Where did that CV go? What recruitment software actually has to do

A hiring manager asked how many candidates were still in play for a role that had been open for six weeks. The HR executive opened her inbox, then a WhatsApp group, then a spreadsheet, then the job board's dashboard, and came back forty minutes later with a number she was not confident in. Two of the candidates on her list had accepted offers elsewhere a fortnight earlier. One had been interviewed twice by different people who did not know about each other. The role was not hard to fill. The pipeline was simply invisible.

Recruitment software is bought to solve sourcing — more candidates, better job boards — and it rarely does. What it can solve, and what most businesses actually need, is visibility: where every candidate is, who owns the next step, and how long they have been waiting for it.

REQUISITION   the role, its approver, its opening date, its brief
PIPELINE      the stages this role's candidates pass through
CANDIDATE     one record, however many times they apply
STATUS        exactly one stage per candidate at a time
OWNER         who has the next action, and since when
NOTES         interview scores against the brief's criteria
OUTCOME       hired, declined, withdrawn, rejected - and why

If any row is missing, the inbox is still the system.

The single-status rule is the one that matters. A candidate is at one stage, owned by one person, with one date on it. That is what lets a hiring manager see, in one screen, that four candidates are waiting on him and have been for nine days. It is also what lets HR see that a role's pipeline has gone quiet at the same stage three times, which is a process finding rather than a sourcing one.

The most common disappointment is a business that buys recruitment software before it has agreed how it hires. Who approves a requisition? Who screens, against what? How many interviews, with whom, scored how? Who makes the offer, and within what band? A system configured over unanswered questions encodes whichever answers the person configuring it assumed, and hiring managers then work around it, and within two months the WhatsApp group is back.

Settle the process first — on paper, with the three or four managers who hire most — and the configuration is a morning's work. This is the same lesson every HRMS implementation teaches, and recruitment is where it bites earliest because hiring is the process managers feel most entitled to run their own way.

1. 1

   Requisition and approval

   A role opens when someone with the authority says so, in the system, against a brief. Not when a manager sends an email.

2. 2

   Screening criteria

   Written into the requisition before the first CV arrives. The screener applies them; the system records which candidates met which.

3. 3

   Interview stages and scorecards

   Fixed per role type. Scores entered independently before the debrief, which is what makes the debrief worth holding.

4. 4

   Offer authority

   Who may offer, within what band, and what needs escalating. The system holds the band; the manager cannot exceed it without a visible approval.

Candidate experience is talked about as a separate goal, and it is mostly a by-product. A pipeline in which every candidate has an owner and a date is one in which nobody sits unacknowledged for three weeks, because the three weeks are visible on a screen with a name beside them. The automated acknowledgement email helps a little; the manager seeing that eleven people are waiting on him helps far more.

The same visibility fixes the double interview from the opening: one candidate record, one pipeline, and the second interviewer can see the first interviewer's scores before scheduling anything.

A candidate who was second choice for one role is often first choice for the next, and a business with no candidate record hires them from scratch six months later — re-sourcing, re-screening, re-interviewing someone it already knows. The system's talent pool is the answer, and it only works if outcomes are recorded honestly at the time: not just 'rejected' but why, and whether they should be approached again. That one field, filled in at the moment of decision, is worth more than any sourcing integration.

The same record answers the question nobody can currently answer: how many people applied to us last year, and what happened to them. Where that number is surprising — hundreds of applicants and a handful of hires, most of them lost at the same stage — the pipeline has found something the HR process audit would otherwise have to dig for.

• A weak brief. A role described as 'dynamic self-starter' attracts the wrong pool and the system will track them efficiently. The brief is recruitment process consulting work, not configuration.
• Slow decisions. The system makes delay visible. It cannot make a hiring manager decide. That is a management conversation the visibility enables.
• Sourcing. Job-board integrations post the same advertisement to more places. If the pool is wrong, it is wrong in more places.
• Interviewer judgement. Scorecards structure the conversation. They do not make a manager who hires on gut feel stop doing so without training.

Time-to-fill is the number everyone reports and it hides more than it shows. The measures a pipeline makes possible are more useful: time at each stage, which finds the bottleneck; drop-off by stage, which finds the step that loses good candidates; offer acceptance rate, which is the first signal that the market has moved; and early attrition of hires by source, which is the only measure of whether the pipeline is producing people who stay.

Report these monthly to the managers who hire, not only to HR. A hiring manager who sees that his stage is where candidates wait longest, in a number beside his name, moves faster than one who is told the process is slow. The pipeline's value is that the delay has an address.

The last one connects recruitment to everything after it. A source that produces hires who leave inside a year is expensive whatever it costs per hire, and only a system that carries the candidate record through to the employee record can show it — which is the argument for recruitment management living inside the same system as onboarding and the employee master rather than as a separate tool.

For a business hiring a few dozen people a year, the choice is between a standalone applicant-tracking tool and the recruitment module of an HRMS. The standalone tools are better at sourcing integrations and worse at everything after the offer — the record has to be re-entered somewhere else. The module is the reverse. Most growing businesses are better served by the module, because the handover from candidate to employee is where records go wrong, and a single system removes the handover.

Either way, the system is worth buying only after the process is agreed. Run one role through the agreed process on paper first. If that works, recruitment management software makes it repeatable and visible. If it does not, no system will, and the paper run has just saved the licence fee.