4M Change Management System — Cost and Payback in a Thai Plant

Ask for a quote for a 4M change management system at a plant in Thailand and the proposal will almost always be built around one idea — moving application and approval onto a screen. The paper 4M change notice becomes a workflow, and who approved what and when is finally on record. Nothing about that is wrong. But on its own it does not shrink the containment scope by a single unit when a change-induced defect reaches the customer, because nobody can name the first lot the change actually affected. This article follows the numbers all the way through, using a Japanese-owned Tier-2 automotive parts plant in Rayong, Thailand as the model — how much the containment scope changes with and without lot linkage, which layer the money sits in, and which layer costs you money to cut.

Some plants install a 4M change management system and the containment scope never gets smaller

Let us pin down the model plant first. A Japanese-owned Tier-2 automotive parts plant in Rayong, Thailand. 420 employees, 6 production lines, output of 12,000 pieces per day (1,500 pieces per hour over 8 hours), 25 operating days per month for 300 days a year, or 3,600,000 pieces a year. IATF 16949 certification is already in place. 4M change notices circulate on paper and the history lives in an Excel ledger. The form came from the parent company in Japan, was translated into Thai, and the layout itself has barely changed in twenty years.

When a plant like this starts talking about installing a 4M change management system, the first set of requirements is almost always the same. Put the application form on the web. Designate approvers in the system. Send a reminder email when an approval stalls. Export an approved notice as a PDF. Make everything searchable for audits. Every one of those is the mirror image of a daily frustration with paper, so it is natural that they come out first.

And after go-live, some of those frustrations really do disappear. Notices stop vanishing under a pile on someone’s desk. Lead time to completed approval drops from 3.2 days to 0.4 days. Auditors stop writing findings about approvers who cannot be identified. So far, so expected.

The problem is what happens next.

Six months after go-live, a customer complaint comes in for a dimensional defect. Trace the process and the likely cause is a machining condition change at one operation. At that point the plant has to answer exactly one question. From which lot did that change start to take effect?

On paper or on a screen, the answer comes back in the same shape. “The change notice was approved on 3 July.” The approval date is known. But the approval date and the moment the change actually reached the product are two different things. It might have been applied the next day at a changeover, or three days later during a different production plan. The notice has a “planned application date” field, and nothing guarantees the floor switched over on plan. Above all, the notice says nothing anywhere about which lot numbers were running at that moment.

When you cannot answer, containment has to be cut by date. Go back to the day the last customer audit closed, or the day the last 100% inspection was performed, and hold everything from there until today as the suspect period. At the model plant that comes to 28 days. 28 days multiplied by 12,000 pieces equals 336,000 pieces. The cost of sorting those 336,000 pieces comes straight out of the plant’s own pocket.

This is the most common disappointment in a 4M change management system rollout. Digitizing the application form speeds up approval, but it does not change the containment scope. The only thing that changes the scope is the link between the change point and the production lot. In a design that just lifts paper onto a workflow, the number of boxes you have to hold when a defect appears is identical before and after digitization.

At the model plant, the share of changes for which the first affected lot can be identified sits at 35% today. For the other 65% the notice records alone are not enough. Sometimes cross-checking the equipment changeover log gives an answer, but that is a human doing a manual reconciliation after the fact, not the system answering. Build the linkage into the design and that share becomes 100%. The gap between 35% and 100% is not a difference in precision — it comes back as a difference in containment scope, measured in baht.

The largest category of 4M change is Man

There is one more thing worth confirming before you look at a quote. Which category of 4M change is actually occurring at this plant, and how often? Decide the scope of your notices without pinning that down with real counts and the highest-volume category ends up outside the system.

The breakdown of 620 changes a year, and what the 380 Man changes contain

The model plant actually generates 620 4M changes a year. The breakdown by category is as follows.

Of the 620, 380 changes — 61% — are Man changes. Machine at 90, Material at 60 and Method at 90 add up to 240, which still does not reach 380. The phrase “4M change management” makes most people picture a die swap or a material lot switch, but by count the lead role belongs to people.

The 380 Man changes split four ways.

The 150 new assignments are unavoidable as long as turnover in Thai manufacturing stays where it is. The 130 reassignments follow naturally from reshuffling line staffing against a moving production plan. The 60 returns after long absence are operators who spent weeks away from the line coming back to the same process. The 40 support cases are people pulled in temporarily from another line or another process, and they are the hardest of all to keep on record.

Here is the practice most plants settle into. Machine, Material and Method are subject to a 4M change notice. Operator changes are not. The floor’s reasoning holds together. People move around almost every day. Filing a notice for all of it would consume the whole day. Staffing is part of the daily production plan, not a change management matter. So notices are limited to occasions when someone touches equipment or conditions.

The result is that only 240 of the 620 annual changes survive in the ledger as formal 4M change notices. The remaining 380 exist nowhere in the plant as a change record. What makes that awkward is that the effort is still being spent — it is just not being recorded. When an operator changes, there is a handover at the morning meeting, a handover memo, a senior operator standing alongside for the first day, and a verbal report to the line leader. The plant pays the labor for 620 events and can retrieve records for 240. That is the real state of a plant running on paper and Excel.

What you lose when operator changes sit outside the notice system

Leaving people out of the notice system costs you more than the name of who built the part. Three things actually disappear.

The first is the operator’s skill certification status. At the moment a newly assigned operator entered that process, what certification level did they hold? Were they certified to work unsupervised, or conditionally certified with a trainer alongside? That information usually does survive on the training record side, but it is not linked in the form of “from this date, at this process, a person at this level took over”. Pulling out training records after a defect appears is useless if you cannot establish who was standing on the line at that hour.

The second is the tacit variation in work conditions. Even under the same work standard, the fine detail of the method shifts when the operator shifts. How the workpiece is set down, the feel of a tightening torque, what gets checked first in a visual inspection. Whatever the standard does not spell out genuinely varies by person. The 60 returns after long absence are especially prone to trouble, because the operator believes they are working exactly as before while process-side conditions may have moved in the meantime.

The third is the time window in which a support worker was present. The 40 support cases are usually decided that morning, communicated verbally, and processed in the attendance system under the original department. In the plant’s records, that person was on another line. Try to slice the suspect window by operator and the support worker was never even in the search population.

When those three are missing, the human factor cannot be examined after a change-induced defect. And because it cannot be examined, the countermeasure lands on the same conclusion every time — retrain to the work standard. Quality record digitization, and cutting across in-process data, was covered in our article on the quality data management system; 4M change management is the entrance to it. Unless what changed and when is on record, no volume of measurement data will support the most basic analysis there is — comparing before and after the change.

The workable landing point is not to file a formal notice for every operator change. It is to stop using a single form and split it by impact. New assignments and returns after long absence become proper notices with first-piece verification and a supervised start. Reassignments and support workers become a “record only” category that requires no approval and is linked to a clock-in at the line terminal. Separate the weight of approval from the existence of a record and you can keep all 380 on file while carrying a lighter approval load than today.

Linkage changes the containment scope by a factor of 50

This is the core of the article. Here is what the presence or absence of linkage is worth in money, using the model plant’s numbers.

28 days and 336,000 pieces versus 4.5 hours and 6,750 pieces

Set the containment target for a suspected change-induced defect side by side, with and without linkage.

Both calculations are simple. Without linkage, 28 days multiplied by 12,000 pieces per day equals 336,000 pieces. With linkage, the window in which the change was genuinely in effect can be pinned to 4.5 hours, so 4.5 hours multiplied by 1,500 pieces per hour equals 6,750 pieces.

336,000 divided by 6,750 is a factor of roughly 50.

Same plant, same defect, same root cause. The only difference is whether the change point and the production lot are linked. That alone moves the number of products you have to hold by a factor of 50.

In money, at 2 THB per piece for sorting, that is 672,000 THB against 13,500 THB. Assume containment happens twice a year at the model plant and the annual sorting cost goes from 1,344,000 THB to 27,000 THB. The gap is 1,317,000 THB. Keep that figure in mind — it reappears in the return calculation later.

And sorting cost is only the visible part of containment. Holding 336,000 pieces means putting that much inventory on shipping hold, telling the customer about a delivery delay, pulling people off the line to sort, and in some cases arranging an external sorting contractor. 6,750 pieces is settled in-house the same day. Whether or not you have to explain yourself outside the plant gates sits inside that factor of 50.

Our article on the cost of building traceability breaks that cost down layer by layer; 4M change management is the part of that structure that covers the change-point side. Make only the product side traceable and you can follow where a part went, but not what was different about the conditions it was made under. Production change traceability is what closes that second half.

The three interfaces linkage requires

So what does it take to link a change point to a production lot? Three interfaces. They feed directly into the cost discussion below, so let us be precise about them.

The first is the interface to production result lots. This is the record held by the production management system — from what time to what time, which lot number ran on which line. Only when the effective time of a 4M change can be overlaid on that timeline does the first post-change lot become definite. Put another way, without this one interface you can slice the effective window by time and never convert it into product numbers.

The second is the interface to equipment operation and changeover. This is the equipment-side record of when a machine stopped, when the changeover finished, and when it restarted. Most of the 90 Machine and 90 Method changes are applied at a changeover. This interface is what lets you capture the moment the equipment actually started running under the new condition, instead of the “planned application date” written on the notice. Its job is to absorb the gap between the plan on the form and reality.

The third is the interface to operator attendance and skills. Keeping the 380 Man changes on record means tying the attendance data held by the time and attendance system, and the process-level skill certifications held by the training records, to a line and a time window. Support workers fall out of the records because attendance is organized by department rather than by line. Move to clocking in at a line terminal when entering a process and the question of who was standing at that process during that window becomes definite, support workers included.

With all three in place, one change notice connects through time and line to three things — lot number, equipment condition and operator. Containment is simply tracing those connections backwards. Conversely, lose any one of the three and the linkage does not hold. Without the production result interface it never lands on product; without the equipment interface the effective time is never fixed; without the operator interface 380 of the 620 annual changes stay out of scope. It only works as a set of three, and that point comes back in the cost discussion.

The cost is built out of four layers

Quotes for a 4M change management system can differ severalfold depending on who wrote them. Most of the reason is that the quotes are not describing the same thing. Split them into layers and they become comparable.

Layers 1 to 4, initial and annual cost

Here is the cost of a full-scope build at the model plant, split into four layers. Figures are in THB.

The breakdown of each layer is as follows.

Layer 1, workflow, 850,000 THB. The workflow platform is 550,000 THB and the design of forms and approval routes is 300,000 THB. The design costs 300,000 THB because one route is not enough. Combining the 4M categories (Man, Machine, Material, Method) with impact level produces 12 route patterns. A new operator assignment and a change of material supplier need different approvers, different verification items, and a different answer on customer notification. Force them through a single route and you get one of two outcomes — heavy approval applied to a light change, stopping the line, or a heavy change slipping through on a light approval. The annual cost is a cloud subscription of 15,000 THB per month over 12 months, or 180,000 THB.

Layer 2, shop-floor input terminals, 780,000 THB. Twelve line terminals at 52,000 THB each come to 624,000 THB, and installation plus network work is 156,000 THB. Twelve terminals across six lines is two per line, positioned at the entry and exit of a process or at the main changeover points. When there are not enough terminals, operators walk to the office every time they file, and that is the moment “I will enter it all later” begins. Annual maintenance is 10% of the initial cost, or 78,000 THB.

Layer 3, linkage to production results, 1,050,000 THB. The three interfaces described above at 350,000 THB each. Production result lots, equipment operation and changeover, and operator attendance and skills. It is the most expensive of the four layers, and as we will see, the one you must never cut. Annual maintenance is 10% of the initial cost, or 105,000 THB.

Layer 4, embedding, 420,000 THB. Procedures and training in Japanese, Thai and English at 260,000 THB, and one month of on-site support after go-live at 160,000 THB. Three languages are needed in a Thai plant because the person filing is a Thai line leader, the person approving is a Japanese section manager, and the party you show at an audit is an English-speaking customer. That month on site exists so that someone is standing on the floor to settle the variation in judgment — “which category do I file this under?” — that always appears right after go-live. The annual cost is local support at 14,000 THB per month over 12 months, or 168,000 THB.

Total initial cost is 3,100,000 THB and the annual cost is 531,000 THB.

When layers get mixed together, quotes stop being comparable

Line up quotes from several suppliers without this four-layer split and comparison simply does not work. Three mismatches are common.

The first is whether Layer 2 terminals are in the quote at all. A supplier quoting software only will mark terminals as customer-supplied. The quote looks 780,000 THB cheaper, but the plant buys them itself later and total spend is unchanged. Worse, once they are customer-supplied it becomes vague who owns the 156,000 THB of installation and network work, and that argument surfaces right before go-live.

The second is whether Layer 3 has been pushed outside the main quote as an option. Three interfaces at 1,050,000 THB is the largest single item of the four layers. Move it out of the body of the quote and into a footnote and the headline total becomes 2,050,000 THB. It looks cheap. But that 2,050,000 THB proposal will never make the containment scope smaller. The next two sections put a number on that difference.

The third is whether Layer 4 is in the quote at all. Procedures, training and go-live support can be dropped from a quote and the system will still run. What was dropped is either done by the plant itself or not done at all. If it is not done, six months later you have a fully formed situation in which people are entering data into the system and the data no longer matches reality.

Re-sorting quotes into layers is worth doing on larger systems too. The layer discipline we set out in comparing MES implementation costs applies just as directly to 4M change control manufacturing projects. Do not compare totals — first make sure the same things sit in the same layers, then compare totals.

The annual benefit is 3,953,400 THB, itemized

After cost comes benefit. Here is the annual saving from a full-scope rollout at the model plant, by source. Figures are in THB.

Let us check the basis for each.

Change-induced defect escapes, 1,920,000 THB. Loss per escape is taken at 480,000 THB, and 7 events a year become 3. The 480,000 THB covers everything other than sorting — writing the customer report, emergency sourcing of replacements, freight, and the engineering effort behind the corrective action. The reason 7 becomes 3 is that the 380 Man changes come into scope and the approval route splits into 12 patterns, so changes that used to slide past as “just part of the day” now trigger first-piece verification.

Sorting cost during containment, 1,317,000 THB. As calculated above, two containment events a year take 1,344,000 THB down to 27,000 THB. This single line item comfortably exceeds the system’s annual cost of 531,000 THB.

Effort to process change notices, 334,800 THB. 620 cases multiplied by 2.4 hours is 1,488 hours, and at 300 THB per hour that is 446,400 THB. After the rollout, 620 cases multiplied by 0.6 hours is 372 hours, and at 300 THB that is 111,600 THB. The 2.4 hours per case is, for the 240 filed notices, the total of writing the paper form, chasing down the approver, collecting the stamp, transcribing into the Excel ledger and filing the original; for the 380 unrecorded changes it is the total of handovers, handover memos and day-one supervision. All 620 already consume effort today. The 0.6 hours is the value once both categories are entered at a line terminal and the only remaining wait is the approval notification.

Line stoppage waiting for approval, 324,000 THB. 18 events a year multiplied by 1.5 hours is 27 hours, at 18,000 THB per hour, or 486,000 THB; this becomes 6 events a year multiplied by 1.5 hours, or 9 hours, at 18,000 THB, or 162,000 THB. Once lead time to completed approval goes from 3.2 days to 0.4 days, halting a changeover to wait for approval largely stops happening.

Preparing material for customer audits and PPAP, 57,600 THB. 8 times a year multiplied by 24 hours is 192 hours, at 400 THB per hour, or 76,800 THB; this becomes 8 times multiplied by 6 hours, or 48 hours, at 400 THB, or 19,200 THB. The 24 hours is the real effort between a customer asking for every 4M change in a given period and the list being finished by cross-checking an Excel ledger against paper files. With linkage in place, it becomes specifying a period and exporting.

From here, the payback period.

• Annual benefit 3,953,400 THB minus annual cost 531,000 THB equals a net annual benefit of 3,422,400 THB
• On a monthly basis, 285,200 THB
• Initial cost 3,100,000 divided by 285,200 equals roughly 11 months

Over five years, the picture is as follows.

• Five-year total cost (initial plus five years of annual cost) equals 3,100,000 plus 531,000 multiplied by 5, or 5,755,000 THB
• Five-year benefit equals 3,953,400 multiplied by 5, or 19,767,000 THB
• Five-year net equals 14,012,000 THB

Looked at alone, an initial cost of 3,100,000 THB is not small for an IT investment at a plant in Thailand. But payback is roughly 11 months and the five-year net is more than four times the initial cost. The largest benefit at 1,920,000 THB and the second largest at 1,317,000 THB are both losses that occur when a change-induced defect appears. In other words this is not an investment to increase output — it is an investment to stop a loss the plant is already incurring every year. That character matters for the decision in the next section, and for the Thai market discussion after it.

Phasing that drops Layer 3 is the option that saves you the least

Layer 3, linkage to production results, is the most expensive of the four at 1,050,000 THB. Naturally, a proposal appears. “Let us start with Layers 1, 2 and 4, and consider Layer 3 in the next phase.” Initial cost drops from 3,100,000 THB to 2,050,000 THB. As phasing, it looks perfectly reasonable.

Here are the numbers side by side.

Check the phased breakdown. Cost is Layer 1 at 850,000 (180,000 a year) plus Layer 2 at 780,000 (78,000 a year) plus Layer 4 at 420,000 (168,000 a year), giving an initial 2,050,000 THB and an annual 426,000 THB. That much is straightforward subtraction.

The problem is on the benefit side.

The 1,317,000 THB of containment saving disappears entirely. Without Layer 3 there is no lot linkage, so the containment scope stays at 28 days and 336,000 pieces. As shown above, that is the second largest item in the benefit table.

The defect escape saving is halved as well. Full scope took 7 events a year down to 3, worth 1,920,000 THB. Without Layer 3 it only goes from 7 to 5, which is two events, or 960,000 THB. The coverage of the notice system widens, so escapes do fall, but without the ability to chase the early post-change lots and stop them quickly, the size of the reduction shrinks.

Audit material is halved too, at 28,800 THB. You can still specify a period and export a list, but the list carries no linked lot information, so anything the customer asks for beyond that goes back to manual work.

The 334,800 THB of processing effort and the 324,000 THB of line stoppage come through in full. Both are benefits produced by the Layer 1 workflow and the Layer 2 terminals, so neither depends on Layer 3.

Add the remaining benefits — 960,000 plus 334,800 plus 324,000 plus 28,800 equals 1,647,600 THB. The 1,317,000 THB of containment saving is zero, so none of it enters the total. Against the full-scope 3,953,400 THB, a little over 40% survives.

Now the net. Net annual benefit is 1,647,600 minus 426,000, or 1,221,600 THB, which is 101,800 THB a month. Payback is 2,050,000 divided by 101,800, or roughly 20 months. Five-year total cost is 2,050,000 plus 426,000 multiplied by 5, or 4,180,000 THB, and the five-year benefit is 1,647,600 multiplied by 5, or 8,238,000 THB, so the five-year net is 4,058,000 THB.

Dropping Layer 3 saves 1,050,000 THB up front. But the five-year net falls from 14,012,000 THB to 4,058,000 THB, a loss of 9,954,000 THB. And payback slips from 11 months to 20 months. You are trading 1,050,000 THB of initial cost for 9,954,000 THB thrown away over five years.

Put another way — cutting the layer that looks most expensive is the cut that saves you the least. On this plant’s numbers, you can state flatly that Layer 3 is the one part of a 4M change management system you must not remove.

Phasing itself is not the problem. The problem is the order in which you cut. If initial cost genuinely has to come down, there are real options — roll the 12 terminals out in stages, shorten the on-site support from one month, build the approval routes starting from the main categories instead of all 12 patterns at once. Each of those can be added later, and none of them breaks the linkage. Layer 3 is the exception, because when you add it later you cannot retroactively link the notice data already created under Layers 1 and 2. The missing period stays a period you can never contain properly.

4M change management Thailand — the points that differ from Japan

Everything so far has been internal to the model plant. Operating in Thailand adds several conditions a plant in Japan never has to design for.

When production is not growing, 4M changes increase rather than fall. Thailand’s automobile production in 2025 was 1,455,569 units, down 0.9% year on year. Passenger cars were 550,456 units (down 1.4%) and commercial vehicles 905,113 units (down 0.6%). Export-bound output fell to 956,230 units (down 5.2%), while output for domestic sales rose to 499,339 units (up 8.6%). The figures were released by the Federation of Thai Industries (FTI) on 28 January 2026 and reported by JETRO on 5 February 2026. FTI has also revised its 2026 production outlook down from an initial 1.5 million units to 1.45 million units (reported 27 July 2026). The reasons cited are uncertainty in the global economy, the situation in the Middle East, trade restriction measures in various countries, and intensifying competition in the EV market.

What that does to a plant is make it impossible to plan long runs of a single model. Lots get smaller and the number of model switches and changeovers rises. More changeovers means more Machine and Method notices, and matching headcount to swinging volumes means more reassignments and more support workers. Production volume falls, and the count of 4M changes does not. It goes up.

And when investment in new lines is frozen, the proposals that clear internal approval are not the ones that add capacity — they are the ones that stop a loss already occurring. As the benefit table shows, of the 3,953,400 THB a year, the 1,920,000 THB of defect escapes and the 1,317,000 THB of containment make up the bulk. The benefit of this investment comes from stopping losses, not from adding capacity. That makes it the kind of investment that is easier to justify in a year when production is flat. That is the argument to put in front of a board that has frozen capacity spending.

IATF 16949 change control and customer notification. Clause 8.5.6 of IATF 16949 requires control of changes, and the automotive requirement includes the plant judging which changes need customer approval and notifying in advance. The practical difficulty is whether the floor can sort, at the moment of filing, which of the 620 annual notices are subject to customer notification. Notify everything and the customer cannot process it; miss a notification and it becomes an audit finding. Designing the Layer 1 approval route across 12 patterns exists partly to embed that sorting into the form itself. Make the choice of category and impact level automatically determine whether customer notification is required and to whom, and the judgment stops depending on the experience of whoever is filing. How the IATF requirement combines with product-side traceability is covered in automotive parts traceability and IATF 16949.

Multilingual notice screens, and lead time when the approver sits at head office in Japan. A three-tier structure is common — a Thai line leader writes the notice, a Japanese manager at the Thai plant approves it, and high-impact changes require approval from an engineering department at head office in Japan. Two things stretch lead time in that structure, translation and time difference. Prepare the same form in Japanese, Thai and English, increase the number of selection fields and reduce free text, and translation waiting largely disappears. For the time difference, decide in advance which categories genuinely need head office approval by limiting them to the top impact levels, and design delegated approval so everything else closes on the Thai side. Getting approval lead time from 3.2 days to 0.4 days requires both of these to be built into the form design. Changing the rules after go-live means rebuilding approval routes, and that is expensive.

How to bring support workers and agency staff into the notice system. The 40 support cases a year include agency-supplied labor at plants in Thailand. These people are often not registered in the plant’s attendance system the way permanent staff are, and in some cases have no ID at all. Leave that alone and the very window you most need on record — the hours when someone unfamiliar with the process was working it — is the one that goes missing. The practical answer is to issue every person, agency and support staff included, an ID that works at the line terminal, and to make clocking in at the process the entry point of the routine. There is no need to require approval. Keep the record reliably and containment can be sliced by operator.

Seven items to settle before you compare quotes

Here is everything above turned into the items the plant should settle before issuing a request for quotation. Send a request to several suppliers with these seven undecided and the figures that come back will not be comparable.

Item 1 — where to draw the line on what requires a notice. For each of Man, Machine, Material and Method, what falls in scope? Above all, how are the 380 Man changes handled? Are all of them approval-bearing notices, or are they split into an approval category and a record-only category? Change the population of 620 a year and both the Layer 1 design effort and the Layer 2 terminal count change with it.

Item 2 — approval tiers and delegated approval. How many approval tiers are there? Which categories require head office approval in Japan? Who acts when an approver is absent? Design without delegated approval and line stoppages waiting for approval remain. Design with it and the scope of delegation has to be set per category.

Item 3 — the granularity of linkage, by lot or by time. Do you link by lot number, or by start and end time? Lot-level is the more straightforward implementation, but it lacks granularity where conditions change within a single lot. The model plant’s 4.5 hours and 6,750 pieces assume slicing by time and mapping onto production result lots. This granularity directly determines the containment scope, which makes it the most important item on the list.

Item 4 — feasibility and direction of the three interfaces. Production result lots, equipment operation and changeover, operator attendance and skills. For each, can the counterpart system export the data, in which direction does the data flow, and at what update frequency? If an existing system has no API, the assumption of 350,000 THB per interface in Layer 3 changes. It is worth checking with the vendor of the counterpart system before you request quotes.

Item 5 — validity period of a change and how it is released. Some 4M changes have an expiry. A support worker is in for today only; a provisional machining condition holds until the next scheduled maintenance. Register changes without a validity period and they pile up until nobody can say which changes are currently in force. Decide who enters the release and when.

Item 6 — the format of documents produced for audits. For the forms you actually submit for customer audits and PPAP, what shape will they take after the system goes live? Do you reproduce the existing format exactly, or move to a new one and agree it with the customer? Proceed without deciding and you create a post-go-live routine in which people export from the system and then re-transcribe by hand into the old format.

Item 7 — unit price and response time for change requests after go-live. Adding an approval route, adding a field to a form, changing the layout of a report — all of these will happen after go-live. What does each cost, how many business days is the response, and how many per year are included in maintenance? What the 531,000 THB annual cost actually contains is decided by this item.

Frequently asked questions

What is a 4M change management system?

It is a mechanism for filing, approving and recording changes to manufacturing conditions across the four categories of Man, Machine, Material and Method. However, a system that only digitizes application and approval and a system that goes as far as linking change points to production lots produce different results. The former shortens approval lead time but does not change the containment scope when a defect appears. The latter, on the model plant’s figures, narrows the containment target from 336,000 pieces to 6,750 pieces, a factor of roughly 50.

How much does a 4M change management system cost?

For the model plant (420 employees, 6 lines, 12,000 pieces per day) at full scope, the initial cost is 3,100,000 THB and the annual cost is 531,000 THB. The breakdown is Layer 1 workflow at 850,000 THB (180,000 a year), Layer 2 shop-floor terminals at 780,000 THB (78,000 a year), Layer 3 linkage to production results at 1,050,000 THB (105,000 a year) and Layer 4 embedding at 420,000 THB (168,000 a year). It varies with plant size, line count and the number of interfaces to existing systems, but when comparing quotes, align what sits in each of the four layers before you look at the totals.

What is the difference between 4M change point management and 4M change management?

In practice they cover almost the same ground, but the emphasis differs. Change management is weighted toward the procedure of filing and approving something you are about to change in advance. Change point management is weighted toward detecting and recording what changed, then monitoring quality either side of it. Things that do not suit prior application but still need to be recorded — operator changes currently left outside the notice system, for example — are easier to run under a change point management approach. When you build the system, the realistic design keeps approval-requiring changes and record-only change points in the same ledger, separated as categories.

What is missing from an Excel 4M ledger?

What is missing is not searchability, it is linkage. An Excel ledger can record who changed what and when. But there is no line between a row in the ledger and the production lots that ran that day. So when a defect is suspected, the first affected lot cannot be identified from the ledger and containment gets cut by date instead. At the model plant, the first lot can be identified in 35% of cases today. For the other 65%, someone is estimating after the fact by cross-checking equipment records — the record itself does not hold the answer.

How many months does it take to pay back a 4M change management system?

Roughly 11 months at full scope for the model plant. The annual benefit of 3,953,400 THB less the annual cost of 531,000 THB gives a net of 3,422,400 THB, or 285,200 THB a month, and the initial cost of 3,100,000 THB divided by that is the result. Over five years, against a total cost of 5,755,000 THB the benefit is 19,767,000 THB, for a net of 14,012,000 THB. Note that in the phased option without Layer 3, the initial cost drops to 2,050,000 THB but payback stretches to roughly 20 months.

If operator changes also require a notice, will the floor still be able to function?

If you make every one an approval-bearing notice, no. The model plant has 380 Man changes a year, and adding the 240 Machine, Material and Method changes gives 620 a year. The realistic design separates the weight of approval from the existence of a record. The 150 new assignments and 60 returns after long absence become proper notices with first-piece verification, while the 130 reassignments and 40 support cases become record-only with no approval, linked to a clock-in at the line terminal. That keeps all 380 on record while only 210 of the 380 require approval. On top of that, per-case effort falls from 2.4 hours to 0.6 hours, so total time across all 620 goes from 1,488 hours a year to 372 hours — the overall workload actually decreases.

What is different about implementing in Thailand compared with Japan?

Three things, broadly. The first is language. The Thai line leader who writes the notice, the Japanese manager who approves it and the English-speaking customer who reviews it at audit mean you need the form in Japanese, Thai and English. The second is the distance of approval. Where high-impact changes need approval from head office in Japan, time difference and translation stretch the lead time. You need to narrow the categories that require the Japanese side to the top impact levels and design delegated approval so the rest closes locally. The third is the treatment of support and agency staff, who are often not registered in the attendance system the way permanent employees are and therefore drop out of the records. Issuing every person, agency staff included, an ID that works at the line terminal is the practical starting point.

Summary

The value of a 4M change management system lies not in digitizing an application form but in linking change points to production lots. Here are the figures from this article, lined up again.

• Without linkage, containment is 28 days and 336,000 pieces; with it, 4.5 hours and 6,750 pieces. The gap is roughly 50 times, or 1,344,000 THB against 27,000 THB a year in sorting cost
• The model plant has 620 4M changes a year, of which 380, or 61%, are Man changes. Left outside the notice system, none of them are recorded
• Cost comes in four layers. Initial 3,100,000 THB, annual 531,000 THB. The benefit is 3,953,400 THB a year and payback is roughly 11 months
• Cut the most expensive layer, Layer 3 at 1,050,000 THB, and the five-year net falls from 14,012,000 THB to 4,058,000 THB, a loss of 9,954,000 THB, while payback slips from 11 months to 20 months
• Thailand’s automobile production was 1,455,569 units in 2025, down 0.9% year on year, and the 2026 outlook has been revised down from 1.5 million to 1.45 million units (FTI). The flatter production gets, the more changeovers and staffing reshuffles occur, and the count of 4M changes goes up rather than down

Before you compare quotes, settle these seven items on the plant side — where the notice scope is drawn, approval tiers and delegated approval, the granularity of linkage, the feasibility of the three interfaces, the validity period of a change, the format of audit documents, and the terms for change requests after go-live. Do that and the quotes that come back will be comparable.

TOMAS TECH is based in Bangkok and implements production management, traceability and FA control systems for Japanese-owned manufacturers in Thailand. On 4M change management, we are happy to talk through just two decisions — where to draw the line on notice scope, and how fine the linkage needs to be — or simply whether your existing production management and attendance systems can be interfaced at all. The exploratory stage, before you take any quotes, is entirely fine. Bring your current notice form and your annual counts and we can start from a view of how much needs to sit in each layer. You can reach us here.

References

• JETRO, “Automobile production in 2025 down 0.9% year on year, below 1.5 million units for a second consecutive year (Thailand)”, 5 February 2026 https://www.jetro.go.jp/biznews/2026/02/75b8dcdaa86b9a1c.html
• newsclip, “Thailand’s 2026 automobile production outlook revised down to 1.45 million units, below the previous year’s result”, 27 July 2026 https://newsclip.be/thai-news/thai-economy/40041
• International Automotive Task Force (official IATF 16949 site) https://www.iatfglobaloversight.org/
• Ricoh, “What is a 4M change (4M change point)? How far should changes be filed? Methods and case studies” https://www.ricoh.co.jp/service/digital-manufacturing/media/article/detail49
• Daikin Industries iTEC, “4M change — change management and change point management in quality control” https://www.itec.daikin.co.jp/manufacture/blog/4m/