Sourcing

What does qualified have to mean before sourcing starts?

Define the acceptance test before the first search, in writing, with the hiring manager in the room. Qualified should be a checkable claim: this person has done this scope of work, at this level, in a context close enough to ours, and there is a plausible reason they would consider a move. Left undefined, qualified drifts toward whoever the search happened to return, and the disagreement surfaces weeks later as a manager rejecting a shortlist without being able to say why. A written definition also makes the work reviewable rather than a matter of taste.

How should sourcing be measured without punishing quality?

Measure at the point where somebody other than the sourcer has made a judgement. Profiles sourced is an activity count the sourcer controls entirely and can inflate. People who passed a recruiter screen, or reached a hiring manager, cost nothing to game because a second person had to agree. Pair that with reply rate, which says whether the targeting and the message are working, and with a slower measure such as offers accepted from sourced candidates. Read all of it against how hard the market is, since identical effort produces different numbers in different markets.

How does sourcing capacity change what a hiring plan can promise?

A hiring plan is a capacity commitment whether or not anyone treats it as one. If the plan calls for a number of specialist hires in a quarter, and the sourcing team can thoughtfully work only a limited number of people a week, the arithmetic either supports the plan or it does not, and the gap will not close through urgency. Doing that calculation out loud before the quarter starts converts a familiar argument about recruiter performance into a resourcing decision: add sourcers, buy external help, narrow the brief, or move a start date.

Should sourcing sit with a dedicated sourcer or a full-desk recruiter?

Two operating models dominate. In the full-desk model one person owns a requisition from intake through offer, doing their own market research, outreach, screening and closing. In the split model a dedicated sourcer generates and qualifies people while a recruiter runs the process from first screen onward. The split buys depth, because research and outreach reward uninterrupted blocks of time that a recruiter fielding interview logistics never gets.

Neither model is correct in the abstract. A team hiring a handful of roles a year cannot justify a specialist and should keep the desk whole. Split roles earn their cost where markets are hard, volume is steady, or the same profile is hired repeatedly. The common failure is a nominal split: someone is titled sourcer but still carries requisition ownership, so process work eats the research time the title was meant to protect.

Why does always-on pipeline building die first when hiring gets busy?

Sourcing runs in two modes. Project sourcing is attached to one named requisition with a hiring manager waiting and a start date behind it. Always-on sourcing builds familiarity with a market the company will keep hiring from, ahead of any specific opening. Both are legitimate work. Only one of them has somebody chasing it, which is why the second is what quietly disappears the moment a busy quarter arrives.

The displacement is predictable rather than a discipline problem, so treat it structurally. Ring-fence the hours in the calendar, give the always-on work a named output that gets reviewed like any requisition, and report on it separately so its absence becomes visible. Teams that leave it as whatever is left over after the urgent work finish every quarter with nothing banked and begin the next search from a blank page.

What actually limits how many qualified profiles one sourcer produces?

The binding constraint is rarely message volume. It is research time per person: reading a background properly, confirming the work is genuinely relevant, finding a reason the approach should interest them, then writing something that is not a template. Add reply handling, scheduling and keeping records straight, and the number of people one sourcer can process thoughtfully in a week is far smaller than an outreach tool implies.

Market depth and brief clarity move that number more than effort does. A well-populated market with an unambiguous brief yields good profiles quickly, while a scarce market with a brief that changes weekly yields rework. This is exactly why raw profile counts make a poor target. Counted alone they reward volume over judgement, and the fastest way to hit the number is to quietly lower the bar on who gets included.

Where does the handoff between sourcer and recruiter leak?

Four leaks recur. A profile arrives with no note on why the person was approached or what was said, so the recruiter restarts the conversation cold. Nobody agreed how fast a handed-over candidate gets picked up, so warm interest goes stale across a long weekend. Rejections come back with no reason, so the sourcer never calibrates. And the candidate meets a second stranger who asks the same opening questions all over again.

The protocol that fixes it is unglamorous. Write the handoff into the applicant tracking system rather than a chat message: what was claimed, what was promised, what the person asked about. Agree a pickup window and hold both sides to it. Send a short reason back to the sourcer for every rejection. Introduce the recruiter by name in the final message, so the candidate is told about the switch instead of discovering it.

Should sourcing be built in-house, outsourced, or bought as research?

Three procurement options behave differently. An in-house team keeps market knowledge, brand voice and candidate relationships inside the company and compounds over time, but scales slowly and carries fixed cost through a hiring freeze. Outsourced sourcing flexes up for a burst of hiring and flexes down again, at the cost of that knowledge walking out when the contract ends. Research vendors sell name lists and structured market data only.

The trap in the third option is that a list is not a pipeline. Names arrive unqualified and unengaged, so outreach, screening and relationship work still land on your team, and buying more names without the capacity to work them merely moves the bottleneck. Whichever route is chosen, agree in writing what a delivered candidate means, who owns the relationship, and what happens to the data when the engagement ends.