Agencies oscillate between the two and stall.
The practical answer for a solo desk is to keep both alive at reduced capacity rather than to alternate — a little prospecting during a busy search, a little sourcing during a quiet week — because restarting either from zero costs more than maintaining it.
Next: what to do with the people you have found — screening, submitting and closing the offer.
Key takeaways
Mostly by contacting people directly rather than waiting for applications. Their own network and past placements first, then direct approach through professional networks, then referrals from other candidates. Job advertising plays a part, particularly for high-volume roles, but it only reaches people who are actively looking — which in most specialist markets is a small and not especially strong subset. A new agency with no brand has to be outbound almost entirely.
Direct approach and referrals, both of which cost time rather than money. Work your own network first, then search professional networks for people doing the job now and contact them individually, then ask everybody you speak to who else does this work well. Paid tools speed this up and change nothing about the method; agencies were built on it before the tools existed.
Sourcing is finding and engaging people who are not applying — building the list and making first contact. Recruiting is everything from there: qualifying, presenting, managing the process, closing the offer. On a small desk one person does both, but they are different skills, and the failure to distinguish them is why people conclude they are bad at recruiting when they are actually not sourcing enough.
By being specific enough that it is obvious you know the work. Candidates in scarce skills are approached constantly, nearly always with generic messages, and they filter on the first line. Name why you are contacting them specifically, say one genuinely distinguishing thing about the role — not "exciting opportunity" — and ask something small. Then follow up, because most replies do not come from the first message here either.
Yes, from the first search, and it does not need to be software at the start. Every person you speak to — placed, rejected, unavailable, uninterested — is worth a record with a note about what they do and when to revisit. The value compounds: within a year, a specialist desk that has kept records can shortlist from people it already knows, which is the difference between filling a role in days and starting from nothing.
Finished this lesson?
Next: Screening, submitting and closing the offer
Progress is saved in this browser only — no account, nothing sent anywhere.