On the shop floor of an automotive parts maker, three pressures arrive at the same time. The gap between forecast orders and firm orders keeps moving, kanban drives small and frequent deliveries, and engineering changes never stop. Whether you have chosen a production management system that fits automotive parts work therefore shows up directly in your on-time delivery rate and in the accuracy of your costing. This article looks at the question through production planning and BOM management rather than traceability, and organizes the requirements an automotive parts maker should nail down along with a practical selection sequence. It also covers how those requirements differ for electronic components, metal processing, and plastic molding.
What changed around the automotive parts industry in 2026
Requirements for a production management system shift as the industry around it shifts. Before going into features, it is worth confirming three parts of the 2026 external environment that are actually affecting production management at parts makers.
Demand is harder to read, and forecast orders no longer mean what they used to
Take Thai vehicle production as an example. According to Federation of Thai Industries figures reported by JETRO, cumulative vehicle production for January to June 2026 was 717,212 units, down 1.0% year on year. Within that total, passenger cars came to 252,371 units, down 4.7%, while commercial vehicles reached 464,841 units, up 1.0%, so the direction differs by vehicle type. Export-bound production was 449,161 units, down 5.4%, while production for domestic sales was 268,051 units, up 7.4%, meaning the sign also flips depending on destination. The Federation of Thai Industries revised its full-year 2026 production outlook down from 1.5 million units to 1.45 million units, cutting the export-bound portion from 950,000 units to 900,000 units.
So the headline is a roughly 1% dip, but break it down by vehicle type and destination and you find near double-digit moves in both directions sitting side by side. That structure feeds straight into a parts maker’s production plan, because company-wide sales can look flat while individual part numbers swing hard. The old habit of “the forecast looks like last year, so order material like last year” stops working. You need forecast history held per part number, and you need to decide order quantities while watching how much that particular part number has moved recently.
Electrification is reshuffling the part number mix
JETRO also compiled data from Thailand’s Department of Land Transport showing that new registrations of passenger battery electric vehicles for January to July 2026 reached 126,439 units, up 88.2% year on year. In July alone, passenger BEV registrations came to 21,159 units, roughly 10% higher than the same month a year earlier. By brand, BYD held 19.3%, CHERY 18.5%, SAIC Motor-CP 14.0%, and GEELY 11.5%, bringing the combined share of Chinese makers to 90.3%.
Seen from a parts maker, this shows up as existing part numbers thinning out while new ones ramp up. Legacy part numbers sitting in a volume trough and new part numbers still going through drawing changes before mass production compete for the same equipment in the same plant. What a production management system has to deliver is not only the ability to run stable high-volume items efficiently, but the flexibility to handle prototype, initial-flow, and mass-production part numbers on the same planning table.
Margins are thin, and costing accuracy is under scrutiny
The automotive industry supply chain survey published by Teikoku Databank on 14 July 2025 analyzed 68,338 supply chain companies connected to ten domestic Japanese automakers. Operating margins averaged 1.4% overall, at 2.8% for Tier 1, 1.3% for Tier 2, and 0.6% for Tier 3 and below. The share of companies posting an operating loss was 31% overall and reached 35% at Tier 3 and below. The ratio between Tier 1 and Tier 3 operating margins widened from 1.3 times in 2018 to 4.7 times in 2024.
An operating margin around 1% means a costing error of a few percent is enough to flip a part number from profit to loss. The practical problem is that in many plants, actual cost still depends on a monthly roll-up in Excel. Finding out after the month closes that a given part number lost money leaves almost nothing you can still do about it. The value of investing in a production management system lies precisely in whether it can compress that lag.
This earnings environment also shows up in how parts makers see themselves. A survey published by PwC Advisory on 27 April 2026 found that about 80% of 157 Japanese automotive suppliers said alliances, meaning collaboration between companies, would be necessary to address the challenges ahead. The same survey noted a wide gap between what companies recognize as a challenge and what they have actually acted on. Whether you pursue collaboration or improve on your own, the prerequisite is the same. You need profitability and load per part number visible as numbers, and building that foundation is what a production management system is for.
For context on Japanese manufacturing as a whole, the 2026 Monozukuri White Paper approved by the Cabinet on 29 May 2026 compares nominal labor productivity per person at USD 75,000 in Japan, USD 226,000 in the United States, and USD 117,000 in Germany. It notes that while about 70% of businesses collect data from their manufacturing processes, only about 40% are actually getting results from using it, and that data linkage between companies has barely progressed in the past two years. Collecting data is no longer unusual. The stage we are in now is one where the difference comes from whether collected data flows back into planning and costing.
What makes selection hard specifically for automotive parts
Scanning a generic feature list for production management systems will not reveal what is difficult about automotive parts. Here are five points where real implementations tend to stumble.
Can forecast orders and firm orders live on the same screen
Automotive parts orders usually arrive first as a rolling forecast covering several months ahead, with only a defined near-term window converting into firm orders. That forecast is refreshed weekly or even daily, and the quantities move every time.
When you evaluate a production management system, the thing to check is whether forecasts are merely parked in a separate table as reference information. If the forecast stays isolated reference data, the requirements calculation runs on firm orders alone and material procurement is permanently a step behind. Feed the forecast straight into requirements instead, and a downward revision leaves you sitting on excess inventory.
The workable requirement looks like this. First, you must be able to set, per part number, how many weeks of the forecast horizon are treated as procurement-relevant. Second, you must be able to see, in one list, the difference between the previously received forecast and the current one. Third, when a forecast line converts to a firm order, the forecast must be consumed automatically so the demand is not double counted. A system missing any of these three only half works against the automotive parts ordering pattern.
Do not overlook the fact that EDI receiving formats differ by customer. For the same forecast information, the file format, the field names, and the granularity of the delivery instruction all vary by automaker. During selection, name specific customers and establish how far the standard functionality goes and where add-on development begins.
How do you reconcile kanban delivery with lot production
Almost every automotive parts plant lives with the same contradiction. The delivery side wants small, frequent shipments driven by kanban, while the manufacturing side wants to consolidate lots because changeover on a press or a molding machine is expensive. A production management system alone will not resolve that contradiction, but how the system is built changes how manageable it is by a wide margin.
What you want to confirm is whether the unit of the delivery instruction can be separated from the unit of the manufacturing order. If the design forces the number of delivery kanban cards to become the manufacturing lot, changeovers multiply and equipment utilization falls. Plan on manufacturing lots alone and the finished goods area overflows. In practice, most plants settle on a middle path, holding an upper and lower finished goods inventory limit per part number and rounding manufacturing lots within that band, so the system needs settings that support it.
Also check whether you can manage the receiving time window and the truck run at the customer. When morning and afternoon runs on the same day require different quantities, a system that can only hold a delivery plan at day granularity sends the floor straight back to Excel.
Can the BOM keep up with engineering changes and 4M changes
BOM management is where automotive parts production management goes wrong most often. Drawings change, material suppliers change, dies change, process sequences change, and these keep happening to part numbers already in mass production. In 4M terms, changes to Man, Machine, Material, and Method arrive continuously rather than as one-time events.
The key here is effective date management. If the system cannot hold the date from which a new structure applies, every switch-over becomes a manual BOM overwrite, and inventory and work-in-process calculations across the switch stop reconciling. Whether the system can hold the old and new structures in parallel and have the requirements calculation switch automatically at a specified date is something you must verify.
Equally important is being able to trace the impact of a change upward. When the structure of one child part changes, you need to list the parent part numbers that use it, and the higher-level finished goods that use those parents. Without that, every change turns into an email round to the departments involved to establish the impact, and things get missed.
We cover in more depth how the freshness of BOM and master data itself determines the quality of the requirements calculation in our article on master data accuracy in MRP systems. Read alongside this one, it makes clear why plants end up in the state of having installed a system whose calculated results nobody trusts.
Linking change history to manufacturing lots so it can be traced afterwards, in other words designing for traceability, is outside the scope of this article. For how to think about traceability including IATF 16949 compliance, see automotive parts traceability and IATF 16949. On the production management side, the first job is producing a correct answer to the question of which structure you should be building to right now.
Can high-mix low-volume production and changeover be built into the plan
Once part numbers grow from the hundreds into the thousands, planning by hand stops working. Yet a plain requirements calculation on its own does not optimize the number of changeovers, and the plan on paper drifts away from what the floor actually executes.
The evaluation point is whether setup time can be determined automatically from product group or die attributes. Running part numbers with the same material, the same color, or the same die base back to back keeps changeover short. A system that holds this definition of similar part numbers in the master and considers it during planning reduces how much the plan depends on the planner’s tacit knowledge.
The second point is whether planning granularity can differ by process. A press process wants daily lot planning, an assembly process wants hourly sequence planning, and the granularity each needs is genuinely different. Force every process onto the same granularity and you produce a plan that is unusable in one of them.
How finely should you capture cost
With cost management, finer is not automatically better. A very common failure is trying to collect actuals at every process, loading the floor with data entry, and ending up with nothing entered at all.
For automotive parts, cost variances concentrate in three or four places, namely material yield, setup time, rework on defects, and outsourced processes. Design the system to capture actuals only there and carry the rest at standard cost, and you get meaningful variance analysis while keeping the entry burden low. Given an industry-wide operating margin in the 1% range, the deciding factor for the investment case is whether you can reach a state where gross margin per part number is visible weekly rather than monthly.
Functional requirement checklist for automotive parts
Here are the points above, organized so you can check them during selection. Print this table and bring it to a vendor demo, and you will not be carried along by the feature-list walkthrough.
| Business area | Requirement specific to automotive parts | Question to ask |
|---|---|---|
| Orders and forecasts | Manage forecast and firm orders separately, consume the forecast automatically when the order is firmed | Can forecast differences be listed against the previously received version |
| EDI integration | Receive customer-specific formats and import delivery instructions | Which customers are already supported as standard, and who builds the additions |
| Production planning | Separate delivery units from manufacturing lot units, manage finished goods min and max | Can the lot rounding rule be configured in the master |
| Changeover optimization | Group similar part numbers by material, color, die, and other attributes | Is setup time calculated automatically from the master |
| BOM management | Structure management with effective dates, old and new structures held in parallel | Does the requirements calculation switch automatically at a specified date |
| Engineering change | List the impact of a change up through higher-level part numbers | Is there a screen to look up parent part numbers from a child part |
| Inventory and kanban | Delivery planning by truck run, work-in-process visibility by process | Can morning and afternoon runs on the same day be planned separately |
| Cost management | Separate variances for material yield, setup, defects, and outsourcing | How many days after the fact is actual cost per part number finalized |
| Shop floor data capture | Actuals entry with a low burden on operators | How many entries and how many seconds does it take, and can we try it on the real machine |
Every item in this table is there to ask how much effort something takes, not simply whether it exists. Most packages can achieve the majority of it with enough configuration, but cost and maintainability depend on whether it is standard functionality or add-on development. If you want to compare from the level of package categories first, reading our guide to comparing and selecting production management systems beforehand will make this table easier to use.
How requirements differ by manufacturing method
Automotive parts is not one thing. Requirements change with the manufacturing method, and it is not unusual for one plant to run several methods under the same roof. Here we compare three representative areas plus final assembly.
Electronic component mounting, metal processing, plastic molding, and final assembly are all major processes in the automotive parts supply chain, but what each one needs from a production management system differs as follows.
| Method and process | What matters in planning | Key points for inventory and actuals | Main driver of cost variance |
|---|---|---|---|
| Electronic component mounting | Component supply units on the mounter and board panelization | Remaining quantity per reel, reverse lot lookup | Component scrap and mounting defects |
| Metal processing and press | Die changeover and material nesting | Coil and material lots, work-in-process between processes | Yield and outsourced process unit prices |
| Plastic molding | Number of mold cavities and molding cycle | Resin drying and mixing lots, post-molding inventory | Color change loss and defect rate |
| Automotive final assembly | Sequence planning from forecasts and synchronized supply of components | Work-in-process on the line and shortage prediction | Waiting time and rework |
The differences in the table above translate directly into the volume of configuration work at implementation. Here is what the three main manufacturing methods mean in practice.
Applying a production management system to electronic component processes
In electronic component and board mounting processes, the unit of component supply is a reel or a tray. If the production management system holds inventory only as a piece count, you cannot see how many pieces remain on the mounter and the line stops mid-run. Panelization is also particular to this area. Without a coefficient for how many pieces come off a single board, the requirements calculation does not reconcile.
Electronic components are also an area with long and volatile procurement lead times. The window from forecast to purchase order typically needs to be longer than in other methods, so being able to vary the number of forecast weeks treated as procurement-relevant per part number becomes close to mandatory in practice.
Applying a production management system to metal processing
In press and machining work, material nesting and process outsourcing are the two big issues. Nesting determines yield through how many pieces come out of one sheet, so holding material consumption in the BOM as a fixed value alone will not make cost reconcile. You need a design that keeps the pieces-per-sheet figure per part number and isolates the difference against actual material input as a yield variance.
Process outsourcing means sending plating, heat treatment, painting, and similar operations outside. Work-in-process that has gone to a subcontractor tends to disappear from your own inventory view, and that is where the hole in the plan opens up. Confirm whether outsourced issue and receipt can be built into the plan as processes, and whether dwell time at the subcontractor can be captured from actuals.
Applying a production management system to plastic molding
In plastic molding, the mold is the lead actor in planning. Several part numbers may share one mold, one mold may yield several different parts at once, and the number of cavities changes how many pieces come out per shot. All of that has to be reflected in the plan. The dividing line is whether the system registers molds as resources and lets you plan against mold availability.
Color change and material change losses matter too. Switching from a clear resin to a dark color, or the reverse, consumes purge material and time. Being able to determine setup time automatically from a color attribute takes a large manual resequencing burden off the planner.
How to run the project from selection to go-live
Once requirements are organized, decide how you will run the project. For an automotive parts maker, the following five stages minimize rework.
The first stage is taking stock of where you are. Count part numbers, customers, EDI formats, engineering changes per month, processes, and subcontractors. Producing those numbers up front changes the accuracy of every later estimate. The monthly count of engineering changes in particular is a critical indicator of how strong your BOM management requirements need to be.
The second stage is prioritizing business requirements. Use the checklist above and sort each item into must-have, nice-to-have, and not needed. Mark everything as must-have and no package will fit, which pushes you toward scratch development. Given the operating margin levels in this industry, the cases where that level of investment pays back are limited.
The third stage is vendor selection. Narrow to about three candidates and compare them against the same requirement table. At this point, always ask whether they have live sites with the same manufacturing method and the same scale as yours. Ask at the granularity of a plant running forecast management on a press process, not the vague claim of extensive manufacturing experience.
The fourth stage is a trial. Using a slice of real data, run one full circuit from receiving a forecast through requirements calculation, manufacturing orders, and actuals entry. Have the people who will actually use it try it here, and time how many seconds a single entry takes. Skip this stage and go straight to production, and you land in the classic failure where data entry never takes hold after go-live.
The fifth stage is phased go-live. Rather than switching every process at once, start with one line or one product group. As the Monozukuri White Paper points out, getting results from collected data is harder than collecting it. Starting small and confirming that planning accuracy and cost accuracy genuinely improved before expanding gets you to a company-wide rollout faster in the end.
Additional considerations for sites in Thailand and ASEAN
When a Japanese automotive parts maker selects a system for a plant in Thailand, a separate set of considerations joins the ones that apply in Japan.
First is language. Work instruction screens and actuals entry screens on the floor need Thai, management reports for the Japanese head office need Japanese, and audit response needs English. Confirm that all three languages can be switched over the same data. Whether part names in the master can be held per language also matters a great deal in daily operation.
Second is accounting and tax. Thai value added tax, withholding tax treatment, and raw material management under BOI privileges all work differently from the Japanese setup. Decide at the outset whether this is handled within the production management system or pushed to the accounting system, otherwise you end up with duplicate data entry.
Third is staff turnover. In Thai manufacturing, personnel changes can happen more often than in Japan. A design that depends on one person’s way of doing things stops the moment that person moves on. Documenting master maintenance procedures and building things so anyone produces the same result matters even more here than it does in Japan.
Fourth is a changing customer base. A seminar on the Japanese automotive parts industry in Thailand, held by JETRO and the Embassy of Japan on 27 January 2026 in Si Racha, Chonburi province, drew about 260 participants in a hybrid format. Presentations at that seminar noted that the growing presence of Chinese automotive suppliers in Thailand could affect the Thai automotive supply chain as a whole. Customers changing means order formats and delivery rules change with them. Choosing a system where adding a new customer takes little work makes that shift much easier to absorb.
For the broader selection procedure for plants based in Thailand, our guide to selecting a production management system for Thai factories covers wider ground. Refer to it if you want a comparison that includes industries beyond automotive parts.
Common failure patterns and how to avoid them
Plants where implementation goes badly tend to follow a few recognizable patterns.
The first is going live without ever handling forecasts. Requirements calculation runs on firm orders alone and material procurement continues in Excel as before. The system then stays a tool for order entry and actuals roll-up, and planning accuracy does not improve. The only preventive measure is to make loading forecasts into the system a mandatory requirement before implementation begins.
The second is treating the BOM as something you build once. The structure that was correct at go-live gets managed through notes on the floor with every engineering change, and six months later the way things are actually built no longer matches the structure in the system. Even with effective date functionality, the same thing happens if nobody has defined who updates it and how.
The third is capturing cost in too much detail. Requiring start and end stamps at every process increases the burden on the floor, entries fall behind, and nobody trusts the roll-up any more. Narrowing actuals collection to the processes where variance actually appears and carrying the rest at standard cost produces a system that gets used.
The fourth is writing the specification from head office requirements alone. Optimizing for the reports the Japanese head office wants produces screens that are awkward for the Thai site, and operation splits into two parallel tracks. Designing from the screens local operators actually touch produces higher utilization.
Frequently asked questions
Can a production management system be used in the electronic components industry
Yes, provided it can manage component supply units such as reels and trays and handle board panelization coefficients. Install a generic production management system as is and inventory can only be held in piece counts, which makes it impossible to predict component run-out on the mounting line. During selection, ask specifically whether the vendor has live sites in electronic components or board mounting.
How do production management system requirements differ between metal processing and automotive parts
Requirements for automotive parts center on forecast management and support for kanban delivery, while requirements for metal processing center on material nesting and management of outsourced processes. A metal processing plant serving automotive parts needs both. Choose a system strong in only one and the other side of the business stays in Excel. The reliable approach is to compare how much of both a system covers with standard functionality, using the same requirement table.
Can a production management system manage molds in a plastic molding plant
It can, if the system registers molds as planning resources. The points to check are allocation when one mold is shared by several part numbers, a mechanism to calculate pieces per shot from the cavity count, and whether maintenance management based on mold shot counts is handled in the same place. Split mold management into a separate system and planning stops being synchronized with mold availability, which makes the plan unexecutable.
Will loading forecasts into the system increase inventory
It will if you pour forecasts straight into requirements. You can hold it down by setting, per part number, how many weeks of the forecast horizon are treated as procurement-relevant, and shortening that window for part numbers that swing a lot. If the system can accumulate the gap between past forecasts and actual firm quantities and hold it as a variability rate per part number, you can set that window based on numbers rather than instinct.
How long does implementation take
It depends on the breadth of requirements, but for a single plant covering forecast management, production planning, BOM management, and shop floor data capture, many cases run from six months to a year from requirements definition to phased go-live. A high volume of engineering changes, many customer EDI formats, or multiple sites will extend it. The figures you counted in the first stage of taking stock become the basis for that estimate.
Do we have to replace all our existing Excel work
No. Making full replacement the goal actually inflates requirements and delays go-live. A realistic split is to move areas where several departments share data, such as requirements calculation, BOM, inventory, and actuals, into the system, while leaving analysis sheets used by one individual in Excel. What matters is standardizing on the system data as the source of truth with Excel reading from it, and leaving no flow that runs the other way.
Summary
In selecting a production management system for an automotive parts maker, what proves decisive is not the number of features but whether the system can withstand three issues, namely the dual structure of forecast and firm orders, the separation of kanban delivery from manufacturing lots, and BOM management with effective dates. In Thailand in 2026, full-year vehicle production has been revised down to 1.45 million units while BEV registrations are up 88.2% year on year, and the reshuffling of the part number mix continues. Given an industry-wide operating margin in the 1% range, the risk of continuing to see cost only on a monthly basis is greater than it used to be. Start by counting your own part numbers, engineering changes, and EDI formats, sort requirements into mandatory and optional, and only then compare candidates. Which of electronic components, metal processing, or plastic molding you run will change what additional requirements you need.
TOMAS TECH implements and supports the PEGASUS production management system for Japanese manufacturers in Bangkok, Thailand, including live sites at automotive parts plants. We are happy to start from the stocktaking stage if you simply want to work out how much of your requirements a package can cover and where add-ons begin. Even if you are only at the entry point of comparing options, please contact us and we will be glad to help.
References
- Vehicle production in the first half fell 1.0% year on year (Thailand) – JETRO Business Brief
- Passenger BEV registrations for January to July 2026 up 88.2% year on year (Thailand) – JETRO Business Brief
- JETRO and the Embassy of Japan hold a seminar on the Japanese automotive parts industry in Thailand – JETRO Business Brief
- Automotive industry supply chain survey, July 2025 – Teikoku Databank
- Survey report on the outlook for the automotive supplier industry, steps toward transformation from a survey of more than 150 companies – PwC Advisory LLC
- 2026 Monozukuri White Paper approved by the Cabinet, improving earnings through capital investment and AI adoption – Project Design Online
