Introduction

Real-time analytics (RTA) and the Industrial Internet of Things (IIoT) remain drivers for contemporary manufacturing, promising enhanced efficiency, predictive maintenance, and dynamic decision-making. While technology vendors often position such solutions as Plug&Play, implementing IIoT-driven analytics in an actual production environment can become a complex undertaking. Systems designed to capture and process vast amounts of sensor data in real time often conflict with the multi-dimensional realities of shop floor operators and engineers.

In the end, real-time analytics will either succeed or fail based on how well it aligns with user needs, business objectives, and existing technical infrastructures. This article presents the integration of IIoT and real-time analytics through the lens of an ethnographic case study in a Swedish manufacturing facility. The case shows that strategic system architecture, harmonized data structures, and user engagement are central to ensuring that RTA brings tangible, sustainable benefits but does not cause unnecessary disruption.

The Promise of IIoT in Modern Manufacturing

Real-time analytics depend on interconnected machines, all working in tandem, to deliver insights on production processes, machine health, and overall efficiency. The idea of the Internet of Things, first given by Kevin Ashton in 1999, has gone through an evolution of its own in industrial contexts and culminated into what we now understand as Industry 4.0. It uses the connectivity of machinery, sensors, robots, and other devices to enable advanced data collection and data sharing across a manufacturing ecosystem. IIoT then adds real-time data processing on top to handle immediate reactions to fluctuations in production performance, resource availability, or unexpected machine failures. In other words, the goal is to have an industrial process that is more transparent, predictive, and efficient, where data-driven intelligence underpins critical decision-making.

Yet, even when the technological advances point toward clear benefits, practical barriers often arise. Many real-time analytics systems are implemented without full consideration of existing workflows, older enterprise resource planning systems, or shop floor skill sets. Raw data, which does not provide actionable insights, can overwhelm operators and technicians, while advanced analytics platforms introduced by IT departments sometimes have little regard for shop floor context. Therefore, the gap between sophisticated IIoT architectures and real-world operational needs remains one of the central challenges for any organization looking to modernize its manufacturing processes.

Real-Time Analytics: A Tool for Rapid Decision-Making

One key battleground lies over the definition, collection, and interpretation of data. An up-to-date RTA system may monitor events-press or press stoppages, for instance-only in terms of discrete cycle times or sensor states, while a legacy ERP would estimate production times according to nominal tasks that include both direct manual and machine-operated tasks. When these various sources of data are not aligned, operators and managers are left with partial or conflicting views of what is taking place on the shop floor. Without a coordinated approach to harmonizing data structures and definitions, analytics mislead more than they inform, hindering the possibility of real-time, data-driven decisions.

In theory, real-time analytics allows the owner immediate insight into subtle changes in operating conditions, such as temperature, vibration, or unexpected power consumption-all possible precursors of impending equipment failure or divergence in process efficiency. Detection in real time permits the manufacturer to take corrective action before damage has occurred or the costs are substantial. As powerful as this type of analytics technology is, for the smooth integration into an operation environment, it should augment rather than disrupt existing production routines.

Implementation Challenges: Many researchers and industry experts now recommend a design-in-use methodology to address the subtleties of IIoT implementation. Traditional approaches often regard system design and system use as two separate phases, assuming that once a technology is developed, the rest is easy deployment. Design-in-use, by contrast, emphasizes a continuous engagement between the system’s evolving configuration and the practical realities faced by its users. The method covers not only what hardware and software is necessary for successful implementation but also focuses on worker-machine interactions that happen daily. Operators understand where the bottlenecks are, anomalies that presage larger issues, and how to effectively embed real-time information into SOPs. Absent this, advanced analytics risk underutilization and improper deployment.

Cultural and organizational factors, at a deep level, may exacerbate technical issues. A new digital tool may require new skill sets or changes in production routines, which can be resisted by workers. On the other hand, existing systems may depend on data definitions different from those used by real-time analytics platforms, which calls for re-categorization of production events and even re-education of personnel. No matter how smart a piece of software or powerful its algorithms are, its effectiveness is always bound by the human context in which it has been set. For this reason, a design-in-use methodology emerges as a practical framework that ensures technology evolves with the workforce, adjusting to real-life constraints and feedback rather than enforcing rigid, preconceived ideas.

The Design-in-Use Approach

In a Swedish manufacturing plant run by Siemens Energy, the authors seek to explore the deployment of an X-top real-time analytics system. The plant manufactures combustion chambers for turbomachinery, requiring long production cycles involving skilled manual labor and variable scheduling demands. This environment represents a particularly interesting test bed for real-time analytics because of its mixture of advanced machinery and labor-intensive processes. The introduction of X-top was intended to harness real-time data to reduce unplanned downtime, facilitate predictive maintenance, and help operators make swift, informed decisions about daily work tasks.

Although promoted as a Plug&Play solution, the system actually took close to a year to be integrated cautiously with the plant’s current ERP (SAP). X-top was theoretically tracking machine cycle times, automatically logging disturbances or deviations to create a digital record operators and engineers could consult to identify problems and refine processes. But SAP used different metrics and assumptions. While SAP expected certain manual tasks, X-top focused strictly on machine cycle times. Because neither system matched the other’s data definitions, operators soon found themselves having to translate, reconcile, or manually update information. The mismatch reduced the effectiveness of real-time analytics and created confusion on the shop floor.

The Swedish Case Study

One of the complications was that the project was from the bottom up-that is, driven by operators and engineers rather than ordered by upper management. On the one hand, this meant that the end users of the system were heavily involved in the development process, which could make for a much better fit when the solution went live, addressing real problems. On the other hand, that meant a slower rollout, little organizational support, and there were sporadic delays whenever technical expertise had to be sought from the outside. As the ethnographic study showed, this is because design-in-use requires iterative work: the research team worked with the plant staff to diagnose problems, fix software updates, redefine stop codes, reconfigure data elements, and then reconfigure them again -.

Although some difficulties emerged in tuning real-time data with production data, several positives also emerged as operators gained deeper knowledge of the machine’s status and performance. X-top enabled them to visualize patterns that previously went unnoticed, such as repeated disturbances of a particular type or persistent gaps between planned and actual cycle times. By iterating on how data was collected and displayed, the plant gained insights that allowed for more precise troubleshooting and better communication between different teams. New roles within the workforce also emerged, such as local system administrators who could manage the analytics dashboard, update deviation codes, and act as points of contact for continuous improvements.

Lessons for Architecture, Data Structures, and Alignment

Three clear lessons surfaced from the case study. First, choosing and setting up the right system architecture is a little more complex than what Plug&Play slogans may suggest. Robust architecture needs to seamlessly integrate real-time data streams, enterprise data, communication protocols, and user-facing dashboards. Omission of any of these aspects threatens to leave the analytics system half-configured, with critical data points stuck in some sort of non-compatible format.

Second, data structures have to be harmonized to ensure that real-time data makes sense when combined with legacy planning systems. When a system like SAP relies on scheduled machine times but an IIoT platform measures actual machine cycle times, the result can be confusion or outright mistrust of analytics. Bridging that divide requires more than a technical fix. It requires communication between IT, operators, engineers, and managers to establish common terminology, ensure consistency in measurement criteria, and clarify how the data will be used.

Third, business objectives and user needs have to be kept aligned. Companies are more often investing in IIoT for its potential impact on return on investment, overall equipment effectiveness, or predictive maintenance goals. Operators, on the other hand, tend to focus on whether the system simplifies their work, reduces paperwork, and helps them solve real problems such as machine stops or incomplete job orders. Only through the convergence of these perspectives-through consistent user engagement and a feedback loop-can an IIoT-based real-time analytics solution become a real asset rather than an underutilized tool.

The Role of Design Ethnography

The research team adopted the design ethnography approach, which differed from traditional ethnography insofar as it involved active deployment and adaptation of technology. Rather than observing the workers’ behavior, they became collaborators, joining workshops and data configuration sessions, thereby suggesting incremental improvements. Following this approach, they detected the different sorts of embedded knowledge that these operators had. Workers had a nuanced understanding of how deviations occurred, which machine anomalies demanded immediate attention, and how to code or categorize these issues in a digital system.

Ethnographic design allowed us to understand that real-time analytics cannot remain static. For RTA systems to provide long-lasting value, they must be able to accommodate changes in job orders, changing hardware configurations, and even evolving production lines. Only when operators are empowered to fine-tune the system-to reduce confusion and make sure analytics align with everyday concerns-can the system truly pay off. Without ownership of the system, it would likely become deadlocked and generate dubious results.

Conclusion

Alignment of business and user values de facto requires an iterative process firmly based on operational realities, as well as strategic objectives. The design-in-use perspective has emerged as a particularly practical model to pursue such alignment, as design and use become intertwined in a single, ongoing process. Finally, empowering operators and technicians to take an active part in system implementation creates a culture of shared responsibility and adaptability that bridges the gap between the powerful capabilities of real-time analytics and the intricate realities of shop floor operations.

These insights are transferable to many industrial settings that aspire to harness IIoT and RTA in a meaningful, sustainable manner. The importance of a co-evolving relationship between technology and its users will become even more apparent as future research digs deeper into system incompatibilities and the refinement of predictive models.