Hiring fifty people is not hiring five people ten times. The process that works for a handful of roles falls apart at volume, usually in the same predictable ways.
What changes at volume
Sourcing has to be continuous, not campaign-based. A campaign produces a burst of applications that is exhausted halfway through the ramp. Continuous sourcing maintains a pipeline that refills as it drains.
Screening has to be identical across every applicant. When one recruiter screens twenty candidates, consistency is automatic. When six recruiters screen six hundred, standards drift — and they drift downward as the deadline approaches.
Onboarding happens in cohorts. Individual start dates do not scale. Training capacity becomes the binding constraint, not sourcing.
Where quality actually collapses
It is rarely a decision. It is drift.
Week one, the bar is clear. Week four, you are behind target, and a candidate who is nearly good enough gets through because the alternative is missing the deadline. Week six, that is the new normal. Nobody decided to lower the standard, but it is lower.
The defence is to make the standard mechanical rather than a judgement call under pressure.
Define the minimum before you start. Not a wish list — the actual floor. What must be true for this person to succeed in the role? Usually three or four things.
Score against it, consistently. A simple rubric applied by everyone beats a sophisticated one applied inconsistently.
Track pass rates by week. If your screening pass rate climbs as the deadline approaches, the standard is slipping. That number is your early warning.
Plan backwards from the training calendar
The usual mistake is planning forward from today. Volume hiring works backwards from the date people need to be productive.
Work out the training capacity per cohort, the number of cohorts available, and the attrition between offer and start — which is always higher than people expect at volume. That gives you how many offers you need per week, which gives you how many candidates you need to screen per week, which gives you how much sourcing is required.
If that number is impossible, you know now rather than in week five.
Build in a reserve pipeline
At volume, no-shows and early attrition are not exceptions, they are a rate. People accept offers and do not start. People start and leave within a fortnight.
Plan for it. Keep a warm reserve of candidates who have been screened and would accept. The cost of carrying that reserve is far lower than the cost of a half-empty training cohort.
Agree weekly intake targets
A single deadline at the end of the programme hides problems until it is too late to fix them. Weekly targets surface a shortfall in week two, when there is still time to widen the specification, adjust the salary, or add a sourcing channel.
Agree them with whoever is delivering the hiring, and review them weekly without exception.
What to watch
Four numbers tell you most of what you need:
- Screening pass rate — rising means standards are slipping
- Offer acceptance rate — falling means your package or your process has a problem
- No-show rate — rising means the gap between offer and start is too long
- Thirty-day attrition — rising means the role was mis-sold, or the screening is passing the wrong people
Any of these moving in the wrong direction is a signal you can act on while the programme is still running.
The thing most people get wrong
Treating volume hiring as a sourcing problem. It almost never is. The constraint is usually training capacity, process throughput, or the offer-to-start gap.
Adding more candidates to a pipeline that is already backed up makes the numbers look better and changes nothing.
Where to start
Before anything else, establish your actual training capacity per week. Everything else is derived from it, and most volume programmes that go wrong were planned without that number.
Sizanid runs volume hiring programmes across contact centre, manufacturing, logistics, and retail. If you have a ramp coming, [tell us the shape of it](/find-talent) and we will tell you honestly whether the timeline works.
