Waiting: telling idle people apart from idle work

Two problems wearing the same label

Idle labour is a cost of capacity that was paid for and not used. Idle material is a delay to a customer and cash sitting on the floor. They behave differently and they have different remedies. Compare the time a part spends being worked on with the total time it spends in the plant and the imbalance becomes obvious. Yet almost every reporting system tracks labour utilisation in detail and material waiting not at all. Fixing the accounting blind spot usually matters more than any individual improvement, because you cannot manage a delay nobody records.

Finding out where work is actually stopped

Three methods complement each other. Walk the floor at random points during a shift and note what is stationary and for how long, tagging pallets with the date they arrived so ageing is visible without a system change. Pull the recorded start and finish stamps for a sample of jobs and look at the gaps rather than the durations. Ask supervisors what they were waiting for yesterday, and write the answers down. The pattern that emerges is usually a small number of recurring causes concentrated at a handful of places, not diffuse inefficiency everywhere.

Delay caused by support functions

A large share of stopped work is waiting for a decision or a service rather than for a machine: quality release on a first-off, a deviation signature, an engineering answer on a drawing query, a fitter, an inspector, a crane driver, a tool from the crib. These delays are invisible in production reporting because the responsible department has no measure attached to them. Start recording response time for each support function and publishing it. The conversation changes immediately, because the delay stops being a production problem caused by an anonymous elsewhere and becomes something with an owner.

The wrong way to remove operator idle time

Filling a gap with additional production creates surplus, occupies space and consumes material, so the plant swaps a modest labour cost for a larger inventory one. Better uses exist, but only if they are agreed beforehand and treated as legitimate work: preparing the next changeover, cleaning that doubles as inspection, cross-training on an adjacent operation, working through an improvement action, completing a maintenance task assigned to the operator. Write that list with the team, put it where they can see it, and make sure a supervisor walking past treats someone doing one of those things as working, not as skiving.

Structural waiting is a design problem, not a disruption

Some waiting is built into how the work is arranged. When consecutive operations have unequal cycle times, the faster one waits every single cycle, permanently, and no amount of firefighting will change it. That is a balancing and design question, owned by engineering and the area manager. Episodic waiting caused by breakdowns, shortages or absence is a different animal with different owners and different remedies. Separate the two before deciding what to do, because attacking a design problem with better reactions, or a reliability problem with a redesigned line, wastes the effort entirely.

Frequently asked questions

How can we measure waiting without installing a tracking system?

Date-tag material as it arrives at each queue and walk the area with a notebook. Within a couple of weeks you will know which queues age and which move, at a cost of a marker pen. A second cheap method is to follow one order through the plant personally and record the clock time at every handover. Neither is statistically elegant, and both produce enough evidence to identify where to look before anyone spends money on data collection.

Is some waiting between operations unavoidable?

Yes. Batch processes such as curing, heat treatment and plating impose their own rhythm, shared expensive equipment cannot serve every operation at once, and a deliberate buffer protecting a constraint is doing a job. The distinction that matters is whether the wait was chosen and sized by someone, or whether it simply happens. Buffers with an owner, an agreed level and a review are engineering. Queues nobody decided on are the thing to attack.

Operators say they are waiting for materials constantly. How should we check?

Ask them to log each occurrence for a fortnight with the part, the time and what they did instead, and keep the log short enough that it actually gets filled in. The pattern usually points to a small number of items and a specific handover, most often between stores and the line rather than between the plant and its suppliers. That distinction matters, because the remedies for a line-side supply problem sit inside the building and can be fixed without involving anyone external.