Process Management System Costs 2026: Thailand & Vietnam Guide

“That order the customer is chasing — which operation is it at right now?” In most plants, answering still takes a phone call and a walk to the shop floor. Output figures are not complete until the next morning’s meeting, and the WIP box has to be hunted down before anyone can confirm it. A process management system is the layer that captures progress and actual output at the order level — a different job from the scheduler that builds the plan and the IoT platform that watches machines. This article breaks down 2026 costs, deployment options and a 90-day path to reliable shop floor data collection for plants in Thailand and Vietnam.

What a process management system is — and how it differs from ERP, MES and APS schedulers

When someone says “we want a process management system,” check first what the words point at. In the same meeting, the admin manager means “something that shows cost and open backlog,” the production control manager means “something that schedules automatically,” and the production manager means “something that shows how far the job has got.” Request quotations before those three are aligned and the comparison sheets will not line up, leaving price as the only criterion.

Factory IT splits into four layers

Separate the layers by the question each one has to answer.

This article deals with the bold row: the progress and actuals layer. The axis is not machine utilization but where the order is right now, where the WIP is physically sitting, and how fast you can commit to a delivery date.

Difference from a production management system (ERP)

A production management system or ERP is the system of record: sales orders, requirements calculation, purchasing, inventory and costing in one stack. These products do include “process management” functions, but the granularity usually stops at start and completion of the manufacturing order as a whole — they will not tell you how many pieces have cleared visual inspection at operation 3. Conversely, a process management system normally carries no costing and no purchasing at all.

Which one you tackle first depends on where the pain is. If backlog does not reconcile with inventory, or you cannot read your product cost, that is an ERP conversation — see our guide to choosing a production management system for a Thai factory.

Difference from a production scheduler (APS)

A scheduler decides how you are going to build; a process management system tells you how the build is actually going. The latter comes first, because a scheduler with no actual data drifts from reality the day after the first Gantt chart is published, and within a month nobody opens it. If you are evaluating planning tools, read our production scheduler (APS) comparison first and separate whether you need planning or feedback.

The field test: can somebody state today’s actual output as a percentage of plan from a screen, without phoning anyone? If not, a scheduler gives you a planning engine with no input source.

Difference from MES

MES (Manufacturing Execution System) is the heavyweight category: progress and actuals plus quality records, traceability, equipment integration and material lot allocation. A process management system is best understood as the progress-and-actuals slice of MES, packaged light.

The MES market keeps growing — a CAGR of 11.7% is forecast for 2026-2033, expanding toward the USD 42.1 billion range, and 8.2% or higher for 2026-2031 within Japan. Adoption is uneven, though: 56.1% of large enterprises use MES versus 21.4% of SMEs, a gap of 34.7 percentage points, or roughly a factor of 2.6. Overseas plants of Japanese groups usually sit closer to the SME end in headcount, and the common outcome is “MES is premature for us” — after which progress and actuals quietly stay on paper.

Difference from electronic forms and IoT machine monitoring

These two are the easiest to confuse with process management.

• Electronic forms (e-forms) move a paper format onto a tablet as-is. Strong where the form itself is the point: check sheets, inspection records, quality documentation. See going paperless with electronic forms.
• IoT machine monitoring collects equipment signals to show utilization, minor stoppages and OEE. Here the machine is the point. See implementing IoT machine monitoring in a factory.
• A process management system puts the order and the lot at the centre. All three look like an operator tapping a tablet, but the aggregation axis differs — and that is decisive.

Because the implementations overlap, collecting operation actuals on the e-forms platform is a realistic architecture. But if you do, the order number and operation code must be mandatory fields. Skip them and the data cannot later be re-aggregated by order, which means rebuilding from scratch.

Why production progress tracking became urgent for Thai and Vietnamese plants in 2026

Projects that sat in the “someday” pile are moving in 2026, and the reason can be stated in numbers.

Demand has recovered — Thailand PMI 53.6, Vietnam PMI 51.8

The S&P Global Thailand Manufacturing PMI stood at 53.6 in June 2026, up from 52.6 in May; anything above 50 signals expansion. Vietnam’s Manufacturing PMI was 51.8 in June 2026 — down from 52.8 in May, but still above 50, with output up for 14 consecutive months.

More orders means more WIP, and when WIP rises, paper-and-Excel control degrades not linearly but exponentially. In a plant running 20 lots, “we cannot find it” is a once-a-day nuisance; at 40 lots the search time and status enquiries more than double, and you run out of staging space as well.

Vietnam’s output is climbing — H1 industrial production +10.8% year on year

Vietnam’s Index of Industrial Production (IIP) for the first half of 2026 rose 10.8% year on year, the strongest reading since 2019. Industrial value added grew 9.86% and value added in manufacturing and processing grew 10.23%. For groups with sites in both countries, it can now be rational to stand the platform up in Vietnam first, riding the volume increase, then roll the template across to Thailand.

Labor cost has risen — the Bangkok minimum wage is THB 400 per day

Thailand’s minimum wage rose from THB 372 to THB 400 per day across all industries in Bangkok on 1 July 2025 — up THB 28, or roughly 7.5%, affecting an estimated 700,000 Thai workers. Four provinces and one district including Chonburi and Rayong had already moved to THB 400 in January 2025. (Currency notes: USD figures use USD 1 ≒ THB 32.5, JPY figures use THB 1 ≒ JPY 4.5. THB 400 per day is roughly USD 12.3; THB 372 was roughly USD 11.4.)

What matters for the investment case is not the direct labor line but that the raise lifts the floor under indirect hourly rates. Transcribing, reconciling and chasing status are hours in which nothing is produced, and paying people to absorb that work gets more expensive every year. Read the other way, the rate used to convert saved hours into money has also gone up — so the same hours saved make a stronger payback case than a few years ago.

Regulatory pressure — ISO 9001:2026

ISO 9001:2026 is scheduled for publication around September 2026, with JIS Q 9001 following in December 2026. The documentation requirement is reorganized around “documented information shall be available as evidence,” and the weight given to effectiveness, evidence and continual improvement increases. Stacking paper daily reports in a storeroom and being able to produce evidence on demand are two different things. (Record design is covered below.)

Five symptoms that process management has outgrown Excel and paper

Before judging the investment, identify which symptom you have. These patterns recur across plants in the region.

Symptom 1: progress is only knowable by phone

Sales asks “what is happening with that order?”, production control phones the shop floor, the line leader goes and looks, then calls back. Five to ten minutes per enquiry; ten a day and an hour has gone. The real problem is not the hour but that the answer arrives 30 minutes to half a day later, with the customer waiting throughout.

Symptom 2: actuals only arrive the next day

Daily reports are handwritten, collected in the late afternoon and typed into Excel the following morning. In that structure the earliest you can notice a delay is the next morning, and a delay on night shift cannot be acted on until the day after. Compressing this detection lag is, in practice, the single largest benefit of a process management system.

Symptom 3: nobody knows where the WIP is

Buffer areas between operations, the queue in front of inspection, the temporary rack for rework. The only way to know which lot is in the box on the trolley is to read the handwritten tag. “Every stock count throws up variances and it takes two full days to trace them” is something we hear in Thailand and Vietnam alike.

Symptom 4: you cannot commit to a delivery date

The customer asks “an extra 10,000 pieces — when can you ship?” and the answer is “let me come back to you tomorrow,” because the current WIP load is not expressed in numbers anywhere. Delivery date management looks like a story about hitting dates; in reality it is about how quickly and how reliably you can answer.

Symptom 5: the production daily report is entered twice or three times

The operator writes on paper, admin types it into Excel, and at month end someone keys a summary into the ERP — the same figure entered three times. The more entries, the less clear which version is authoritative, and a brand-new job appears: reconciling “the Excel number does not match the ERP number.”

Self-diagnostic checklist

Score one point for each statement that applies (maximum 10).

How to read your score

Buying software at a score of 9 or above will fail. In a plant where the physical goods carry no identifiers, no system can be fed with data.

The process management system feature map: from work orders to delivery date management

The progress-and-actuals layer breaks into five blocks. Information flows left to right.

1. Work order system — digitizing work instructions

Explode the manufacturing order into operations and dispatch what each team or machine has to build, when and how many.

• Replace the paper work order with a tablet display, so re-sequencing and priority changes take effect immediately
• Attach drawings, work instructions and cautions so they display in context
• Record the start time automatically when the operator taps “start”

The classic mistake is reproducing the paper work order’s layout on screen. Paper conveys information through *everything visible at once*; a screen conveys it through *filtering*. This calls for redesign, not migration.

2. Shop floor data collection — capturing actuals and digitizing the production daily report

Record good quantity, scrap quantity, operator, start and end times and stoppage reason where they occur. The choice of input method drives cost and adoption more than any other single decision.

Recommended combination: quantities by scan or machine signal, reason codes by button on a tablet. Automating everything sends retrofit costs through the roof; keying everything by hand means nobody enters anything after three months.

3. Production progress visualization — real-time progress on screen

Three screens are the minimum viable set.

• Order progress list: one row per order number, a horizontal bar showing which operation is complete, delay shown by colour
• WIP by operation: waiting quantity and dwell time at each operation — where the bottleneck becomes visible
• Today’s summary: planned quantity, actual quantity, achievement rate and scrap rate by time block, displayed permanently on a large shop floor monitor

What goes on the shop floor monitor must be readable in three seconds. Design the management dashboard and the shop floor display as two different artefacts.

4. WIP tracking and material control

The “put a number on the physical goods” layer — the foundation of the whole system, and the part most often skimped on.

• Lot tags: print the lot number as a QR code and attach it to the box, trolley or pallet
• Location management: assign addresses to buffer areas and record where it was put with a scan
• Status control: waiting / in process / awaiting inspection / passed / in rework / ready to ship

Do not make locations granular on day one. Area level — Buffer A, Buffer B, inspection queue rack — is enough to start. Manage down to individual shelves and the scans per put-away grow until the shop floor stops complying.

5. Delivery date management and delay countermeasures

Only once progress and WIP are in place can you have a real conversation about dates.

• Calculate the expected completion date from the sum of standard times for remaining operations plus current load by operation
• Alert on any order whose slack against the due date drops below a threshold
• For orders that slipped, record which operation it sat at and for how many hours, and feed that into prevention

The biggest lever is not forecast accuracy but how many hours earlier the delay is detected. Three days of warning can be absorbed with overtime or a re-sequence; finding out on the day leaves only expedited freight.

Priority of features

Do not implement all of this at once. The order that works is:

Breaking process management system costs into five layers

Quotations cannot be compared side by side because each vendor loads different layers at a different level of detail. Always decompose into the following five layers and fill the gaps before comparing.

Assumptions for the model

The figures below are a model case. They represent a general range; actual amounts vary with requirements, site size and vendor.

• Exchange rates: THB 1 ≒ JPY 4.5 and USD 1 ≒ THB 32.5. Every conversion in this article uses these rates.
• Scenario A: about 50 employees, a pilot limited to one line. Work instructions and actuals capture only, no ERP interface.
• Scenario B: about 150 employees, a standard rollout across three lines, including progress visualization, WIP tracking and an ERP interface.

As a reference point from the Japanese market, cloud products are quoted from the tens of thousands of yen for setup and per month, on-premise from the millions of yen for setup with monthly maintenance in the tens of thousands. For production management systems, examples start at JPY 1.5 million for companies with 100 employees or fewer and JPY 5 million for large rollouts including overseas sites. The baht figures below are totals including local devices and local labor.

Five-layer cost table (first year, THB)

Converted at the rates stated above:

Recurring cost from year two

What gets missed in each layer

Layer 1, licensing. Confirm the metering unit: per named user, per device or per site. Per-user pricing is usually unfavourable in a shift-based plant; if three shifts share one terminal, per-device pricing is cheaper.

Layer 2, setup and master data. The most under-estimated line by a wide margin. Standard times per operation either do not exist or were last updated ten years ago. Building 500 part numbers x 5 operations = 2,500 rows of standard time is your work, not the vendor’s. A suspiciously cheap setup quotation usually means that homework has been pushed back to the customer.

Layer 3, shop floor devices. The forgotten item is the wireless environment. In a hall full of metal racking and machines, placing APs with office-building instincts guarantees dead zones — insist a physical site survey is in the setup scope. Tablets that are not dust-, splash- and drop-resistant also need replacing within a year.

Layer 4, integration. Manual CSV import is sufficient in year one. Building the automated interface after the actuals data has stabilized means far less rework — hence layer 4 is zero in Scenario A.

Layer 5, support and training. In Thailand and Vietnam, budget training not as the cost of one event but as an annual retraining cost, for reasons in the regional section below.

Building the ROI case — labor savings alone will not pay for it

Start with the hours saved. Assume a fully loaded indirect labor cost (including social security) of THB 24,000 per month (approx. USD 738) over 160 working hours, giving an hourly rate of THB 150 (approx. USD 4.6). The Bangkok minimum wage is THB 400 per day, but production control and administrative staff sit above that level, hence this rate.

Annually that is 24,600 x 12 = THB 295,200 (approx. USD 9,083 / JPY 1,328,400).

Now the honest part. Scenario B’s recurring cost is THB 360,000 per year, so labor savings alone leave you THB 64,800 (approx. USD 1,994) in the red every year. The pitch that “digitizing the daily report pays for itself in headcount” does not survive a plant of this size.

The investment stands up when it removes the hard cash cost of late delivery. Suppose detecting delays three days earlier cuts expedited air shipments from twice a month to half a time a month, at THB 40,000 (approx. USD 1,231) per shipment.

The payback profile looks like this. Year one counts only half of the annual benefit, to allow for the ramp-up period.

The cumulative position turns positive about two years and seven months after go-live. From year two onward the net benefit runs at THB 655,200 (approx. USD 20,160 / JPY 2,948,400) per year.

So the first task in an evaluation is not comparing features but measuring your own hard cash cost of late delivery: the last 12 months of expedited freight, weekend overtime premiums, subcontracting fees paid to cover shortfalls and customer penalties. If that number is small, this is not an urgent investment for you.

Do not calculate ROI on the pilot

The same arithmetic on Scenario A gives 47 hours saved per month on one line (the 24 hours of monthly consolidation and head-office reporting do not shrink with line count, so 164 – 24 = 140 hours is spread across three lines) x THB 150 = THB 7,050 per month, or THB 84,600 per year (approx. USD 2,603), below the THB 174,000 running cost. The pilot does not pay for itself in cash terms.

That is not a failure, because the purpose of a pilot is not financial return. Its three deliverables determine how accurate the full rollout quotation will be:

1. Measured entry times, validating the assumptions in the quotation
2. Actual effort consumed building master data, which sharpens layer 2 dramatically
3. How long adoption took, and which operating rules were abandoned

Start “let’s just try one line” without agreeing this framing with management, and three months later somebody will conclude “it hasn’t delivered anything” and stop it.

Comparing deployment options: cloud, on-premise, custom build, MES

How to choose between them

Cloud package. Five to ten operations, conventional process technology, one or fewer IT staff on site, a target of going live in three months, and an eye on replicating to other sites later.

On-premise package. The ERP already sits on an internal server and actual data has to move both ways frequently. (Resilience against network outages is not the argument it used to be — several cloud products handle this with local caching.)

Custom build. The reasons to choose this shrink every year. Reserve it for a process technology nobody else has that is genuinely a source of competitive advantage. The main risk is not cost but that a transfer or resignation leaves nobody able to touch it — the large-scale version of a locally written Excel VBA tool becoming a black box the day its author leaves.

MES. Automotive, medical devices, electronic components — cases where complete lot-level traceability is mandated by the customer or by regulation. Otherwise MES is too heavy.

How to test a product that calls itself a “process management system”

The question list for the demo. A vendor who cannot answer these on the spot probably has limited implementation experience in the progress-and-actuals layer.

1. When one order splits across parallel operations and merges again, how is percent complete calculated?
2. When rework sends a lot back to an earlier operation, how is the actual quantity treated?
3. When a lot is split or merged, is traceability preserved?
4. When an operator forgets to press “start,” can the timestamp be corrected afterwards — and is that correction logged?
5. If the network drops for 30 minutes, what happens to entries on the shop floor terminals?
6. Is the UI language switched per user or per device?
7. Can we maintain part number and routing master data ourselves, or is every change a vendor job?

The situations behind questions 4 and 5 will occur in your plant, guaranteed. A product with no answer to them loses data credibility within three months of go-live.

A 90-day roadmap to get shop floor data collection running

Narrowed to actuals capture only, 90 days is achievable. The following assumes the three-line scale of Scenario B.

That is 45 internal plus 42 vendor person-days, or 87 person-days in total. The internal 45 is roughly a production control manager giving two days a week for three months. If you cannot secure that, move the start date. A process management system run in someone’s spare time always stalls on incomplete master data.

What Phase 1 (days 0-30) covers

• Draw the process on paper. Walk the floor and draw how material actually moves — not the org chart, not the layout drawing, but the flow of material. This always uncovers two or three operations that exist in reality but on no diagram: rework, re-inspection, outside processing.
• Decide the numbering rule. Digit structure of the lot number, when it is issued, who attaches it. The principle: issue the number when the manufacturing order is released, attach the tag at the first operation.
• Measure the cash cost of delays. The numerator of the ROI case: 12 months of expedited freight, weekend overtime and subcontracting.
• Pick one pilot line. Not the busiest, not the quietest, but a line with a middling number of operations and a cooperative line leader.

What Phase 2 (days 31-60) covers

• Run paper and system in parallel. Retiring paper immediately leaves you with nothing when the system goes down.
• Measure entry compliance daily. Untracked, gaps appear from week two. Target: 95% or higher for seven consecutive days.
• Survey the wireless network physically. Not on a drawing — walk the floor with a device and plot the dead spots.
• Time the data entry. Ten measurements with a stopwatch; if one entry takes over 30 seconds, cut input fields.

What Phase 3 (days 61-90) covers

• Set and announce the paper retirement date in advance. Parallel running is comfortable, so left alone it continues forever.
• Establish that the previous day’s actuals are confirmed by 09:00. The practical goal of the whole implementation.
• Document the operating rules: who enters what and when, and who is contacted when something is abnormal. This carries straight over into ISO 9001:2026 readiness.

What comes after day 90

From day 91, decide once you have data. Most plants follow this order:

1. Revise standard times, recalculated from actuals. This is the point at which planning accuracy finally improves.
2. Analyse WIP and dwell time by operation to identify the bottleneck operation.
3. Automate the ERP interface, once manual CSV operation has stabilized.
4. Add automatic calculation of expected completion dates.
5. If justified, extend into scheduling or IoT machine monitoring.

Five failure patterns and how to avoid them

Pattern 1: starting before the physical goods carry numbers

Symptom. The labelling routine never takes hold, so operators key order numbers by hand and the data fills with typos.

Avoidance. Start issuing and attaching lot tags before the system project, on paper if necessary. Run one month with every lot carrying a number, even handwritten. A plant that cannot sustain that will not sustain it with software either.

Pattern 2: too many input fields

Symptom. “While we’re at it” produces 30 defect reasons and 20 stoppage reasons, and the shop floor selects “Other” for everything.

Avoidance. Start with no more than six options each — a number that fits on one screen. After three months, interview the floor about what is hiding inside “Other” and promote only the frequent ones. It is a sequencing issue: cutting options later is easy; adding them and then cutting them back is politically hard.

Pattern 3: no baseline measurement

Symptom. After go-live somebody asks “so how much better is it?” and nobody can answer.

Avoidance. Measure the before values in Phase 1 — at minimum these five:

Pattern 4: forcing the head office format onto the shop floor

Symptom. Carrying every field used by the head office system leaves the design misaligned with how the local plant actually runs.

Avoidance. Separate the data entered locally from the data reported upward. Design local entry around shop floor reality and generate the report by aggregating from it. Without that separation, operators are made to enter data “for head office” and the routine becomes an empty ritual.

Pattern 5: leaving everything to the vendor and retaining no master data capability

Symptom. Every new product requires a vendor request, which means extra fees and waiting time.

Avoidance. Confirm before signing whether part number and routing master data can be maintained in-house, and make Phase 3 completion conditional on your own staff having registered records themselves. Receiving a manual does not transfer capability.

Issues specific to sites in Thailand and Vietnam

Multilingual UI — can three languages coexist?

At a typical plant here, operators work in Thai or Vietnamese, supervisors mix in English, and the expatriate management team works in Japanese or English — three layers at once. Points to verify:

• Can the language be switched per user? Per-device switching breaks down under shift rotation.
• Can master data names be held in multiple languages? If part and operation names are fixed independently of the display language, operators still cannot read them.
• Are error messages translated too? Plenty of products stop at that boundary.
• Thai and Vietnamese strings are wider than Japanese, so check on real hardware that button labels do not wrap and break the layout.

Turnover — treat training as an annual cost

Training is never a one-off here: every new hire and every team reshuffle triggers retraining. Build this into the design from the outset:

• Simplify operation until it is obvious by looking. Anything requiring a manual will not stick.
• Post an A4 one-pager beside each terminal, in the local language with photographs.
• Appoint two key users per line. With only one, the knowledge leaves when they do.
• Produce training content as video so it can be reused indefinitely.

The cost table books training in year one, but budget on the assumption that some level recurs every year.

Check BOI and DEPA investment incentives

Thailand’s BOI offers SMEs incentives including exemption from import duty on machinery and corporate income tax exemption of up to 200% of the investment amount (excluding land and working capital). The minimum SME investment is THB 500,000 (approx. USD 15,385). DEPA, the Digital Economy Promotion Agency, separately supports digital-sector activity including software development.

Scenario A here is THB 535,000 and Scenario B is THB 1,551,000, so both exceed the THB 500,000 level in headline terms. Whether the incentives apply depends on the business category, the application class and how eligible investment is defined — including whether software cost counts toward the investment amount, which varies case by case. Confirm with BOI/DEPA or your advisor before requesting quotations, and compare totals assuming whichever incentive applies. It changes how easily the capital request gets approved.

Record design with ISO 9001:2026 in mind

In ISO 9001:2026, due around September 2026, the documentation requirement is reorganized around information being “available as evidence,” and effectiveness and continual improvement carry more weight. In design terms that becomes:

Whether a change history is kept when actuals are corrected after the fact comes up surprisingly often in audits, and is frequently missing from lower-cost products. That is question 4 on the demo list above.

Designing the report to the Japanese head office

Head office usually wants monthly production quantity, utilization, scrap rate and on-time delivery rate. The key points are to separate local entry design from the head office reporting format and to agree the definitions in writing first.

If “scrap rate” means scrap ÷ input at head office and scrap ÷ good output locally, the two numbers are simply different things. Because a process management system starts producing numbers, the mismatch becomes visible and turns into an argument shortly after go-live. Get production quantity, scrap rate, utilization and on-time delivery rate defined in writing during Phase 1.

Frequently asked questions (FAQ)

What is a process management system?

It records where a manufacturing order or lot currently sits and how many pieces are complete, and makes that visible as progress. Five blocks: work instructions, shop floor data collection, progress visualization, WIP and material control, delivery date management. It differs from a production management system (orders and cost), a scheduler (plans) and IoT machine monitoring (equipment) in the unit of information it handles.

What is the difference between a process management system and a production management system?

A production management system is the system of record for sales orders, requirements calculation, purchasing, inventory and costing; its units are orders, part numbers and money. A process management system works in manufacturing orders, lots and operations, at a granularity that answers “how many pieces have cleared operation 3.” Production management systems include process management functions, but usually only to the level of starting and completing the manufacturing order. If you need both, sequence by where the pain is: ERP first if backlog and cost do not reconcile, process management first if progress and delivery dates are unreadable.

How much does a process management system cost?

Japanese market benchmarks run from the tens of thousands of yen for cloud setup and monthly fees, and from the millions of yen for on-premise setup. In this article’s model case (THB 1 ≒ JPY 4.5, USD 1 ≒ THB 32.5), year one is THB 535,000 (approx. USD 16,462 / JPY 2,407,500) for a one-line pilot and THB 1,551,000 (approx. USD 47,723 / JPY 6,979,500) for a three-line rollout. Recurring cost from year two is THB 174,000 (approx. USD 5,354 / JPY 783,000) and THB 360,000 (approx. USD 11,077 / JPY 1,620,000) per year respectively. All indicative, and variable with requirements, site size and vendor.

Can we start by digitizing the production daily report?

Yes — the most realistic starting point. But the order number and the operation code must be included as input fields. Digitize the report purely as “the quantity produced that day” and it cannot later be re-aggregated into order-level progress, so the work has to be redone. Make the digital daily report and the operation actuals capture a single entry, and you remove double entry and gain progress visibility at once.

How far should WIP tracking go?

Area-level locations plus lot labels are enough to begin with. Assign addresses to buffer areas, attach a QR code to each lot, scan on put-away. Managing down to individual shelves or trolleys multiplies the scans per put-away and the shop floor stops complying. Refine later, once three months of operation produce a concrete complaint that area level takes too long to search.

How long before delivery delay countermeasures show results?

The earlier-detection benefit appears as soon as data entry is embedded, roughly 60-90 days in. Having the previous day’s actuals confirmed by 09:00 lets you catch on the same day delays that used to surface several days later, so they can be absorbed with overtime or a re-sequence. Predictive functions such as automatic completion-date calculation are realistic only after three to six months of data, because standard times must be revised from actuals first. Financial payback, in the Scenario B model case, works out at about two years and seven months.

How do we know when we have hit the limits of Excel production tracking?

Three tests. Is the same figure entered two or more times? Has the number of people who can open the file effectively dropped to one? Is work spent confirming which version is the latest? If any apply, the management cost of Excel is already approaching the running cost of a system. Score 6 or above on the self-diagnostic checklist and it will break when volumes rise.

Does a small or mid-sized plant need MES?

MES adoption is 56.1% among large enterprises versus 21.4% among SMEs — the smaller the company, the less it has penetrated. If traceability is not mandated by a customer or by regulation, MES is too heavy. Cover the progress-and-actuals layer with a process management system first, and extend to MES when quality records genuinely become a requirement.

Conclusion

A process management system is not a tool for building plans or watching machines. It is the tool for knowing where the order is, where the WIP is, and how soon you can commit to a date. The 2026 demand environment supports acting now: the Thai manufacturing PMI rose to 53.6 in June from 52.6 in May, Vietnam held above 50 at 51.8 in June with output up for 14 consecutive months, and Vietnam’s first-half IIP was up 10.8% year on year. Meanwhile the Bangkok minimum wage rose about 7.5% to THB 400 per day from July 2025, so absorbing indirect work with people keeps getting more expensive.

Three points drive the decision. First, use the four-layer model to identify which layer your problem lives in. Second, decompose cost into five layers and close the gaps, especially setup, master data and the wireless environment. Third, do not put labor savings alone in the numerator of the ROI: in the model case they do not cover the annual running cost, and only with the hard cash cost of late delivery included does the case pay back in about two years and seven months. Which is why the first task is not product comparison but measuring the last 12 months of delay-related costs.

On execution, narrowed to shop floor data collection, 90 days is enough: days 0-30 map the process and put numbers on the goods, days 31-60 run one line in parallel at 95% entry compliance, days 61-90 retire paper and lock in the previous day’s actuals by 09:00. Whether you can commit 45 internal plus 42 vendor person-days — 87 in total — is the real condition that sets your start date.

TOMAS TECH is happy to talk even while you are still deciding which operation to tackle first, or whether the lot tag routine should come before the software at all. We are based in Bangkok supporting Japanese-affiliated plants across Thailand and ASEAN, and our factory IT experience — including the PEGASUS production and energy management system — lets us discuss a practical sequence that accounts for local wage levels, multilingual operation and BOI/DEPA assumptions. Bring your checklist score and the last 12 months of delay-related costs, and the first conversation can go straight to specifics. Get in touch here.

References

• Thailand Manufacturing PMI (June 2026 = 53.6, May = 52.6): https://www.fx.co/en/forex-news/3046093
• S&P Global PMI: https://www.pmi.spglobal.com/
• Vietnam Manufacturing PMI (June 2026 = 51.8, May = 52.8): https://english.vov.vn/en/economy/vietnam-manufacturing-pmi-stands-at-518-in-june-signaling-solid-start-to-h2-post1311352.vov
• Vietnam H1 2026 Index of Industrial Production, +10.8% year on year: https://english.news.cn/20260703/2caf0183320e4753b36cee7cb0f792dc/c.html
• Thailand minimum wage increase (1 July 2025, Bangkok THB 372 to 400 per day): https://www.jetro.go.jp/biznews/2025/07/b21007a1ac8f7fca.html
• MES market, CAGR 11.7% for 2026-2033: https://www.gii.co.jp/report/grvi2040618-manufacturing-execution-systems-market-size-share.html
• MES market expanding toward USD 42.1 billion: https://seisan-navi.com/20260501-mes421/
• Japan MES market, CAGR of 8.2% or higher for 2026-2031: https://www.atpress.ne.jp/news/2654316
• MES adoption rate (large enterprises 56.1% / SMEs 21.4%): https://www.cct-inc.co.jp/koto-online/archives/28
• Cost benchmarks for process management systems: https://hnavi.co.jp/knowledge/blog/process-management-cost/
• ISO 9001:2026 key changes (BSI): https://www.bsigroup.com/ja-JP/products-and-services/standards-services/iso-9001-2026-key-changes-and-guidance/
• ISO 9001:2026 transition guidance (SGS): https://www.sgs.com/ja-jp/showcases/iso-9001-2026-key-updates-and-transition-guidance
• Investment incentives in Thailand (BOI / DEPA, JETRO): https://www.jetro.go.jp/world/asia/th/invest_03.html
• BOI investment incentives for the digital industry: https://www.boi.go.th/upload/content/Investment%20incentives%20for%20Digital%20Industry%20(BOI).pdf