How to Assign Job Orders and Candidate Leads to Recruiters

Assign job orders by desk ownership first, then by capacity. The working model most staffing agencies land on is: client-owned accounts route to the owning recruiter automatically, everything else routes by specialty and territory to a qualified pool, and inside that pool the order goes to whoever has the fewest open reqs rather than whoever is next in line. Candidate leads route separately and faster, because a candidate who filled in a form is perishable in a way a signed job order is not. The failure mode to design against is not unfairness. It is a req sitting unclaimed for two days while three recruiters each assume someone else picked it up.

Job orders and candidate leads are two different routing problems

Agencies often try to run both through one rule set and end up serving neither well. The economics are different.

A job order is a commitment you already won. It can wait an hour for the right owner, and putting it with the wrong recruiter is expensive: they burn a day sourcing in a specialty they do not know, the submission quality drops, and the client notices. Job orders should route for correctness, weighted toward skill match, client relationship, and whether the recruiter has room to actually work it.

A candidate lead, or an inbound client inquiry from your website, is perishable. Response time is close to everything. These should route for speed, to whoever is available right now inside a reasonable qualification filter, with a fallback that never leaves the lead sitting. The numbers behind that urgency are in the speed-to-lead statistics. Getting these two backwards, routing job orders fast and sloppy while candidate leads wait for the perfect owner, is a common and quietly costly mistake.

The five inputs a real assignment rule uses

Order matters. Ownership overrides everything, specialty and territory define the eligible pool, and capacity plus availability pick the person inside it. Run capacity first and you will hand an aerospace req to whoever happens to be free.

Match the routing to your desk model

360 desks. One recruiter owns the client and the candidate side. Routing here is mostly about ownership and protecting it. New business is the only real rotation, and it should be weighted, because a recruiter with fifteen active reqs does not need the sixteenth more than the person with four.

Split desks. Account managers hold clients, recruiters fill reqs. This is where routing earns its keep: the job order arrives attached to an account manager, and then has to find a recruiter by specialty and load. Two routing decisions, not one, and most ATS setups only automate the first.

Pod or team desks. A pod owns a vertical. The order routes to the pod automatically, then inside the pod by load. Simplest to configure, and the one that most often degrades into cherry-picking if there is no assignment step inside the pod.

High-volume light industrial. Requisition counts are large, margins are thin, and speed to fill dominates. Straight load balancing with an availability check usually beats anything clever, because the reqs really are close to interchangeable.

Why the ATS usually is not enough

Bullhorn, JobDiva, Ceipal, and the rest all support an owner field and some form of workflow automation, and they will happily assign a record. What they generally do not do is decide fairly. The gap is consistent across the category: the ATS knows who owns a record but not who is busy, who is out today, or whose turn it is. So agencies fill the gap three ways, and all three have a cost.

• A shared queue everyone watches. Fast when the team is small and disciplined. It becomes cherry-picking as soon as it is not: the easy contract roles get claimed in minutes and the hard reqs age. Nobody is accountable because nobody was assigned.
• A manager who distributes by hand. Highest quality decisions, worst scalability. It also means every assignment is a judgment call the manager has to defend, and it stops entirely when they are on a client call.
• A spreadsheet rotation. Works for a quarter. Then someone goes on leave, the counter drifts, and there is a monthly argument about whose turn it was.

None of these is stupid. They are all reasonable responses to an ATS that does not model capacity. They just do not survive headcount growth. Modelling capacity is the missing piece, and workload balancing is what replaces the spreadsheet counter with a live count of open reqs per desk.

Speed to first contact is the number that moves

Every one of the manual approaches above adds latency between "lead arrives" and "a human is working it," and that latency is where placements are lost. A candidate who filled in a form on Tuesday at 2pm and hears from you Thursday morning has already talked to two other agencies. The same is true for inbound client inquiries, which are rarer and worth far more.

Two things fix it: automatic assignment with a real fallback so nothing waits for a claim, and taking work off the first conversation so recruiters can have more of them. On the second, a growing number of agencies now let an automated first-round screen handle the initial qualification conversation and route only the candidates who clear the bar to a recruiter, which changes the arithmetic on how many inbound leads one desk can absorb.

Rules worth writing down

1. Ownership is a field, not a memory. If the owning recruiter is out, the rule needs to know who covers, or ownership becomes a single point of failure.
2. Weight the rotation. A recruiter in week three should get a smaller share than one in year three. Equal splits punish the people carrying the desk.
3. Count reqs by weight, not headcount. One req for twelve warehouse roles is not one unit of work. If your capacity model cannot express that, it will drift out of touch with reality within a month.
4. Always have a fallback tier. "No qualified recruiter is available" must resolve to somebody, even if it is a manager. Unassigned is the one outcome with no upside.
5. Make the reason visible. When a recruiter can see the rule that gave a req to a colleague, the fairness argument ends. When they cannot, it never does.
6. Rebalance on departure, not eventually. When someone leaves or a desk shifts, the open reqs need redistributing by rule that week, not absorbed by whoever is nearest.

Automating it beside your ATS

Assigner sits beside Bullhorn, JobDiva, or whatever you run, and handles the assignment step your ATS does not: it distributes job orders and candidate leads by round-robin, weighted round-robin, specialty, territory, live open workload, priority, and availability, respects working hours and time zones, and names the rule behind every assignment so the split desk argument has an answer. It routes internal ops tasks and support requests through the same engine, and it starts at $12 per user per month. It is a companion to your ATS, not a replacement for it, and it is not a live two-way sync.

There is more on the agency setup on our staffing and recruiting agency lead routing page. If cherry-picking out of a shared queue is the specific problem you are trying to end, stopping cherry-picking from shared queues covers the mechanics, and weighted round-robin distribution covers how to handle ramping recruiters fairly.

Frequently asked questions

How do you assign job orders to recruiters?

Route client-owned accounts straight to the owning recruiter, filter everything else to a pool qualified by specialty and territory, then pick inside that pool by who has the fewest open reqs and is actually available. Ownership overrides rotation, specialty defines eligibility, and capacity breaks the tie. Always define a fallback so no req is left unclaimed.

Should job orders be assigned round robin?

Only in high-volume desks where reqs are genuinely interchangeable, such as light industrial staffing. For specialized recruiting, plain round robin ignores skill match and current load, so it hands hard reqs to whoever happens to be next. Weighted round-robin inside a specialty-filtered pool, with a capacity check, is the better default for most agencies.

Can Bullhorn assign job orders automatically?

Bullhorn can set an owner through workflow automation and supports distribution lists, so it can assign a record. What it does not do natively is decide fairly across a team using live open req counts, recruiter availability, and PTO. Agencies typically fill that gap with a shared queue, a manual manager step, or a routing layer beside the ATS.

How fast should you respond to an inbound candidate lead?

Treat it in minutes, not days. Inbound candidate and client leads decay quickly because most people contact several agencies at once, so the first agency to make real contact has a structural advantage. The practical fix is automatic assignment with a fallback tier, so no lead waits for someone to claim it from a shared queue.

What is the difference between a 360 desk and a split desk for routing?

On a 360 desk one recruiter owns both the client and the candidate side, so routing is mostly about protecting ownership and rotating only new business. On a split desk, account managers own clients and recruiters fill reqs, which means two routing decisions per order. Most ATS configurations automate the first and leave the second manual.