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 \u2014 consisting of setup matrices, standard operating procedures, and integration with Heijunka \u2014 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.<\/p>\n
If your department frequently changes production\u2014multiple times a day, on lines that have to keep up with an ever more heterogeneous mix\u2014you 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.<\/strong><\/p>\n SMED Single Minute Exchange of Die<\/a> \u2014 is the answer that the Toyota Production System gave to this problem more than fifty years ago. But in the way it is often \u201capplied\u201d 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.<\/p>\n This article talks about something else: Advanced SMED<\/strong>, what really works in production contexts to high mix<\/strong>, 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.<\/strong><\/p>\n There's a paradox many manufacturing companies are experiencing right now: The market demands increasing flexibility, ever-smaller batches, and ever-tighter delivery times.<\/strong> 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.<\/p>\n The point is precise: in a high-mix context, setup time isn't a fixed value\u2014it'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.<\/p>\n 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.<\/strong>.It's not a machine problem, it's a system problem.<\/p>\n The \u201cbasic\u201d version of SMED\u2014the one many are familiar with\u2014is 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<\/strong> The issue is that, in the reality of many departments, a significant portion of \u201cinternal\u201d 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.<\/p>\n So far, the logic is clear. The problem is that stopping here is enough to produce an episodic improvement, not a structural one.<\/p>\n Advanced SMED goes a step further in three key areas.<\/strong> First of all, work on the variability<\/strong> 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.\u00a0Secondly, it builds analytical tools \u2014 the\u00a0setup matrices<\/strong> \u2014 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.<\/span>\u00a0Thirdly,\u00a0Integrate SMED with production leveling<\/strong> \u2014 Japanese Heijunka \u2014 and with advanced planning and scheduling (APS) systems to make the system truly responsive, not just faster locally.<\/p>\n Not all changeovers are the same. The advanced SMED method begins with an analysis of the past six months\u201c history: how many changes per machine, of what type, and with what duration. Multiplying the frequency and duration gives the \u201dreal cost\" of each type of change, expressed in lost hours.<\/p>\n 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.<\/p>\n 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,<\/strong> Quality and process engineering.<\/strong> Not to multiply meetings, but because each change has different dimensions\u2014organizational, technical, qualitative\u2014that no single function can see entirely.<\/p>\n 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.<\/p>\n For each micro-phase detected, the team asks itself three precise questions: Can I delete it?<\/strong> Can I outsource it<\/strong>, meaning to move it by machine while it's running? Can I reduce it<\/strong> Or to simplify it? It's not free-form brainstorming \u2014 it's a systematic analysis that produces actionable items ranked by impact and implementation cost.<\/p>\n 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.<\/p>\n 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 \u201ctype change\u201d is not enough; it requires a parametric model<\/strong> able to accurately estimate every single variation in format or material<\/strong>. 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.<\/p>\n 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 levelingHeijunka<\/strong><\/a>and integrates with advanced planning and scheduling systemsAPS<\/strong>Together, they allow for real-time optimization of changeover sequences, reduction of lead time, and drastic reduction of inventory.<\/span>\u00a0The result is not \u2018working faster,\u2019 but creating a system capable of responding precisely and reliably to market variations.<\/strong><\/p>\n 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.<\/p>\n 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 \u2192 improvement \u2192 gradual return to chaos.<\/p>\n Advanced SMED addresses this risk through three tools. The first is the Structured documentation<\/strong>New standard sequence, setup checklist for each machine\/format, OPL (One Point Lesson) on critical steps, updated layout. Not bureaucracy \u2013 living tools used daily by operators.<\/p>\n The second is the Daily department meeting<\/strong>Fifteen 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.<\/p>\nWhy high-mix production makes changeover the real bottleneck<\/strong><\/h2>\n
What distinguishes advanced SMED from a superficial application?<\/strong><\/h2>\n
How to apply: step-by-step operational procedure<\/strong><\/h2>\n
The starting point: choose the right changeover to work on<\/strong><\/h3>\n
The team: SMED is not a solo job<\/strong><\/h3>\n
Video analysis: the detail that changes everything<\/strong><\/h3>\n
The three questions that guide every improvement<\/strong><\/h3>\n
Setup Matrices: The High-Mix Dashboard<\/strong><\/h2>\n
Heijunka and SMED: the virtuous cycle that slashes lead time<\/strong><\/h3>\n
Standards, monitoring, and expertise: the third level that makes all the difference<\/strong><\/h2>\n