Outgrace, home
OperationsExplainer

Why changeover time sets the unit cost of small batches

Setup time is paid once per batch, not once per part. Whoever spreads that cost over fewer parts is setting the price of each one, often without seeing it.

Alarm clock, the summer time changeover, time change, wintertime, summer time, pointer
Photo: PIRO4D / Pixabay

Take a machine that needs an hour to change from one job to the next. It does not care whether the next run is twelve parts or twelve hundred. The hour costs the same either way. That fixed cost has to be carried by the parts in the batch, and in a plant that runs many short jobs it is carried by far fewer parts than most managers assume. Changeover time is therefore not a scheduling detail. It is one of the main reasons a small-batch part can cost more to make than its quoted price allows.

How it works

The arithmetic is simple. Add up what a changeover costs: the labour of the people doing it, the time the machine produces nothing, the scrap or rework from the first parts off the new setup, and any tooling or fixtures used up in the process. Call that total S. If the batch holds Q good parts, each part carries S divided by Q of setup cost on top of its material and running cost. Halve the batch and the setup share per part doubles. Nothing about the machine has changed; only the number of parts sharing the cost has.

This is why small-batch shops feel squeezed in a way their material costs do not explain. A plant can hold material prices steady, keep labour rates flat and still see margins fall because the job mix moved towards shorter runs. The fall shows up as a gradual erosion of gross margin on certain product families, usually the ones with the most changeovers, and it is rarely traced back to the setup line of the schedule.

Not every part of S is easy to see. Idle labour is visible if someone watches the clock. Scrap from first-off parts is visible if someone records it against the job. Lost running time is harder, because a machine that is down for setup takes capacity away from the next order, and that opportunity cost rarely appears in the job cost card. A sensible test is to ask, for each changeover, whether a customer order was delayed or a second shift was added because of it. If the answer is yes, part of the cost sits outside the batch and has to be added back before the unit price means anything.

Accounting rules add a second layer. Under IAS 2, the inventories standard that the International Accounting Standards Board adopted in April 2001, the cost of inventories includes conversion costs, and conversion costs are defined as direct labour plus production overhead, according to the IFRS Foundation's summary of the standard. Whether a given changeover cost belongs in inventory or in the period's expense depends on how the costing system allocates it, and that allocation is a question for your accountant. The standard also says write-downs to net realisable value, and all inventory losses, are recognised as expense in the period they occur. A batch made to stock that does not sell is therefore not just wasted capacity. It can become a loss in the accounts.

Trusted manufacturing advisors in the NIST Manufacturing Extension Partnership network, pe
nearly 1,400
Date the IASB adopted IAS 2 Inventories, per the IFRS Foundation
April 2001

Who it affects

The most direct audience is the small and medium-sized manufacturer. NIST's Manufacturing Extension Partnership exists to strengthen manufacturers of that size in the United States, and it delivers its help through state-designated MEP Centers, according to NIST. The network names defense, aerospace, automotive, energy, electronics and medical devices among the critical industries it serves. These are the sectors where job shops, contract machinists and short-run assemblers are common, and where a single product family can involve many changeovers in a week.

The second audience is the finance function of any company that holds inventory under IFRS reporting. For them the question is not how to make the machine faster but how the cost of getting there is recorded. A plant manager who shortens changeovers changes the cost per unit, and the controller needs to know how that change flows through the inventory valuation and the period's expense.

The public data is thin for this subject. The Census Bureau's Annual Survey of Manufactures has been folded into the Annual Integrated Economic Survey, and the survey pages we checked contain no setup, changeover or batch-size data. The benchmark tables give establishment statistics but no cost or batch figures. A leader looking for an industry benchmark for changeover time will not find one in these sources. Measuring your own plants is the only reliable comparison available.

What leaders get wrong

  • They treat changeover as a scheduling problem and leave it to the production planner, when its cost lands in the price, the margin and the inventory valuation.
  • They reduce batch sizes to cut stock without first measuring the setup cost per part, and then discover that the unit cost rose faster than the stock fell.
  • They copy a changeover reduction that worked in another plant without checking whether its product mix, tooling and operator skills match their own.
  • They count only the minutes the machine is stopped and ignore the first-off scrap and the lost capacity, which is usually where the larger part of the cost sits.
  • They run long batches of slow-moving products to avoid changeovers, and do not check the write-down exposure on the stock they create.

What to do about it

Start by measuring, not reorganising. For the three machines that cause the most changeovers, time each changeover from the last good part of one job to the first good part of the next, and record it for a month. Then cost it: labour, idle time, scrap and any lost orders. Divide by the average batch size to get the setup cost per part, and set that figure beside the material and running cost on the job card. If the setup share is small, the batch rules are probably fine. If it is large, the batch rules and the changeover methods both deserve a closer look.

When you look for reductions, start with the setup tasks that can be done while the machine is still running, such as staging tools, materials and fixtures before the stop. In our experience these are cheap to test and easy to reverse, and they often remove more time than a redesign of the machine. Make the case for any larger change with the measured figures, not with an estimate. The sources we reviewed for this guide say nothing about the methods themselves, so the specific techniques that lean and industrial-engineering texts describe need checking against those texts before they go into a plant plan. Where you want outside help, the NIST MEP Centers offer value stream mapping and hands-on consulting, though the network does not name changeover reduction as a specific service. Ask the centre directly whether it has done this work, and what it measured. Finally, ask your accountant to show you, in writing, how setup labour and idle time are allocated into inventory, and whether any batch made for stock carries a write-down risk. Those answers decide whether the changeover saving you find shows up in the accounts at all.

Newsletter

The Briefing

One email a week: the leaders we profiled, the ideas worth stealing, and nothing else.

Weekly, free, one click to leave. Read our privacy policy.