Copyright © 2026 Bonfiglioli Consulting. All rights reserved. P.I. 02646871208 Privacy Policy | Cookie Policy | Adjust Preferences
In a high-mix production context, changeover ceases to be a mere operational cost and becomes the real limitation to the flexibility of the entire system. This article explains why a superficial application of SMED is not enough, and how the advanced method — consisting of setup matrices, standard operating procedures, and integration with Heijunka — allows for stable and repeatable reduction of time and variability, without heavy investments. And why, to achieve lasting results, building structured internal expertise is worth much more than a thousand impromptu workshops.
If your department frequently changes production—multiple times a day, on lines that have to keep up with an ever more heterogeneous mix—you probably already know what it means to watch the clock during a setup change and wonder where those forty-five minutes of production went. The answer, in most cases, is not in the machine. It's in the method.
SMED Single Minute Exchange of Die — is the answer that the Toyota Production System gave to this problem more than fifty years ago. But in the way it is often “applied” in companies, it remains an unfinished project: times slightly improved, some good ideas from operators gathered in a workshop, and then, slowly, a return to old habits.
This article talks about something else: Advanced SMED, what really works in production contexts to high mix, when each change is slightly different from the other, and when budgets for new equipment are not available or not justified. A rigorous, replicable, and - above all - sustainable method over time.
There's a paradox many manufacturing companies are experiencing right now: The market demands increasing flexibility, ever-smaller batches, and ever-tighter delivery times. But every time we try to respond with a more complex schedule, the department gets bogged down with changeovers. And then we go back to producing large batches, keeping high inventories, and promising deliveries that we then struggle to meet.
The point is precise: in a high-mix context, setup time isn't a fixed value—it's a chaotic variable. Each change has its own duration; it depends on the operator, the product sequence, and how many pieces of equipment are already ready. In the absence of standards, planning can only defend itself: by inflating lot sizes, increasing buffers, and adding hours.
Shigeo Shingo, the Toyota industrial engineer who theorized SMED, had precisely identified this trap: the long time it took to change equipment. It not only limits productivity but also compromises the entire production architecture..It's not a machine problem, it's a system problem.
The “basic” version of SMED—the one many are familiar with—is limited to distinguishing between internal and external activities. Internal activities can only be performed when the machine is stopped; external activities can be carried out while the line is still in production. The fundamental intuition The issue is that, in the reality of many departments, a significant portion of “internal” activities has no technical reason to be internal: they are done with the machine stopped out of habit, due to lack of preparation, because no one has ever questioned the sequence.
So far, the logic is clear. The problem is that stopping here is enough to produce an episodic improvement, not a structural one.
Advanced SMED goes a step further in three key areas. First of all, work on the variability not just on average, but at the setup stage: a changeover that lasts twenty minutes on average but can range from ten to forty minutes is a process problem, not just an efficiency problem. Secondly, it builds analytical tools — the setup matrices — which allow each changeover to be broken down into its determining variables (format, tooling, material, machine parameters) and to calculate times parametrically, which is very useful for planning in high-mix contexts. Thirdly, Integrate SMED with production leveling — Japanese Heijunka — and with advanced planning and scheduling (APS) systems to make the system truly responsive, not just faster locally.
Not all changeovers are the same. The advanced SMED method begins with an analysis of the past six months“ history: how many changes per machine, of what type, and with what duration. Multiplying the frequency and duration gives the ”real cost" of each type of change, expressed in lost hours.
This analysis reveals two ideal candidates for the first project: a change with a greater overall impact on OEE and a change complex enough to represent the broadest company-wide setup cases. Working on these two cases allows for significant results to be achieved quickly and for a generalizable method to be developed.
One of the reasons internal SMED projects fail is that they are assigned to one or two people, often with technical expertise but lacking the big picture.. The method requires a cross-functional team that includes the department manager, the production planning and control department, the tooling department, Quality and process engineering. Not to multiply meetings, but because each change has different dimensions—organizational, technical, qualitative—that no single function can see entirely.
Video analysis represents the most powerful step for breaking old habits. It's not about simple timing, but a microscopic breakdown of every movement that reveals irrationalities invisible to the naked eye. The true value of this phase is not just technical, but also psychological: seeing reality on film allows operators' resistance to transform into active participation, shifting the focus from personal opinions to a shared standard sequence.
For each micro-phase detected, the team asks itself three precise questions: Can I delete it? Can I outsource it, meaning to move it by machine while it's running? Can I reduce it Or to simplify it? It's not free-form brainstorming — it's a systematic analysis that produces actionable items ranked by impact and implementation cost.
Most of the most effective actions have a very low cost: moving a toolbox, defining fixed reference values to eliminate iterative adjustments, preparing a cart with everything needed for the next changeover while the line is still in production. Nothing technologically sophisticated, but capable of significantly and stably reducing setup time.
In a high-mix context, equipment changeover time ceases to be a fixed data point and becomes a chaotic variable. To manage this complexity, optimizing “type change” is not enough; it requires a parametric model able to accurately estimate every single variation in format or material. Setup matrices act as a decision dashboard, allowing planning to move away from intuitive logic and towards scientific calculation, ensuring customers receive reliable delivery dates.
Reducing setup times is not an exercise in itself, but the key to redesigning the entire production architecture. When the cost of change plummets, the need to produce large batches to amortize it also decreases. This is where SMED merges with production levelingHeijunkaand integrates with advanced planning and scheduling systemsAPSTogether, they allow for real-time optimization of changeover sequences, reduction of lead time, and drastic reduction of inventory. The result is not ‘working faster,’ but creating a system capable of responding precisely and reliably to market variations.
When SMED and Heijunka work together, WIP goes down, lead time shortens, and responsiveness increases. Not because you work faster, but because the system is better orchestrated.
A SMED project that achieves improvements but doesn't establish a system to sustain them is destined to regress. This is well known by those who have experienced the cycle: workshop → improvement → gradual return to chaos.
Advanced SMED addresses this risk through three tools. The first is the Structured documentationNew standard sequence, setup checklist for each machine/format, OPL (One Point Lesson) on critical steps, updated layout. Not bureaucracy – living tools used daily by operators.
The second is the Daily department meetingFifteen minutes, every morning, in front of a physical or digital board. The previous day's anomalies are analyzed, decisions are made on how to address them, and attention is kept high on adhering to setup standards. In this simplicity lies enormous strength: SMED ceases to be a project and becomes a daily practice.
The third is the skill matrix: a skills map for each operator for each phase of the new setup standard. Identified gaps guide targeted training plans, both in the classroom and through on-the-job coaching. The operational autonomy of the employees is the true guarantee of long-term stability—not an IT system, nor a constantly present supervisor.
There's a question that many production managers ask themselves: “But can't we do it ourselves? We already have capable people.”
The honest answer is: maybe yes, but at what speed, with what methodological depth, and at what rate of errors that compromise the results?
SMED projects initiated internally without structured training tend to remain superficial: some good ideas for individual setups are found, but the complete methodology isn't built – matrices, monitoring, integration with planning, scalability to other lines. A changeover is optimized without knowing how to replicate it. The average is improved without addressing variability. Documentation is produced that no one updates.
A structured course brings something different: the method complete and verified on different realities, tools, the opportunity to compare with case studies from other sectors, and to receive feedback from those who have successfully led dozens of SMED projects. It means returning to the company not only with ideas, but with an ready operational plan and the expertise to replicate it independently tomorrow on another line.
At this point, it's useful to ask a concrete question: within your company, is SMED already a codified method—with matrices, standards, daily monitoring, and distributed skills—or is it still a series of good intentions applied on a case-by-case basis?
If the answer is the second one, it's not a failure – it's simply the most common starting point. Building the method takes time, but most importantly, it requires seeing it work in its entirety, not just reading about it in a manual.
Learn the structured method: enroll in the Lean Factory School® course SMED Method Course – Lean Factory School®
In departments with low levels of standardization—which are the norm in many Italian SMEs—reductions of 30–50% in setup time are realistic even after the first cycle of analysis and simulation, without investing in new equipment. The point is not the percentage itself, but rather building a system that makes it stable and replicable across multiple production lines.
This is precisely the context in which advanced SMED, with its parametric setup matrices, expresses its maximum potential. Instead of optimizing each changeover individually, setup time is broken down into its determining variables, and a model is built that applies to all changeovers. This makes planning in high-mix contexts much more precise and manageable—especially when integrated with advanced planning and scheduling (APS) systems capable of translating the parametric times from the matrices directly into optimized production plans.
No. SMED is often one of the first Lean tools to be introduced in production, because it produces visible results quickly and builds consensus around the continuous improvement method. The presence of 5S and Visual Management helps, but it is not an indispensable prerequisite.
Direct involvement in video analysis is the most effective response. When operators see themselves working, they independently identify irrationalities and become the first advocates for improvement proposals. Standardization becomes “our sequence,” not “the one imposed on us.”.
Internal training tends to stop at basic techniques, without developing the complete methodology—matrices, monitoring and integration with Heijunka, advanced planning systems (APS), and scalability across multiple lines. A structured course offers the complete methodological framework, real-world case studies from different sectors, and guidance from experts who have applied SMED in dozens of production environments. The result is autonomous internal expertise, not dependence on repeated consultations.
Copyright © 2026 Bonfiglioli Consulting. All rights reserved. P.I. 02646871208 Privacy Policy | Cookie Policy | Adjust Preferences