Standard work: the current best method, agreed by the people who run it

Sequence, rate and what sits between the steps

Three things need fixing before a job counts as standardised. The order of work elements, because a different order changes both the time taken and the failure modes. The rate the operation must sustain to meet demand, which determines how the content is divided between stations. And the quantity of material held within the cycle, since an operator who keeps extra pieces at the bench is running a different process from one who does not. Documents that describe only the sequence are common and incomplete; without the rate and the in-process quantity, two operators can both follow the standard and produce very different results.

Written from watching, not from intention

The standard should describe the best method currently observed, which means watching several people do the job and taking the elements that work. Engineers writing from the process design produce something the floor cannot follow, and it is quietly ignored. Expect disagreement about whose method is adopted, and settle it with observation rather than seniority. This part of the exercise is also where much of the value appears, because comparing how three experienced operators do the same task reveals both a better method and the reasons each of them developed a private workaround.

A standard exists in order to be replaced

The point of writing down the current method is to give improvement something to measure against. If a standard has not been revised since it was issued, either the process is perfect or nobody is using it, and the second is far more likely. That implies a revision route quick enough to be worth taking: who may change what, who confirms it, how the change is communicated to other shifts, and how quickly the document at the machine is updated. When revision takes weeks, people improve the job and stop bothering to record it, and the written standard drifts away from reality.

What breaks a standard on a real floor

Standards are usually written for the ideal condition and then meet material from a different supplier, a substitute component, a machine behaving differently after a repair, a mixed model sequence, or a new starter who has not built up the manual skill assumed by the timings. The response should not be to pretend these conditions do not occur. Write the recognised exceptions and what to do about them, define who to call, and record how often each exception is used. A standard covering only the good day guarantees that most shifts are run on undocumented judgement.

Confirming the standard without policing the operator

A supervisor watching a job should be checking whether the method still works, not catching somebody out. When work is done differently, the useful first question is why, since the answer is usually a real obstacle: a part that will not seat, a tool that has gone missing, a container in the wrong place. Deviation treated as information keeps that intelligence flowing. Deviation treated as a disciplinary matter drives it underground, and the plant loses its earliest warning that a process has changed. The exception is deliberate departure from a safety or regulatory requirement, which is a different conversation entirely.

Frequently asked questions

Does standard work apply to high-variety or one-off production?

The content of a specific job cannot be standardised when every job differs, but the surrounding elements can. How a setup is performed, how material is checked on receipt at the station, how a first-off is presented, how the workspace is arranged, how a job is handed over between shifts: all of these repeat regardless of the part. Plants making engineered products often get more from standardising these routines than a repetitive plant does from standardising a cycle.

Who should own the standard, engineering or the production team?

The production team should own the content, because they perform it and will otherwise ignore it, with engineering owning anything that touches process parameters, tooling or product characteristics. In practice a joint arrangement works: the team proposes and drafts, an engineer confirms the technical elements, and the supervisor approves and issues. What must be avoided is a document controlled so tightly that a small improvement cannot be reflected in it for weeks.

How does standard work relate to time standards used for costing?

They come from the same observation but serve different purposes and should not be conflated. Costing standards need an allowance structure and a degree of stability so that financial comparisons mean something, while the operational standard needs to change whenever a better method is found. Keeping one document to serve both makes improvement awkward, because every change triggers a costing revision. Link them by review at a sensible interval rather than by making them the same record.