Lots of Production Data, but Decisions Are Still Too Slow? How Smart Manufacturing Can Break Down Information Silos

As manufacturing becomes increasingly digitalized, many factories have already accumulated large volumes of data. Equipment may continuously generate operating information through PLCs and sensors, Manufacturing Execution Systems (MES) may record work orders and production progress, quality systems may store inspection results, and Enterprise Resource Planning (ERP) systems may manage orders, materials, inventory, and other enterprise resources.

On the surface, manufacturers appear to have access to more information than ever before.

However, once we look more closely at actual shop-floor operations, another situation is still common: equipment has its own data, production departments maintain their own reports, quality teams use separate systems, and maintenance records may be stored on yet another platform. When managers need to answer questions such as “Why is today’s output behind schedule?”, “Can this order still be delivered on time?”, or “Which problem is currently affecting the production line?”, they may still need to contact different departments, download data, organize Excel files, and compare information manually.

The problem, therefore, is not necessarily that a company “does not have enough data,” but that its data is not truly connected.

NIST has long identified Systems Integration and Interoperability as important issues in Smart Manufacturing. Smart Manufacturing requires more than simply connecting equipment. Different machines, software platforms, and manufacturing systems need to be able to exchange and use information. NIST research on Smart Manufacturing standards also emphasizes that information is critical to building more adaptive manufacturing systems, and that one important role of standards is to ensure that the right information can be accessed at the right time to support subsequent action.

Therefore, as Smart Manufacturing becomes more deeply embedded in enterprise operations, the question is no longer only “What additional data should we collect?” but also “Can the data we already have actually be transformed into information that supports decision-making?”

Why Can Information Still Remain Disconnected Even After a Factory Has Become Digitalized?

Digital transformation in manufacturing is rarely completed all at once.

A company may first implement ERP to manage finance and orders, then add MES to improve shop-floor management. Equipment suppliers may provide their own monitoring software, the quality department may use an independent Quality Management System (QMS), and the maintenance team may operate a separate Computerized Maintenance Management System (CMMS).

Each system solves a specific problem and may create value on its own.

The challenge is that when these systems are introduced at different times and by different vendors, they may use different data formats, equipment names, product codes, time records, and information architectures.

For example, MES may show that output for a particular work order is behind schedule, but if the system cannot quickly access equipment downtime information, the production manager may still need to check another equipment system. An equipment platform may show an abnormal parameter, but if that information cannot be connected with quality results, it may be difficult to determine immediately whether product quality has already been affected. ERP may show that an order is approaching its due date, but if the latest production progress is unavailable, the actual completion status may still need to be confirmed manually.

All of these systems contain data, yet the data may still remain trapped in separate “Data Silos.”

NIST’s Smart Manufacturing Systems Design and Analysis Program notes that many Smart Manufacturing applications are still concentrated at the plant level, using information technology, Sensor Networks, Computerized Controls, and Production Management Software to improve efficiency. More advanced Smart Manufacturing requires expansion toward the enterprise level so that real-time control and Data Analytics can extend across broader enterprise systems.

Therefore, the next stage of factory digitalization is not simply adding more systems, but gradually connecting the systems that already exist.

Having Equipment Data Does Not Mean Managers Have Usable Information

Equipment connectivity is often an important starting point for Smart Manufacturing.

Through PLCs, Sensors, the Industrial Internet of Things (IIoT), and other equipment communication methods, manufacturers can continuously collect information such as machine speed, temperature, pressure, vibration, operating status, alarms, and other signals.

However, once equipment begins generating large amounts of data, companies still need to answer another question: what does this data actually mean?

Suppose a machine stops for 18 minutes at 2:15 p.m. Looking only at the equipment data, managers can see that the machine stopped. But what they may really want to know is: Which work order was being produced at that time? What caused the downtime? How much output was lost? Did it affect downstream processes? Could this delay the order?

These questions cannot be answered by equipment data alone.

Equipment information needs to be connected with work orders, production schedules, products, quality data, maintenance records, and even order information before it can move from simple Machine Data toward Manufacturing Information that truly supports operational decisions.

NIST research on IIoT and Smart Manufacturing notes that IIoT creates new forms of interaction among hardware, software, and people, while standards play an important role in ubiquitous connectivity, interoperable integration, and more advanced Smart Manufacturing capabilities.

In other words, equipment being “able to transmit data” is only the first step. What matters more is whether different systems can understand and use that data.

Why Do MES, ERP, and Equipment Systems Need to Be Gradually Connected?

MES and ERP serve different management levels within manufacturing companies.

ERP typically manages orders, purchasing, materials, inventory, and enterprise resources, while MES operates closer to the shop floor and records work-order execution, output, production progress, equipment, and personnel. Below these systems are PLCs, Sensors, Machine Controllers, and other forms of Operational Technology (OT), which capture actual equipment operating conditions.

If these levels remain disconnected, information delays can easily occur.

For example, ERP may already contain a new customer order, but the production schedule may need to be entered manually into another system. A work order may already be completed on the shop floor, but ERP inventory records may still need to wait for manual updates. Equipment may already have stopped, but management may not notice until a production report shows that output is below target.

When information needs to be repeatedly entered, exported, and reorganized manually, it not only takes more time but can also increase the risk of inconsistent data.

Systems Integration in Smart Manufacturing, therefore, is not simply about putting all systems onto the same screen. It is about enabling information to be exchanged within the appropriate process.

Order information can be translated into production requirements, production progress can be fed back into management systems, equipment status can be linked with work orders, and quality results can be connected with actual process conditions.

At that point, companies are no longer looking only at “the data inside each individual system,” but are beginning to form production information that spans multiple processes.

The Real Challenge Is Not Just Connecting Systems, but Ensuring They “Understand the Same Thing”

System integration often sounds like an IT problem: creating APIs, connecting databases, or enabling software platforms to exchange data.

In manufacturing, however, the more difficult challenge often involves the meaning of the data itself.

Suppose two systems both record “equipment status.” One uses Running, Idle, and Down, while another uses codes such as 1, 2, 3, and 4. Without a shared definition, even if data is successfully transmitted, the receiving system may not know what each value actually means.

The same issue can occur with equipment names, work-order numbers, product codes, quality defect classifications, and downtime reasons.

Interoperability, therefore, is not just about whether systems can transmit data to one another. It also includes whether information can be consistently interpreted and used across different systems.

NIST research on Smart Manufacturing Standards notes that Smart Manufacturing integration needs to span different lifecycle dimensions, including products, Production Systems, and Business processes. Relevant standards can help ensure that the right information is provided at the right time to support decisions and action.

This is also why manufacturers need to address Master Data, Naming Rules, Data Definitions, and other data-governance issues in addition to technical integration.

If different departments still define the same information differently, even a large number of Dashboards may simply display previously fragmented problems in one place without creating a shared operational language.

Real-Time Dashboards Are Important, but “Seeing” Is Not the Same as “Being Able to Decide”

Many smart factory projects establish Production Dashboards, Andon Systems, or Digital Command Centers to help managers understand production conditions more quickly.

These tools can certainly improve information transparency, but the Dashboard itself is not the ultimate objective of Smart Manufacturing.

What matters is whether managers can quickly understand the cause of an abnormality and take action after seeing it.

For example, a Dashboard may show that output on a production line is below plan. If it only shows “current achievement rate: 82%,” the supervisor may still need to investigate the cause.

However, if the system also connects equipment and production information, managers may be able to see that a bottleneck machine has accumulated 45 minutes of downtime today, including 30 minutes caused by waiting for materials; WIP has already begun to accumulate upstream; and based on the current production rate, a specific work order may not be completed according to the original schedule.

At that point, data truly begins to become information that can support action.

The value of Smart Manufacturing is not to create more KPIs, but to shorten the time between “a problem occurs → information is obtained → the cause is understood → action is taken.”

Poor Data Quality Can Undermine Even Advanced AI

Once manufacturers accumulate large volumes of production data, the next step often seems to be AI, Machine Learning, or other forms of Advanced Analytics.

However, the value AI can create still depends heavily on the quality of the underlying data.

If equipment timestamps are inconsistent, downtime reasons are frequently missing, the same product uses different codes across systems, or sensor data is repeatedly interrupted, analytical models may be built on incomplete or inconsistent information.

NIST research on manufacturing data collection, management, and reuse identifies Data Collection, Data Curation, Data Management, Connectivity, and Interoperability as important issues in Smart Manufacturing data environments.

Therefore, before asking “How can AI analyze our production data?”, manufacturers may first need to answer several more basic questions: Where does the data come from? Who is responsible for maintaining it? Are definitions consistent across systems? How are missing values handled? Are changes recorded and traceable?

These tasks may not appear as advanced as AI models, but they are an important foundation for scaling Smart Manufacturing.

In other words, a Smart Manufacturing data strategy should not focus only on “collecting more.” It should help data become increasingly reliable, connected, understandable, and reusable.

System Integration Does Not Mean Moving All Data Into One Place

When companies discuss Data Integration, another misunderstanding often appears: does all equipment and system data need to be moved into one massive database?

In practice, the purpose of integration is not necessarily to centralize all data in one location. It is to ensure that the information needed can be accessed and used when it is needed.

Different types of data may have different requirements. High-speed sensor data may need to retain large volumes of time-series information. Quality data may need to be associated with product batches. ERP, meanwhile, manages orders and enterprise resources.

What matters is whether the company has established a clear information architecture and understands which data need to be exchanged, which need to be retained long term, which need to be available in real time, and how different systems identify the same equipment, products, and work orders.

NIST’s Smart Connected Manufacturing Systems Group also identifies heterogeneous data fusion, standards-based digital threads, and interoperability in connected smart manufacturing systems as important research areas. The goal is to help manufacturers use data from different sources to create knowledge and improve decision-making.

Therefore, the core of Data Integration is not “building the largest database,” but enabling data to cross existing system boundaries and support operations.

Looking Back at the Previous Three Topics, Why Is Data Integration Critical to Smart Manufacturing?

If we look back at the first three problems in this Smart Manufacturing series, all of them ultimately lead back to data integration.

When manufacturers want to improve production efficiency, they need to connect equipment downtime, Cycle Time, work orders, output, and scheduling information in order to identify the real bottleneck.

When manufacturers want to reduce quality abnormalities, they need to connect process parameters, equipment status, material batches, and quality results in order to analyze process variation.

When manufacturers want to improve equipment reliability, they need to combine Sensors, historical failure records, maintenance data, and production schedules so that Predictive Maintenance results can actually be translated into maintenance action.

Smart Manufacturing, therefore, is not a collection of isolated technology projects.

IIoT, MES, ERP, AI, Machine Vision, Predictive Maintenance, and Digital Twin technologies all ultimately depend on data exchange and system integration to create more complete operational information.

This is one of the key barriers manufacturers need to overcome when moving from “point-by-point digitalization” toward true Smart Manufacturing.

How Did Ford Otosan Yenikoy Use a Unified Data Architecture to Enable Real-Time Data Flow?

The World Economic Forum’s (WEF) Global Lighthouse Network also includes examples of manufacturers moving from isolated digital applications toward more complete Connected Manufacturing.

Ford Otosan’s Yenikoy plant in Turkey faced challenges including global supply-chain disruptions, increasing demand for customized commercial vehicles, and market volatility. To manage greater production complexity, the plant established a Fully Connected, Data-Driven Value Chain and introduced more than 60 internally developed digital solutions.

These applications included IoT, AI, Machine Learning, and Digital Twin technologies, but one particularly important foundation was the use of a Unified Data Architecture to enable real-time data flow.

According to the WEF’s 2026 Global Lighthouse Network data, Ford Otosan Yenikoy’s transformation resulted in Production Volume increasing to twice its previous level, production complexity increasing 12-fold, Labor Productivity improving by 44%, and Quality improving by 6%.

These results were achieved under Ford Otosan’s specific factory conditions, product mix, technology architecture, and transformation environment. They do not imply that every manufacturer implementing a Unified Data Architecture will achieve the same level of improvement.

What makes this case especially relevant is that the company did not treat IoT, AI, Machine Learning, and Digital Twin technologies as separate, independent projects.

When different digital applications are built on a shared architecture that supports real-time data flow, equipment, production, and management information have a greater opportunity to cross system boundaries and support more complex manufacturing environments.

This illustrates that what Smart Manufacturing really requires is not “the largest number of digital tools,” but the ability for different digital tools to work together within the same operational system.

System Integration Must Also Address Cybersecurity and Data Governance

As equipment, MES, ERP, Cloud, AI, and other systems become increasingly connected, manufacturers gain access to more real-time information. At the same time, however, the number of connections between data and systems also increases.

For this reason, Smart Manufacturing cannot focus only on Connectivity while ignoring Data Governance and Cybersecurity.

For example, which equipment data can be accessed by which systems? What access rights do different users have? Can changes to data be traced? How should appropriate security boundaries be established between equipment networks and enterprise networks?

The purpose of data integration is not to allow everyone to access all data. It is to ensure that the right people and systems can access the information they need under the appropriate conditions.

NIST research on Smart Manufacturing Information Governance also notes that manufacturing environments need appropriate Information Governance to support trust in systems and information.

Therefore, when manufacturers eliminate information gaps, they also need to establish data responsibilities, access rights, standards, and governance mechanisms at the same time.

Being “connected” is not enough. Companies also need to know whether the data can be trusted, who is allowed to use it, and how it can be used securely.

Where Should Manufacturers Start With Smart Manufacturing Data Integration?

For many manufacturers, system integration may sound like a much larger project than implementing a single piece of equipment.

However, companies do not need to integrate every factory and every system from the beginning.

A more practical approach is still to start with one clearly defined operational problem.

If the biggest problem is that managers cannot see production output in real time, the company can first establish connections among equipment status, work orders, and actual output. If quality issues are difficult to trace, it can begin by integrating quality results with key process parameters. If equipment downtime frequently affects delivery performance, equipment status, maintenance information, and production schedules can be connected first.

The next step is to identify which data are actually needed to solve that problem and where those data currently reside.

Which data come from equipment? Which are stored in MES? Which require ERP? Do different systems use the same identifiers for products, equipment, and work orders? How frequently is the data updated?

Once these questions become clearer, the company can then decide whether to use APIs, IIoT Platforms, Data Platforms, Cloud, Edge Computing, or other technologies to establish the required data flow.

In other words, the sequence of system integration should not be “connect all systems first and then see what we can do.” It should be:

First determine which operational problem needs to be solved, and then decide which data need to be connected.

This is fully consistent with the core logic of the entire Smart Manufacturing series.

The Key to Smart Manufacturing Is Not More Systems, but Making Information Flow

Smart Manufacturing is often associated with AI, Robotics, Digital Twins, IIoT, MES, Cloud platforms, and automated equipment.

However, if every technology remains isolated within its own system, equipment data cannot be connected with work orders, quality information cannot feed back into the process, and maintenance data cannot be integrated with scheduling. In that case, even a company with many digital tools may still need people to manually organize information before decisions can be made.

True Smart Manufacturing requires technologies to become connected.

From production efficiency in the first article, to quality stability in the second, and equipment maintenance in the third, all three topics point toward the same fundamental issue: manufacturers need greater transparency into what is happening on the shop floor and need to use data to identify problems and take action more quickly.

For this reason, the maturity of Smart Manufacturing should not be measured only by “how many technologies have been implemented.”

What matters more is whether equipment, production, quality, maintenance, and enterprise-management information can gradually form an information flow that supports operational decision-making.

Recent World Economic Forum Global Lighthouse Network research has also reflected this direction. The 2025 research noted that the volume of data captured and stored during digital transformation can increase by two to three orders of magnitude, making effective integration and management of technology and data increasingly important. By 2026, WEF further emphasized that leading manufacturers are moving from isolated technology applications toward embedding AI and other digital capabilities directly into core operations and value chains.

For manufacturers that are just beginning their Smart Manufacturing transformation, this does not mean they need to immediately build a fully integrated smart factory.

A more practical approach is still to start with a real operational problem: first make the required data visible, then connect the relevant information, and only then gradually use analytics, automation, and AI to improve decision-making speed.

Because the ultimate problem Smart Manufacturing needs to solve is not “What new technology is the factory still missing?” but rather:

Can the company understand more quickly what is happening, why it is happening, and what should be done next?