How are setup times reduced in production?

Changeovers are one of the most underestimated areas of loss in manufacturing. They are planned. They are visible in the schedule. They are often not counted as a "stop". But a changeover where the machine is idle for three hours is three hours of unused capacity, regardless of whether the stop was planned or not.

For many factories, changeover time is the single greatest lever to increase capacity, meet varying demand, and improve OEE. And it is an area where improvements can often be made without major investments. It is about methodology, measurement, and sustained effort.

This guide walks you through how to reduce changeover times based on the SMED methodology, what is required for the change to last, and why measurement is crucial from day one. Throughout, we assume that working to shorten changeovers is a continuous improvement process. A platform that makes losses visible ensures that the work can be driven systematically over time, not as a single project that loses momentum after six months.

Why is changeover time a strategic issue?

Changeover time is strategic because it determines three things for a factory or production manager: capacity on the existing machinery, flexibility to meet customer demand, and competitiveness in a market with varying volumes.

Capacity on the existing machinery. Every hour spent on changeovers is an hour not producing. Halving changeover times on a machine that is changed over five times a week, from four hours to two, yields ten extra production hours per week. Per year, this amounts to 500 hours, which represents a significant portion of a line's total capacity.

Flexibility to meet customer demand. Long changeover times force long production runs. Long runs tie up capital in inventory, make the factory sensitive to changing customer requirements, and degrade delivery performance. Shorter changeovers do the opposite: you can run smaller batches, deliver more frequently, and hold fewer finished products in stock.

Competitiveness in a market with varying volumes. When demand changes rapidly, the factory that can change over quickly wins. When competitors with longer changeover times hesitate, you have already started producing the new item.

This strategic perspective means that changeover work should not be treated as an operational improvement exercise. It is a management decision that affects how the factory can act in the market.

What is SMED and how does the methodology work?

SMED stands for Single-Minute Exchange of Die and is a method for shortening changeover times in three steps: separate internal and external changeover, convert internal tasks to external, and streamline the remaining ones. The term was coined by Shigeo Shingo in Japan during the 1950s and 60s, and the original ambition was to be able to change over tools in less than ten minutes. The methodology has since become a cornerstone of lean and has been used successfully in all types of manufacturing, from the automotive industry to food production.

SMED is built on three steps.

Step 1: Separate internal and external changeover

The first, and often largest, step involves classifying every task in the changeover.

Internal changeover includes work that can only be performed when the machine is completely stopped. This applies, for example, to tool changes, adjustments inside the machine, or other tasks that require a safe stop with lockout-tagout procedures.

External changeover is work that can be performed while the machine is still producing. This includes fetching tools, preparing materials, reading instructions, and setting up the workspace.

In many factories, a large part of the external work is done while the machine is stopped. Moving this work to before or after the actual stop can in itself shorten the changeover time by 30 to 50%, without requiring any individual activity to be performed faster.

Step 2: Convert internal to external

The next step is to see which internal tasks can be converted to external with small changes. Pre-heated tools can be installed directly without waiting. Quick-connect couplings mean that a process that previously required disassembly can be connected in just a few seconds. Standardized fixtures eliminate adjustment steps.

This step often requires small investments in equipment or workspace design. These are almost always minor compared to the value of the capacity that is freed up.

Step 3: Streamline both internal and external

The third step is about making each remaining task as fast and reliable as possible. Parallel work where multiple people can perform different tasks simultaneously. Standardize the working method so that two operators can perform the changeover in the same way. Visual aids and checklists help ensure nothing is forgotten or done in the wrong order.

This is where the fine-tuning happens. The first 50% of the time savings often come from steps 1 and 2. The rest comes from persistent work in step 3, often over several months.

What is the most common mistake in changeover work?

The most common mistake is assuming you know how long changeovers take without measuring them. In many factories, measurements show that changeovers take significantly longer than management believes.

There are several reasons. Often, only the time the machine is stopped is counted, not the time required to ramp up to full speed with approved quality. Minor stops and adjustments during the first hours after the changeover are often missed. And changeover times often vary greatly depending on the product, day, shift, and operator, making the average figure misleading.

Before you can reduce changeover time, you must know what it actually is. This requires measurement that captures:

• From the last approved unit of the previous product

• To the first approved unit of the next product

• Including warm-up, fine-tuning, and quality verification

• Per product transition, per shift, per operator

This data reveals something unexpected. The difference between a fast and a slow changeover for the same product transition is often large. This variation is valuable because it shows that the potential for improvement already exists in the factory, among your own operators. The only question is how to get everyone to perform at their best.

What does operator involvement mean in changeovers?

Operator involvement means that operators are driving the improvement work, not just receiving instructions. Factories that succeed in long-term changeover improvements have something in common: the operators lead the work.

There are two reasons for this. The first is that operators see details that no one else sees. Which tools are missing on site? Which adjustment steps are required on a specific line but not on another? Which tasks prevent parallel work? This knowledge is not found in time studies from a consulting team. It is found on the shop floor.

The second is that the change only lasts if the operators own it. A standardized working method dictated from above will dissolve within a few months. A standardized working method built and refined by the operators themselves lives on.

At Bostik in Helsingborg, which manufactures adhesives and sealants with over 80 products on the same site, this was central to their improvement work. Through systematic DMAIC work with the operators, changeover times on the filling machines were reduced by 70%. And OEE improved by 40%. One of the clearest insights during the work came from a seemingly simple observation: two operators working together on one changeover completed it faster than two working on separate machines. The platform's measurements showed the difference, and it then became the new standard way of working.

What should a system do to support changeovers?

A system should do four things: measure changeovers consistently, make the variation between changeovers visible, link the measurements to products and downtime causes, and support the work in daily operations. This turns changeover work into a systematic process instead of a one-off project that loses momentum.

Measure changeovers automatically and consistently. From the last approved unit to the first approved unit, per product transition, per shift, per operator. Manual measurements wear out. Automatic ones last.

Make the variation visible. An average changeover time says very little. The spread between fast and slow changeovers of the same product transition says a lot. This should be visible directly in the platform, not require a separate BI analysis.

Link to products and downtime causes. Some product transitions are costly. Others are cheap. Being able to prioritize the most expensive product transition, or the one that occurs most frequently, is the difference between structured improvement work and random efforts.

Support the work in daily operations. Changeover times should be displayed on dashboards in morning meetings and weekly improvement meetings. Otherwise, it becomes a monthly report that no one acts upon.

This is close to the core of what an OEE system should do. Measurement is the means. Improvement is the goal. The platform becomes the engine for daily work, not a reporting system that lives on the side.

How do you get started with shortening changeover times?

Get started with six steps. Choose a critical product transition, measure for three weeks without changing anything, video record the changeover, classify each task as internal or external, introduce a standardized working method and test run it, and only scale once it works.

1. Choose a critical product transition. Not the easiest. Not the hardest. Choose one where changeovers occur frequently and where an improvement would free up visible capacity. That is where proof can be built.

2. Measure for three weeks without changing anything. Establish an honest baseline. Include the variation between operators and shifts.

3. Video record the changeover. The value of this cannot be overstated. What operators believe happens, and what actually happens, often differ dramatically. The video becomes the basis for analysis.

4. Classify each task as internal or external. Using stopwatches and the video recording as a guide, gather the operators around a table. Discuss every task. Determine what can be moved, what requires equipment, and what just requires discipline.

5. Introduce a standardized working method and test run. Write it down. Put up visual aids. Test run ten changeovers using the new method. Measure. Adjust.

6. Scale when it works. Not before. Once the first product transition is faster and the results hold across shifts, spread it to the next product transition and the next.

It is quick to win a lot. It takes perseverance to not go back to old habits.

How does Good Solutions work with changeovers?

The platform from Good Solutions is built to drive improvement work in daily operations. This also applies to changeovers. Automatic measurement captures the time from the last approved unit to the first approved unit, per product transition and per shift. The variation is visualized in dashboards used by operators, production leaders, and improvement teams in their ongoing meetings. Downtime causes and product data are linked so that the right efforts can be prioritized.

The platform is delivered with expert support from delivery consultants with production experience. The training program for operators and superusers ensures that the working methods are implemented. A dedicated Customer Success Manager follows the customer over time.

Among the results is Bostik, which shortened changeover times on filling machines by 70% and improved OEE by 40%. System 3R, a Swedish company in tooling and automation solutions, has an explicit goal of reducing changeover times to under ten minutes, in line with the original SMED ambition.

Read more about how others have increased their factory productivity

FAQ

How much is it possible to shorten changeover times?
Experience from various industries shows that an improvement of 30 to 50% in the first year is realistic with a systematic SMED methodology. Factories that work persistently for 2 to 3 years often achieve more, sometimes 70% or more. The largest portion is won early on, by separating internal and external changeovers. The rest comes from persistent work on each sub-task.

Do we have to invest in new equipment to shorten changeover times?
Not primarily. A large part of the improvement comes from methods and organization. Moving external work from downtime to runtime usually only requires new ways of working. Later investments in quick-connect couplings, pre-heated tools, or standardized fixtures can yield additional results, but they should come after the basic work is completed.

Who should lead the changeover work?
Improvement management, production management, and operators together. Improvement management is responsible for the methodology and follow-up. Production management ensures time and resources for the work. Operators drive the practical application. Without operator involvement, the change will not last.

How do we measure a changeover correctly?
From the last approved unit of the previous product to the first approved unit of the next product, including warm-up, fine-tuning, and quality verification. The measurement should occur automatically, per product transition, per shift, and per operator, so that the variation is visible. Manual measurements are a reasonable first step but do not hold up over time.

How do we know that the improvement lasts?
Through continuous measurement. Changeover times should be visible on dashboards in daily operations, not just reported monthly. When the variation between changeovers begins to increase again, it is an early sign that working methods are slipping. At that point, the work around the standard should be reinforced, rather than questioning the measurement.