Production Management System Comparison 2026 – Costs and How to Choose Without Failing

You already know that Excel spreadsheets and paper daily reports have hit their ceiling. But the moment you start a production management system comparison, every vendor uses different terminology and a different way of quoting price, and it becomes unclear what you are even supposed to line up side by side. This article is written for plant managers, administration heads and IT staff at Japanese-affiliated manufacturers with sites in Thailand and ASEAN. It covers how to compare by system type rather than by product name, how to think about cost and payback, why implementations fail, and the issues that only appear at an overseas plant. The goal is not to rank specific products, but to help you build your own decision criteria.

What a production management system actually manages

Before comparing anything, you need a shared definition of what this system is supposed to take over. If that scope stays vague and you invite several vendors in, each proposal will cover a different footprint, and you end up in the sterile exercise of comparing prices alone.

The four domains a production management system covers

What is generally called a production management system handles the following four domains.

The first is production planning. Based on orders and demand forecasts, it decides what to make, when, and how much. This is where the master production schedule feeds material requirements planning (MRP), which is then broken down into manufacturing orders and purchase requisitions. In an Excel-run factory, building this plan and reflecting the changes that come in every day is the task most prone to becoming person-dependent.

The second is purchasing and inventory. This covers ordering based on requirements, goods receipt and inspection, and visibility into both the quantity and valuation of raw materials, work in process and finished goods. Inventory and shop floor progress are two sides of the same coin. In a plant where progress is invisible, you cannot see inventory accurately either.

The third is shop floor progress and production results. This records which operation a manufacturing order has reached, along with output quantity, defect quantity and labor hours. Digitizing daily reports and operator worksheets belongs here.

The fourth is costing. It compares standard cost against actual cost and shows profitability by part number and by order. Because the numbers here are only as trustworthy as the data quality in the previous three domains, in practice costing tends to be the last area to be built out.

The difference between ERP, a production management system and MES

The single most confusing point in any evaluation is where ERP ends, where the production management system begins, and what MES is for. If you organize the explanations published by firms such as Unifinity and Layers Consulting, the cleanest way to understand the positioning is as a hierarchy.

ERP manages enterprise resources as a whole and supplies the information used for management decisions. It sits at the corporate business layer. Its main purpose is to unify company-wide core data across accounting, sales, purchasing and HR, and manufacturing is just one of many modules.

The production management system handles planning and control, and sits at the planning and control layer. It plans by day and by order, then compares plan against actual and manages the variance.

MES (Manufacturing Execution System) is the shop floor execution layer. It distributes work instructions, tracks progress in real time, collects quality data and monitors equipment operation, with information refreshed by the minute or by the hour. MES sits between the planning layer (ERP) and the control layer (SCADA, DCS and PLC). Its role is to translate upstream plans into something the shop floor can execute, and to feed what actually happened back upstream.

The important point is that there is no strict boundary between the three. Coverage overlaps depending on the product. Some ERP manufacturing modules reach well into production management territory, while some products that call themselves production management systems have strong MES-style data collection. If you move forward without clarifying how far ERP covers and what MES is responsible for, you get either duplicated investment from overlapping functions, or the reverse, an integration gap where a task falls through and nobody owns it. The first job in any comparison is not lining up products. It is deciding internally where to draw that line of responsibility.

Where does a production scheduler (APS) fit?

The other commonly confused category is the production scheduler, or APS. Where MRP inside a production management system calculates how much is needed and by when, an APS goes further and assigns work along a time axis, deciding which machine runs which job from what hour to what hour, under finite constraints such as machine capacity, changeover and manpower. In high-mix low-volume plants with frequent changeovers, or plants with a clearly identified bottleneck operation, a production management system alone cannot guarantee that the plan is executable, and running an APS alongside it becomes a reasonable decision. Our article on production scheduler comparison (APS) digs into that division of labor, including planning granularity and how constraints are handled. If planning is your bottleneck, reading that first will make the comparison criteria in this article far more concrete to apply.

Comparing the four types of production management system

If you begin your comparison with product names, you end up counting circles in a feature matrix. In practice it is faster and more stable to narrow candidates down by type first. Here we compare four types.

Cloud (SaaS)

You use an environment prepared by the vendor over the internet. The biggest advantage is that no server procurement is required, so initial cost stays low and you can go live in a short time. Version upgrades are handled by the vendor, which suits sites that cannot keep maintenance staff in house.

The trade-off is a strong assumption that you fit your operations to standard functionality, so building heavily customized workflows is generally difficult. If you plan to use it at a plant in Thailand, you also need to verify your internet connectivity and which country hosts the servers holding your data before you sign. Because billing is monthly, cost grows linearly as user counts or the number of sites increases, and that characteristic matters when you look at long-term total cost of ownership.

On-premise package

You install a purchased software package on your own servers or on cloud infrastructure you contract yourself. The advantages are a high degree of customization freedom and easier integration design with existing accounting systems and equipment. Licenses are often bought outright, so the cost per year falls the longer you use it.

On the other hand, the initial investment is large, and infrastructure operations such as server and OS refreshes and security patching stay on your side. There is also the risk that after piling up customizations you can no longer upgrade, leaving the system effectively frozen in place. Customization is almost always quoted separately from the base product price, so in any comparison it is essential to make vendors present two figures separately: the price as standard, and the price in the state that satisfies your requirements.

Scratch development

You build from zero to match your own operations. When your production model is unusual, or when a proprietary management method that competitors cannot copy is itself your competitive advantage, scratch development of a production management system can be the rational choice. Because you do not have to change existing operations, resistance from the shop floor is also smaller.

That said, cost and duration expand significantly. You should also be prepared for continuous maintenance and enhancement after go-live, and for a structure that depends on the specific people or vendor who understand the specification. The judgment “20 percent of our requirements do not fit the package, so we will go scratch” is also a decision to build the other 80 percent yourself for the sake of that 20 percent, and then maintain all of it forever. The work of separating whether that non-fitting 20 percent is truly a source of competitive advantage or merely inherited habit should come first.

Industry-specific and mid-range packages

These are packages built around the commercial practices and process structures of a specific industry, such as metal machining, plastic molding, food, or electronic component assembly. Because your own operational vocabulary already works in the standard functionality, fit rates are high and the volume of customization stays low. For mid-sized and smaller plants, using an industry-specific package as is often comes out lower in both total cost and duration than heavily modifying a general-purpose package.

The weakness is that once you have an operation that falls outside that industry’s norm, the system suddenly cannot handle it, and the range of available products is itself limited. If the vendor is small, you also need to confirm whether a realistic support structure can be secured for your Thailand site.

Comparison table for the four types

You will find plenty of “best production management system” ranking articles online, but those rankings simply reflect the criteria of the publisher. Production management systems are the kind of software where the same product is a huge success at one plant and abandoned at another. Rather than a ranking, we recommend narrowing candidates by the fit between your production model and the system type.

Cost benchmarks for a production management system

The cost of a production management system needs to be understood in three layers: initial cost, running cost, and the hidden costs that never appear on a quotation. When competing quotes come back two or three times apart, most of that gap comes from differences in how these three layers are divided up.

Cost benchmarks by type

These ranges vary by source. OSK puts cloud subscriptions at JPY 30,000 to 150,000 per month and annual maintenance for packages at 5 to 15 percent of implementation cost, while summaries from SFA JOURNAL and BOXIL give cloud subscriptions of JPY 30,000 to 100,000 per month, initial cost from free to JPY 500,000, and annual package maintenance of 10 to 20 percent of implementation cost. Because the sources disagree on annual maintenance, quoting 5 to 15 percent versus 10 to 20 percent, this article does not assert either one and instead uses the expression roughly 10 to 20 percent of implementation cost. For budgeting purposes it is safer to plan at the upper end, meaning 20 percent. Note that the ROI calculation later in this article adopts 15 percent, the upper bound of the 10 to 15 percent band that both sources include.

As a sense of scale, Prevision concludes that for make-to-order small and mid-sized manufacturers with 10 to 200 employees, JPY 1 million to 5 million is the realistic line for an on-premise package. Put the other way around, if a proposal at that scale comes back in the tens of millions of yen, you have grounds to suspect the customization scope has become excessive.

The hidden costs that never appear on a quotation

The three most commonly overlooked costs in any comparison are the following.

On top of these, your own internal labor is a real cost. The time your shop floor key people spend on requirements definition and acceptance testing never shows up on a quotation, but in practice it reaches a scale you cannot ignore. If you choose a system “because it is cheap” without factoring this in, normal operations stop running and the project stalls.

If you want to start with shop floor progress visibility alone, our article on process management system cost and selection breaks down the relationship between functional scope and price in finer detail for a narrowed scope. If you are considering a phased approach that starts with one process before pursuing global optimization, reading it alongside this one will make budget allocation easier to judge.

How to think about return on investment (ROI)

Just as important as “how much will it cost” is “how many years until it pays back”. Below is one worked calculation with every assumption stated. The figures are a model case and do not guarantee actual results. Use it as a template you can recalculate with your own numbers.

Assumptions for the calculation

Throughout this article the exchange rate is standardized at 1 THB = 4.5 JPY.

Calculating annual savings

1. Reduction in indirect labor hours

The work hours in scope are 8 people x 1.5 hours x 250 days = 3,000 hours per year. Cutting that by 60 percent gives 3,000 hours x 60 percent = 1,800 hours per year freed up. At an hourly rate of THB 200, which is JPY 900, that is 1,800 hours x JPY 900 = JPY 1.62 million per year in savings.

2. Reduction in inventory holding cost

Compressing average inventory of JPY 12 million by 15 percent gives JPY 12 million x 15 percent = JPY 1.8 million less inventory. However, that JPY 1.8 million is a one-time cash flow improvement, not a recurring effect on the profit and loss statement. What hits the P&L is the holding cost portion: JPY 1.8 million x 10 percent per year = JPY 180,000 per year. Confusing these two and booking “JPY 1.8 million of benefit every year” is the most common form of inflation in ROI calculations.

3. Reduction in expedited response cost

Expedited freight and extra shipments run at 12 times x JPY 50,000 = JPY 600,000 per year. If progress visibility cuts that by 50 percent, the saving is JPY 600,000 x 50 percent = JPY 300,000 per year.

Total annual savings come to JPY 1.62 million + JPY 180,000 + JPY 300,000 = JPY 2.1 million.

Payback period

Because annual maintenance costs JPY 600,000, the net annual benefit is JPY 2.1 million – JPY 600,000 = JPY 1.5 million. Dividing the initial investment of JPY 4 million by that gives JPY 4 million / JPY 1.5 million = approximately 2.67 years, or about 2 years and 8 months to payback. Prevision cites 1 to 3 years as one benchmark for payback at small and mid-sized manufacturers, and this calculation falls inside that range.

Sensitivity when the assumptions break down

The problem is that this calculation assumes the 60 percent reduction is actually achieved and that inventory and delivery benefits materialize. Now calculate the case where adoption on the shop floor never takes hold after go-live, the reduction stops at 40 percent, and the inventory and expedited freight benefits do not appear at all.

3,000 hours x 40 percent = 1,200 hours, and 1,200 hours x JPY 900 = JPY 1.08 million per year. Subtracting the JPY 600,000 maintenance cost leaves a net JPY 480,000. JPY 4 million / JPY 480,000 = approximately 8.33 years, or about 8 years and 4 months. That approaches the practical life of the system, so as an investment decision it does not hold up.

For the same investment amount, whether adoption sticks or not moves payback between 2 years 8 months and 8 years 4 months, a factor of more than three. The conclusion to draw from this sensitivity analysis is that what determines ROI is not the feature gap between products but the degree of operational adoption.

Total cost of ownership compared with cloud

Assume the same benefit (JPY 2.1 million per year) can be obtained with a cloud system, and compare against a cloud option at JPY 1 million initial and JPY 100,000 per month (JPY 1.2 million per year). The net annual benefit is JPY 2.1 million – JPY 1.2 million = JPY 900,000. Dividing the JPY 1 million initial cost by that gives JPY 1 million / JPY 900,000 = approximately 1.11 years, or roughly 13 months to payback. Cloud is clearly superior on speed of start-up.

The picture changes, however, when you look at total cost of ownership (TCO).

At the 3-year mark cloud is JPY 1.2 million cheaper, at 5 years the two meet at exactly JPY 7 million, and at 7 years the on-premise package is JPY 1.2 million cheaper instead. In other words, comparing cloud against on-premise without first deciding how many years you intend to use it cannot yield a verdict. A useful guideline is cloud if your product mix or site structure is likely to change significantly within five years, and an on-premise package if you expect to run the same plant on a ten-year horizon.

Five reasons production management system implementations fail and how to avoid them

Cross-referencing the failure cases documented by sources such as IT Trend and the Ninomiya IT Management Institute reveals a common structure behind failed production management system implementations.

Cause 1 Choosing a system that does not match your needs

The most frequent case is a mismatch between your production model and the assumptions built into the system. Put a system designed for make-to-stock mass production into a make-to-order, high-mix low-volume plant, and it cannot absorb the specification changes and urgent insertions that occur every day. The result is a forced workflow that bends operations to fit the system, and workload increases instead, which is exactly backwards.

The countermeasure is to articulate your production model in words at the very start of the comparison. Make-to-stock or make-to-order. One-off engineering or repetitive production. Lot control or serial control. How you hold work in process between operations. Write down these four points, present them to vendors, and explicitly ask whether they have references for this model and whether it is covered by standard functionality or requires customization.

Cause 2 Deciding without input from the people on the shop floor

If selection proceeds only at head office or in the administration department, and the people who will actually key in the data are never consulted, you invite resistance from the shop floor. A production management system cannot output anything without the data the shop floor enters. A system whose users are not convinced ends up with no data going in, output numbers nobody trusts, and eventually nobody looking at it at all.

The countermeasure is to formally place shop floor key people on the project team at the requirements definition stage, and to allocate their hours as part of their job. What matters is not “hearing their opinions” but “having them participate in decisions”. Handing the shop floor the authority to decide details such as entry screen layout and the timing of daily report entry changes their sense of ownership dramatically.

Cause 3 Weak executive interest stalls company-wide alignment

A production management system spans manufacturing, purchasing, sales and accounting. Cross-departmental decisions inevitably arise, such as which department owns which master data, and at what point sales order information is confirmed, and these cannot be settled at the working level. If executives take no interest and leave it to the shop floor, that company-wide alignment bogs down and the project stalls.

The countermeasure is to name an executive as project owner at the point of the investment decision, and to set up a monthly forum to review progress and issues. In particular, listing cross-departmental questions from the outset as “issues to be decided at the management meeting” keeps the working level from being caught in the middle.

Cause 4 Structural factors of complexity, compound causes and organizational silos

Behind individual failures sit more structural factors. Production management is inherently complex as a business function, failures rarely have a single cause and instead come from several intertwined causes, and management, the shop floor and the IT department are siloed from each other. This combination of complexity, compound causes, and the divide between management, the shop floor and IT is what pushes up the difficulty of implementing a production management system.

The countermeasure is not to try to solve everything at once. Divide the scope into phases and break them at points where results are visible: phase 1 is shop floor progress visibility only, phase 2 adds inventory and purchasing, phase 3 adds costing. To address the silos, an effective step is a standing meeting where management, the shop floor and IT all look at the same numbers.

Cause 5 Live but not used

The most serious state is a system that is technically running but not actually used. The classic scene is a plan still being built in Excel right next to an expensive system. This state tends to be recorded as a project success, so the failure never becomes visible and several years pass.

The countermeasure is to stop treating go-live as the goal and start measuring adoption. At 3, 6 and 12 months after go-live, measure adoption KPIs such as data entry rate, entry lag time, how many Excel files remain, and how many people actually look at system screens to make decisions. As the ROI calculation in the previous chapter showed, adoption is a variable that moves the payback period by a factor of more than three. Not measuring it is the same as not measuring whether the investment succeeded.

How to choose a production management system without failing

Here we translate everything above into a checklist you can actually use in a comparison. Simply throwing the same questions at every vendor and lining up the answers raises the quality of the comparison enormously.

Criterion 1 Fit with your production model

Make-to-stock or make-to-order. Repetitive or one-off. Lot or serial. Confirm how far standard functionality goes and where customization begins, using actual screens rather than circles in a feature matrix. Always hand over your own representative part numbers and process routing for the demo and have it shown with that data. A demo on generic sample data provides almost nothing to judge by.

Criterion 2 Division of roles based on the ERP and MES distinction

As described earlier, the hierarchy is ERP at the executive layer, the production management system at the management layer, and MES at the shop floor execution layer, with MES positioned between the planning layer and the control layer (SCADA, DCS and PLC). Draw a diagram to confirm how far up and down that hierarchy each candidate product reaches, and where it overlaps with your existing ERP or equipment-side systems.

Concretely, put the following questions to every vendor in identical form. Which system holds the master record for item master and inventory quantity. Where is costing performed. Where are production results from equipment received. If you sign a contract while the answers to these three remain vague, you end up post-go-live in a state where the same data is being keyed into two places.

Criterion 3 Integration with inventory and purchasing

Production planning and inventory are inseparable. Unless the loop closes, with requirements calculation feeding ordering, receipts feeding inventory, and inventory feeding back into planning, planning accuracy will not improve. If you already run inventory control on a separate system, you need to decide which one holds the master record. Because the right configuration for this point changes with system type, we recommend reviewing the inventory-side perspective in our inventory management system comparison before deciding whether to bring inventory inside the production management system or to integrate with a dedicated system.

Criterion 4 Granularity of shop floor progress recording

The granularity at which you capture production results is a trade-off between data entry burden on the shop floor and the usefulness of the data. Once a day by order, at every operation handoff, or automatically from equipment. The finer the granularity the more analysis you can do, but relying on manual entry exhausts the shop floor and the entry rate drops. Decide concretely, operation by operation, which of barcode, QR, handheld terminals and equipment signals you can use.

Criterion 5 Multilingual support and fit with local operations

At plants in Thailand and ASEAN, language support for screens, reports and manuals is a practical lifeline. Rather than accepting a blanket “Thai supported”, check individually how far translation extends across screen labels, error messages, printed reports, online help and operating manuals. Cases where only error messages remain in English are genuinely common, and they are a reason the line stops.

Criterion 6 Data exit and vendor lock-in

Confirm the exit before you sign. Can master and transaction data be exported with standard functionality, in what format, is there a charge, and how long a grace period do you have to retrieve data after the contract ends. Confirming this keeps migration to a future system a realistic option.

Criterion 7 Support structure and coverage hours

When a failure occurs, who responds, within how many hours, and in what language. If the only option is a support center at the Japanese head office, the 2-hour time difference with Thailand and the language barrier can mean you cannot reach anyone during the hours your line is down. Confirm at SLA level in the contract whether there are response staff on the ground, and if not, what the alternative arrangement is.

Checklist for the comparison

The purpose of this checklist is not to find the outstanding product. It is to screen out the vendors who cannot answer, early. In particular, a vendor whose proposal says nothing about post-go-live enhancement or adoption support is likely in a sell-and-forget posture.

Implementation flow and typical timeline

You often see the figures “1 to 3 months for cloud, 3 to 6 months for on-premise packages” for implementation duration, but in most cases these refer to the build period after signing with the vendor. In reality you need a period before that for internal assessment and product selection. Broken down by phase, it looks like this.

Phases 3 through 5, meaning from post-contract build to go-live, add up to 11 to 24 weeks, roughly two and a half to five and a half months, which lines up almost exactly with the commonly cited 3 to 6 months for on-premise packages. Phases 0 through 2 for internal preparation and selection take another 8 to 16 weeks, roughly two to three and a half months, so you should plan for 19 to 40 weeks overall, roughly four and a half to nine months. With cloud, phase 3 shortens substantially so you can be up and running 1 to 3 months after contract, but note that this does not shorten the internal preparation in phases 0 through 2.

It is tempting to skip phases 0 and 1 to shorten the schedule, but projects that skip them have a high probability of rework from phase 3 onward. If you issue customization requirements without having documented your current business flows, requirements keep appearing late and growing. Seen from the other side, phases 0 and 1 can be done entirely in house, before you even start selecting a system. Since this work can proceed before the budget is confirmed, starting here in the early stage of your evaluation is the most efficient move.

One more point: set the go-live date away from your peak season. At a plant in Thailand, the periods immediately before and after Songkran (April) and the year-end and New Year holidays leave you short of people, and neither training nor a parallel run will work. Requests to align with the start of the fiscal year are common, but if your phase 1 scope stops short of costing, a cutover mid-year is perfectly feasible.

Issues specific to implementing a production management system in Thailand and ASEAN

From here we cover the issues that do not arise in a domestic Japanese implementation but do at sites in Thailand and elsewhere in ASEAN.

Talent retention and the shortage of management-level staff

According to the JETRO FY2023 Survey on Business Conditions of Japanese Companies Operating Overseas (conducted August to September 2023), 40.4 percent of Japanese-affiliated companies in Thailand face a talent shortage, and by job category general management staff are the most severe case, with 79.8 percent reporting a shortage. This has decisive implications for a production management system implementation, because system operation is sustained by management-level staff who understand the business and can read the numbers. If that layer is thin, or if turnover is a given, then a design that does not depend on person-specific operation becomes essential.

Concretely, effective design choices include embedding entry procedures as on-screen guidance, preparing manuals in Thai, and encoding the decision criteria for exception handling as rules inside the system. A design that “makes sense if the right person looks at it” loses its function the moment that person resigns. Note that this survey covers FY2023 and the current situation may differ.

Rising labor costs and the timing of the investment decision

The same FY2023 JETRO survey reported Thailand’s 2023 pay increase rate at 3.8 percent year on year, below Indonesia at 5.7 percent and Malaysia at 4.6 percent. At the same time, the minimum wage was raised to THB 330 to 370 from January 2024 (an average increase of 2.4 percent), and in April 2024 an increase to THB 400 was implemented for large hotels in designated zones. The new government has pledged to raise it to THB 600 by 2027. Indeed, 72.8 percent of Japanese-affiliated companies named rising labor costs as the largest risk in the investment environment.

In this environment, the meaning of reducing indirect workload grows year by year. The ROI calculation in the previous chapter used an hourly rate of THB 200 (JPY 900). If that rate rises, the same number of hours saved produces a larger saving and the payback period shortens. Put the other way around, the longer you keep running on manual processes, the more quietly your costs keep swelling. That said, wage levels will continue to move, so always recalculate with your own most recent actual rate when making an investment decision.

BOI incentives and Industry 4.0 investment

In Thailand, incentives from the BOI (Board of Investment) heavily influence investment decisions. Standard incentives include corporate income tax (CIT) exemption of up to 8 years, followed by a 50 percent reduction for 5 years after the exemption period ends. Combined with the EEC (Chachoengsao, Chonburi and Rayong), CIT exemption of up to 15 years is available, and the technology upgrade under Activity 10.1 carries an additional 3 years of CIT exemption.

Further, for investment in AI workforce training, an extension of CIT exemption is available at 1 year for 1 percent of total annual payroll, 2 years for 2 percent and 3 years for 3 percent, with training costs eligible for a 200 percent deduction. Import duty exemption for qualifying equipment is also set broadly, covering AI hardware, servers, GPUs, image recognition cameras, IoT sensors, industrial robots and edge computing devices. Examples of qualifying Industry 4.0 upgrades include predictive maintenance, visual inspection, AI-based production planning and demand forecasting, energy management optimization, and autonomous transport such as AGVs.

Even when a production management system implementation on its own does not qualify, combining it with elements such as automatic collection of production results from equipment, visual inspection, or AI-based planning optimization may bring it within scope. Because BOI incentives vary greatly by business content and application conditions, always confirm your specific case with the BOI or a specialist. The description in this article is a general summary and does not guarantee eligibility for any individual case.

Looking at Thailand’s investment environment as a whole, the country is shifting toward investment-led growth heading into 2026, and the “Thailand Fast Pass” policy has been announced to accelerate private sector projects worth over THB 300 billion. The BOI’s manufacturing priorities have also shifted toward Industry 4.0 (smart factories, AI-enabled production and automation), with the EEC Automation Park functioning as an implementation hub for robotics and Industry 4.0. Thailand’s digital transformation market is estimated at approximately USD 10 billion as of 2025 and is projected to grow at a CAGR of approximately 8.75 percent through 2031.

Multilingual operation and standardizing training

Issues frequently raised at Japanese-affiliated manufacturers in Thailand include delays on the shop floor that never improve, paperwork that remains on paper and is inefficient, and training for Thai staff that is person-dependent and cannot be standardized. Of these, standardizing training is directly connected to a production management system implementation, because once business procedures are embedded in the system, the training content itself becomes standardized.

Conversely, if you install the system but the procedure documents stay in Japanese, training continues to be oral transmission via the handful of staff who read Japanese. We recommend explicitly including Thai-language operating manuals and role-based training content (for data entry staff, leaders and administrators) as deliverables of the implementation project.

The changing role of the local site

The role of a Thailand site has shifted from the earlier model of expanding abroad for cost competitiveness toward technology deployment, market access and site consolidation. This affects production management system requirements too, because the type you should choose changes depending on whether the goal is efficiency at a single site or whether you have multi-site management including neighboring countries in view. If you anticipate multi-site rollout, you should check at the first quotation whether the licensing model makes the cost of additional sites grow linearly.

Local support structure and the time difference

Finally, a word about life after go-live. Whether routine support such as fault response, master data configuration changes and adaptation to legal changes can be received during local business hours and in the language of the shop floor is a point never considered in a domestic Japanese implementation. The time difference between Japan and Thailand is 2 hours, and the Japanese start of business corresponds to 7 a.m. in Thailand, so mornings overlap. The problem is the other end: the Japanese end of business (18:00) is 16:00 in Thailand. If trouble occurs late in the Thai working day or during the night shift, and your only contact point is in Japan, a gap opens until the next morning. On top of that, public holidays differ between Japan and Thailand, so there are many days a year when one side is closed while the other is operating. This is a point to verify as a practical matter of support structure before you sign.

Summary

A production management system comparison that begins by lining up products tends to leave you lost. The key points organized in this article are as follows.

First, draw the line between the scope of ERP, the production management system and MES before anything else. While that stays vague, you have no foundation for judging which proposal fits your company.

Second, narrow down by type rather than by product name. Once you decide which of cloud, on-premise package, scratch development or industry-specific package fits your production model and scale, the candidates naturally reduce to a handful of vendors.

Third, view cost in three layers of initial, running and hidden. Annual maintenance is roughly 10 to 20 percent of implementation cost, presented as 5 to 15 percent by some sources and 10 to 20 percent by others. The double cost of a parallel run and the data extraction fee on termination do not appear on a quotation.

Fourth, calculate ROI with your own numbers and with the assumptions stated. In this article’s calculation, annual savings were JPY 2.1 million and net benefit JPY 1.5 million, giving payback in about 2 years and 8 months, but without adoption that deteriorated to 8 years and 4 months. What drives the payback period is not the feature gap but the degree of adoption.

Fifth, failure causes come down to five: a production model mismatch, non-participation by the shop floor, weak executive interest, structural silos, and the state of being live but not used. Every one of them can be prevented in the preparation before selection, not after implementation.

Sixth, plants in Thailand and ASEAN add variables that do not exist domestically: talent retention, multilingual support, BOI incentives and local support structure. A design that does not depend on person-specific operation matters even more here than it does in Japan.

Before you start comparing, begin phase 0 current state assessment and phase 1 requirements definition. It looks like a detour, and it is the shortest route.

Where to turn before you evaluate

In the early stage of an evaluation, it is common to get stuck on how to organize your own requirements long before you get to choosing a product. TOMAS TECH is based in Bangkok and implements production management systems and factory IT for Japanese-affiliated manufacturers across Thailand and ASEAN, working in Japanese, Thai and English. We are glad to help at the stage of mapping business flows, separating roles with your ERP, or putting a rough budget together. Because we are based locally, we can also walk you through what support looks like after go-live. If you would like to start with just organizing the issues, please get in touch via our contact page.

Frequently asked questions

How much does a production management system cost?

For cloud, the guideline is JPY 0 to 1 million in initial cost and JPY 30,000 to 150,000 per month (another survey gives free to JPY 500,000 initial and JPY 30,000 to 100,000 per month). For on-premise packages, license plus infrastructure runs from JPY 1 million to over JPY 10 million, with annual maintenance at roughly 10 to 20 percent of implementation cost. Scratch development ranges widely from JPY 5 million to several hundred million. For make-to-order small and mid-sized manufacturers with 10 to 200 employees, JPY 1 million to 5 million for an on-premise package is cited as the realistic guideline. On top of that, budget for the costs that never appear on a quotation: the double cost of running old and new in parallel, post-go-live customization, and data extraction fees on termination.

What is the difference between MES, ERP and a production management system?

It is easiest to organize them as a hierarchy. ERP manages enterprise resources as a whole and handles the information used for management decisions, at the executive layer. The production management system handles planning and control, at the management layer. MES (Manufacturing Execution System) distributes work instructions, tracks progress in real time, collects quality data and monitors equipment operation, at the shop floor execution layer. MES refreshes information by the minute or by the hour and sits between the planning layer (ERP) and the control layer (SCADA, DCS and PLC). Coverage overlaps depending on the product, however, so unless you clarify how far ERP covers and what MES is responsible for at implementation time, functional duplication and integration gaps become likely.

How long does it take to implement a production management system?

The commonly cited figures of 1 to 3 months for cloud and 3 to 6 months for on-premise packages usually refer to the period from post-contract build through go-live. In reality you need roughly two to three and a half months before that for current state assessment, requirements definition and product selection. For an on-premise package, planning for roughly four and a half to nine months from internal preparation through go-live makes the schedule easier to build. Note that the internal preparation portion can be done in house before the budget is confirmed, so starting it early in your evaluation is efficient.

Which is better, scratch development or a package?

If your production model is unusual and that uniqueness is a source of competitive advantage, scratch development can be the rational choice. However, development costs run from JPY 5 million to several hundred million, the timeline is several months to over a year, and maintenance and enhancement continue after go-live. As a decision procedure, first run a fit and gap against an on-premise package, then separate whether the parts that do not fit are a source of competitive advantage or simply inherited habit. If it is the latter, moving the business process toward the standard is more advantageous on cost, timeline and maintainability alike. Considering an industry-specific package can also raise your fit rate while keeping the volume of customization down.

Can we use the same system in Thailand as we do in Japan?

Technically yes, but several points need checking. The extent of Thai language support for screens, reports, error messages and operating manuals; for cloud, where data is stored and how good the site’s internet connectivity is; support for local accounting and tax requirements; and the coverage hours and language of support. Support in particular is a risk: with only a support center at the Japanese head office, the 2-hour time difference and different public holidays can mean you cannot reach anyone during the hours your line is down. The JETRO FY2023 survey also found that 40.4 percent of Japanese-affiliated companies in Thailand face a talent shortage, with 79.8 percent reporting a shortage of general management staff, so it is important to assume turnover and design in a way that does not depend on person-specific operation.

References

• OSK, “A Thorough Guide to Production Management System Cost Benchmarks” (in Japanese) https://www.kk-osk.co.jp/promotion/it_solutions/guide022_seisan.html
• Prevision, “Production Management System Costs” (in Japanese) https://www.smart-go.net/prevision/column/seisan-kanri-cost.html
• Prevision, “Production Management System Failures” (in Japanese) https://www.smart-go.net/prevision/column/seisan-kanri-shippai.html
• IT Trend, “Production Management System Failure Cases” (in Japanese) https://it-trend.jp/production_management/article/failure_case
• Ninomiya IT Management Institute, “Production Management System Column” (in Japanese) https://n-pe.jp/blog-category/production-management-system/column122/
• JETRO, “FY2023 Survey on Business Conditions of Japanese Companies Operating Overseas (Asia and Oceania)” (in Japanese) https://www.jetro.go.jp/biz/areareports/special/2024/0303/f5b4d6344434b2a9.html
• Pertama Partners, “Thailand BOI Manufacturing and Industry 4.0” https://www.pertamapartners.com/funding/thailand-boi-manufacturing
• Thailand Business News, “Top Business Opportunities in Thailand for 2026” https://www.thailand-business-news.com/investment/268723-top-business-opportunities-in-thailand-for-2026
• Unifinity, “The Difference Between Production Management Systems and MES” (in Japanese) https://www.unifinity.co.jp/blog/20260420/
• Layers Consulting, “ERP and Production Management” (in Japanese) https://www.layers.co.jp/column/column32/
• SFA JOURNAL, “Recommended Production Management System Comparison” (in Japanese) https://next-sfa.jp/journal/production-management-system/recommended-production-management-system/
• BOXIL, “Production Management System Price Benchmarks” (in Japanese) https://boxil.jp/mag/a9525/