Labour planning in manufacturing: matching people to the build plan

Translating the build plan into skills, not headcount

A headcount total conceals the problem. What matters is how many people qualified for each restricted task are rostered on each shift: the licensed forklift drivers, the setters for a particular machine family, the welders holding the relevant qualification, the operators signed off for a regulated process. Build the requirement from the routing and the plan, then compare it against a current skills record. Most plants discover that one or two qualifications are held by very few people, and that the whole schedule quietly depends on those individuals being present and willing.

Planning for the absence you already know about

Holiday, training, medical appointments and known leave are predictable and should be in the plan before the schedule is confirmed, rather than emerging as surprises on the day. The unplanned portion has a historical rate that can be estimated by area and by season, and building an allowance for it is more honest than assuming full attendance and covering the shortfall with unbudgeted overtime. Where absence is concentrated in one area or on one shift, treat that as a management signal rather than a planning parameter, because chronic absence usually reflects something about the work or the supervision.

Choosing between overtime, agency labour and hiring

Each lever suits a different duration of gap. Overtime covers short peaks but degrades quickly as fatigue accumulates and becomes an expectation that is hard to withdraw. Agency labour covers medium periods but carries a training cost every time the individual changes, and quality and safety performance usually reflect how well the induction is done rather than the arrangement itself. Hiring suits a sustained increase and takes time to become productive. The mistake is defaulting to whichever lever is administratively easiest, so decide against the expected duration of the gap and record the reason.

Productivity is a function of who is on the shift, not just how many

Replacing an experienced operator with a new one does not maintain output, and a shift heavy with recent starters will run slower, scrap more and need more supervision. Labour plans that treat people as interchangeable units produce schedules that fail for reasons that then get blamed on the individuals. Track the experience mix by shift alongside headcount, and adjust the expected rate when the mix is unusually thin. This also gives an honest basis for the conversation about how quickly a new intake can be absorbed without damaging the plant's delivery.

Keeping the qualification record trustworthy

Labour planning depends entirely on a record of who is qualified for what, and that record decays: qualifications expire, people move areas, training is delivered but never signed off, and the person maintaining the matrix leaves. Give it an owner, define what qualification means for each task, and audit a sample against the people actually doing the work. Regulated tasks carry a legal dimension, since operating certain equipment without current authorisation exposes both worker and employer. An out-of-date record is worse than none, because it produces confident plans that place unqualified people on restricted work.

Frequently asked questions

How far ahead should labour be planned?

Far enough ahead to act on the gap. If closing it means training somebody, the horizon must cover the training period plus the time to competence, which for skilled roles is considerably longer than the notice period for overtime. Run two horizons: a short one that rosters named people against shifts, and a longer one that compares required qualifications against the expected workforce so recruitment and training can start before the shortage bites.

Should we plan labour to the peak or to the average?

Plan the permanent workforce below the peak and cover the difference with a deliberate flexibility mechanism, because staffing to the peak leaves people underemployed in normal periods and staffing to the average guarantees chronic overtime. What matters is choosing the flexibility mechanism in advance — annualised hours, a trained relief pool, agreed short-term contracts — and agreeing it with the workforce, rather than improvising each time demand rises.

How do we stop the plant depending on a handful of key individuals?

Identify the dependency first by listing tasks that only one or two people can perform, then decide which of them would actually stop production. For those, fund cross-training with protected time rather than expecting it to happen alongside a full workload, and rotate the qualified people through the task often enough to keep the skill current. Recognise that the incumbent may resist, since scarcity has value, so make competence transfer part of how the role is assessed.