Some would suggest Operational Intelligence (OI) is simply derived from two applications: Business Intelligence (BI), to support the high-level decision-making that spans beyond manufacturing data sources, and Enterprise Manufacturing Intelligence (EMI), to provide the real-time floor-level data for the factory people. The question I would like to discuss is: “Is this a myth or a reality?”
On the surface, this statement appears to make sense. After all, EMI vendors will tell you that EMI has the ability to visualize plant data in real-time, directly from Automation equipment or Manufacturing Execution Systems (MES). These systems focus on accessing various data sources in the plant and rendering their data into dashboards that can be analyzed in near real-time.
On the other hand, BI vendors will tell you that Operations Intelligence is really a subset of Business Intelligence (this was discussed in a previous blog entry). Offline tools can drill down and report across various functional domains, which is then aggregated as Business Intelligence. Vendors offering this capability focus on statistical analysis, comparison, trending and even predictive analysis. While manufacturing data is often included, it tends to be aggregated up to the ERP level, so not visible in real-time.
It might seem that combining EMI and BI provides a complete solution. Let’s now put this approach to the test.
The Data Consistency Test
Consistent master data management and data contextualization is always going to be difficult with multiple applications. While it’s true that MES does provide certain EMI functionalities, such as aggregation, contextualization, visualization and propagation (to ERP), in reality, functional redundancies in your MES and EMI systems can lead to data integration and consolidation challenges. To make matters worse, what about data from the warehouse and other functional sources beyond the production line? How will that data be captured and integrated?
Moving beyond a specific plant, there is often the challenge of integrating multiple MES applications and instances, implemented in various locations at different times. This labyrinth of IT systems can make data aggregation and master data management at the enterprise level a serious challenge.
Therefore, your OI solution must achieve data consistency across each of these scenarios, suggesting OI should be viewed as operating somewhere within unified and widely deployed Manufacturing Operations Management systems. Standard-fare EMI is not up to the task.
The Process Management Test
Among other components, OI provides Business Process Management to perform model-driven execution of business processes. Some MES systems provide EMI and BPM-like tools, but most are limited to certain manufacturing processes, tending to marginalize warehouse and logistic operations. Some ERP systems offer Intelligence bundled with BPM tools, however these are often limited to modeling back office workflows or providing systems orchestration at best, staying far from actual plant-floor operations.
The Truth Test
Multiple, disparate applications inevitably deliver different versions of the “truth,” with factory staff only seeing their data (in EMI) and corporate executives only seeing their data (in BI), leading to an inevitable disconnect. Since EMI only provides insight into manufacturing processes, BI will typically be used to analyze logistics and other “shop-floor” processes, leading to analytical schizophrenia.
In theory, you can combine EMI + BI to achieve a basic OI solution. But, it would be difficult, and the potential for data inconsistency is big, as are the challenges to achieve and sustain seamless integration. To overcome these challenges, manufacturers should consider thinking more holistically, viewing OI as part of a broadly implemented BPM-based platform for manufacturing operations management. Attributes of this type of approach includes:
- A common master data management architecture, shared by manufacturing operations across all locations and functions, to solve the challenge of isolated silos and complex integration
- Data captured and measured consistently, as it is embedded within the manufacturing operations platform, at all the core MI functions
- Consistent and flexible process modeling and execution including data collection and measurement covering processes in and outside the MES or automation layer, such as warehouse, quality or maintenance operations
- Visibility from different perspective (levels), but based on the same data, so not decoupled from manufacturing operation’s granularity
- Seamless, consistent access to both operational and financial data, from manufacturing’s perspective
Each of these capabilities are essential for an Operational Intelligence (OI) solution, but cannot be met by simply combining Enterprise Manufacturing Intelligence (EMI) and Business intelligence applications. A broader approach with a wider perspective can better deliver the necessary performance and operational excellence that enterprise manufacturers now require. The myth that EMI + BI = OI has been busted!