MES (Manufacturing Execution System): Digitizing the Factory Shop Floor

Mr. Agus Wijaya owns PT Sentosa Plastik Sejahtera, a plastic extrusion and injection molding plant in Sidoarjo, East Java, that he has run since 2011. The 3,200-square-meter facility now runs 18 production machines across three shifts, serving food-and-beverage and pharmaceutical clients across East Java and Bali, with annual revenue of roughly Rp 34 billion.

The plant's real problem was not a shortage of orders — it was not knowing what was actually happening on the floor. Every shift, supervisors logged output, downtime, and rejects by hand on paper checksheets, which a production admin then retyped into Excel the next morning. When Mr. Agus reviewed the performance report at the end of May 2026, he discovered his plant's average OEE (Overall Equipment Effectiveness) was only 51%, far below the 85% widely considered world-class. The reject rate on the injection molding line reached 7.8%, but it was only caught three weeks after the batch had already shipped, because manual reject reports often arrived late and varied from operator to operator. Unplanned downtime consumed an average of 96 hours per month across the 18 machines — equivalent to roughly Rp 410 million in lost revenue per month based on revenue-per-machine-hour.

Things came to a head in June 2026: his largest pharmaceutical client returned a batch of bottle packaging for a dimensional defect, and the quality team needed four days just to trace which machine, which shift, and which operator had produced the faulty batch — because production records were still sitting in an unsorted stack of paper. The real problem was never the machines themselves; it was the absence of a system that captured shop-floor data in real time and in a structured way. That gap is exactly what a Manufacturing Execution System, or MES, is built to fill.

What MES is, and how it differs from ERP

MES is software that sits between the production plan and what is actually happening on the shop floor. Under the ISA-95 industry framework, manufacturing systems are layered: Level 4 is ERP and business planning (what to produce, when, how much, with which materials and budget), Level 3 is MES (what is actually happening on the production floor right now), and Levels 2/1/0 are SCADA, PLCs, and the sensors attached directly to machines.

The distinction can be simplified this way: ERP answers planning questions — which sales orders came in, when they're due, what raw materials need purchasing, how much stock is in the warehouse. That data is typically refreshed daily or per shift. MES answers execution questions — which machine is running right now, who is operating it, how many units have been completed against this work order's target, why machine number 7 stopped for 12 minutes a while ago, and whether this hour's output is still within quality tolerance. That data updates second by second or minute by minute, pulled directly from machines and operators, not from a report re-compiled the next day.

If you're not yet familiar with the broader scope of manufacturing ERP — production planning, purchasing, inventory, and accounting modules — we've covered that separately in our article on manufacturing ERP in Indonesia. This article deliberately focuses on a different layer: MES on the shop floor, which sits below ERP and fills the data gap between the plan on paper and the reality on the factory floor. MES doesn't replace ERP — the two complement each other, and mature factories typically run both with two-way integration.

The real cost of running a factory without MES

• Hidden downtime. Without automated logging, machines that stop for 10-15 minutes multiple times a day are often never recorded because each stoppage seems "too small to bother logging." The monthly accumulation, however, can equal the lost capacity of an entire machine.

• Rejects discovered too late. When rejects are logged manually and compiled daily or weekly, the root cause of a quality issue — a machine temperature drift, or a bad raw-material batch from a particular supplier — is only found after hundreds or thousands of defective units have already been produced, sometimes already shipped to the customer.

• Decisions made on gut feeling, not data. Shift supervisors often rely on memory and verbal handovers to decide which machine needs priority maintenance or which operator needs retraining, because there is no easily accessible, comparable historical data.

• Unmeasured OEE means improvement efforts have no direction. Without a per-machine, per-shift OEE baseline, the engineering team can't tell whether the core problem is availability (machines stopping too often), performance (machines running below rated speed), or quality (too many rejects) — so improvement investments frequently target the wrong thing.

• Weak traceability. When a customer complaint or an ISO 9001, BPOM, or HACCP certification audit comes in, a factory without digital traceability can need days to trace a single batch's raw-material origin and process history — and may not be able to answer at all, risking fines or lost contracts.

Key MES features

• Real-time OEE dashboard. Automatically calculates availability, performance, and quality per machine, per line, and per shift, so production managers can see which machine is holding back output without waiting for a monthly report.

• Automated downtime and reject tracking. Every time a machine stops, the operator selects a reason code (mold change, material jam, waiting for operator, maintenance) via a tablet or physical button, so downtime root causes are neatly categorized and their weekly trends can be analyzed.

• Digital work instructions for operators. SOPs and work drawings are displayed on a screen next to the machine, always synced to the latest version from head office — replacing paper SOPs that get lost, stained with oil, or fall out of date.

• Batch/lot traceability. Every product unit can be traced back to the raw-material batch number, the machine, the operator, and the process parameters used, so when a recall or complaint occurs, the QC team can narrow the affected scope within minutes instead of days.

• Machine and IoT integration (PLC/SCADA/sensors). MES pulls data directly from machine counters and temperature, pressure, and speed sensors via an IoT gateway, so output and downtime data no longer depend on manual operator entry that's prone to error.

• Quality and reject tracking integrated with process statistics. Reject data is automatically grouped by defect type and shown as trend charts, making it easier for the quality team to apply statistical process control to catch deviations before they become major problems.

• Two-way MES-ERP integration. Work orders planned in ERP flow down automatically to MES as shop-floor work instructions, and conversely, actual material consumption and output data from MES flow back to ERP to accurately update warehouse stock and production costing.

• Shift and labor tracking. The system records which operator worked which machine during which shift, along with their productivity — useful both for performance evaluation and as a basis for production incentive calculations.

Build vs buy (SaaS vs custom)

For factories with relatively uniform, newer machines — say, a modern automotive line running standard Siemens or Rockwell PLCs — global MES platforms such as Siemens Opcenter, AVEVA (formerly Wonderware), Delmia Apriso, or SaaS platforms like Tulip can be a fast way to get started. Pricing is usually per machine node or per user per month, and the standard feature set is fairly mature.

In practice in Indonesia, though, most mid-sized factories run a mix of older machines with no digital connectivity and newer smart machines, plus a local ERP or accounting system (Accurate, Jurnal, or a custom-built system) that doesn't always have a ready-made connector to global MES platforms. Foreign SaaS licensing costs also grow with every machine added, and local support is often limited to different time zones and business hours.

That's why many mid-sized Indonesian factories end up choosing a custom-built MES: developed to match the exact number and type of machines they actually own, integrated directly with the ERP and accounting systems already in use, with per-node licensing costs that don't balloon as the factory grows, and full ownership of the data on their own server or chosen cloud. The trade-off is a longer initial development timeline compared with simply subscribing to a SaaS product.

Cost and development timeline ranges in Indonesia

As a rough guide based on similar projects in Indonesia:

• Pilot MES for 5-10 machines (basic OEE dashboard, downtime and reject tracking, simple PLC integration): roughly Rp 180-350 million, 3-4 months of development.
• Mid-scale MES for 20-40 machines with batch/lot traceability, digital work instructions, and full two-way ERP integration: roughly Rp 550 million to Rp 1.2 billion, 6-9 months of development.
• Enterprise-scale MES for multi-plant or complex multi-line operations: roughly Rp 1.5-3 billion and up, 10-14 months of development.

Beyond software costs, budget for IoT gateway and retrofit sensor hardware at roughly Rp 3-8 million per machine, depending on whether the machine already has a modern PLC or needs additional external sensors. Annual support and maintenance typically runs 15-20% of the initial development value. For a more detailed breakdown of packages and cost components, see our pricing page.

Case study

PT Bintang Tekstil Nusantara, a weaving and textile finishing plant in Majalaya, Bandung Regency, operates 32 weaving machines and 6 finishing machines. Before implementing MES, the plant looked like many mid-sized textile factories: average OEE of just 47%, unplanned downtime of 140 hours per month across lines, a 9.2% reject rate in the finishing process, and manual production reports that typically reached the production manager only the following day.

In mid-2025, the plant began a custom MES pilot on 8 critical weaving machines over three months, before rolling it out fully to all 38 machines in early 2026. The system covers a real-time OEE dashboard, downtime tracking with reason codes, yarn-to-fabric batch traceability, and integration with the ERP already used for purchasing and accounting.

Eight months after the full rollout, as of June 2026, the results were: OEE rose from 47% to 68% (up 21 points), unplanned downtime fell 55% to 63 hours per month, finishing reject rate dropped from 9.2% to 4.1%, and quality issue detection time, which used to average 3 days, is now under 10 minutes via automated dashboard alerts. Management estimates the project paid for itself in about 14 months, based on savings from reduced rejects and additional output gained from reduced downtime.

Metrics to monitor after implementation

• OEE (Overall Equipment Effectiveness) per machine, per line, and per shift, compared against the pre-MES baseline and industry benchmarks.
• MTTR (Mean Time To Repair) and MTBF (Mean Time Between Failures) to measure maintenance team effectiveness and flag machines needing more attention.
• Reject rate and first pass yield per production line, to confirm quality improvements are consistent rather than a temporary blip.
• On-time work order completion rate, the percentage of work orders finished within the time target planned by ERP.
• Changeover or setup time between production batches, particularly important for factories with high product variation.
• Traceability data completeness, the percentage of production batches recorded completely from raw material to finished good with no missing data.

Where to start

If your factory is still relying on paper checksheets and Excel recaps like PT Sentosa Plastik Sejahtera before it cleaned house, the most realistic first step isn't building an MES for the entire plant at once. Start with a pilot on one or two of your most critical or most problematic lines, measure the impact over one to two quarters, then expand to other lines using a pattern you've already proven. Our team at AFSS regularly designs custom MES systems tailored to the real machine mix on the ground — a blend of old and new equipment — and integrates them directly with the ERP and accounting systems you already run. Check estimated costs on our pricing page, or go ahead and submit your project to discuss your factory's specific needs.