Ask almost anyone what the term “business intelligence” means and you’ll likely get a fairly standard answer — that it’s the process used to aggregate and analyze information so people can make better business decisions. That’s a broad definition, however, and if you drill down for more information, the finer details cloud the picture.
For example, business intelligence (BI) is different from data warehousing, which is little more than data storage. It’s also related to, but different from, data mining, which analyzes data to let you spot trends and draw conclusions. Reporting tools deliver information in a timely fashion, allowing users to generate reports or queries in order to get specific details. On-line analytical processing (OLAP) analysis adds even more dimensions, allowing you to compare and contrast information against time or other factors so you can uncover trends. Scorecarding tracks your key performance indicators (KPIs) and alerts you if pre-determined thresholds have been crossed, and executive dashboards give you a convenient way to add context to information so it’s more easily understood.
But even if you know and understand all of the foregoing, those definitions really only apply to the business systems in an enterprise, not really to the plant floor operations of a manufacturing company. On the production lines and in the plant front office, there’s yet another version of BI that’s often referred to as “enterprise manufacturing intelligence,” or EMI. This sounds a lot like business intelligence, since they both refer to gathering and analyzing information from multiple applications — but based on their deployment they’re quite different.
Business intelligence has historically involved transaction-oriented business applications, such as enterprise resource planning (ERP), supply chain management (SCM), finance, and enterprise asset management (EAM). As critical as BI solutions are for helping management run a company better, they can’t provide all the intelligence needed to manage manufacturing operations optimally because they cannot deal with the different protocols and formats used for the time series data generated by plant floor systems. For that you need EMI applications, which similarly aggregate information from multiple applications and maintain key relationships between different data elements in them — but which are targeted to plant floor production data sources rather than enterprise applications.
The ultimate problem is that no matter which “intelligence” solution you look at, you can’t mix the two worlds. It’s difficult to synchronize time series data from real-time production operations with the transaction-oriented world of enterprise applications. The availability of these parallel types of systems still hasn’t solved the problem of integrating “silos” of information within companies.
In the past, people have customized their own linkages using specialized adapters and connectors to extract data from control systems and feed it to the business systems — but these require extra cost, specialized programming and hundreds of hours to implement, and their total cost of ownership is high because of the added maintenance needed. Until now, there has been no off-the-shelf way to combine the two in order to aggregate all data, distribute it easily over the Internet and overlay it with the business context that makes it more understandable.
Business intelligence software, such as Incuity EMI from Incuity Software, in Mission Viejo, Calif., is designed specifically for manufacturing companies. These types of BI solutions accommodate both the time series data that describes plant and equipment behavior and the transactional data found in traditional business systems. Users do not have to generate more data but can better understand and deploy the information they already have in the myriad applications dispersed throughout a plant.
Information can be accessed over any network, using standard browsers, and processed as requested, allowing users to analyze data from within the application software they already use to do their jobs.
Data from disparate sources can be quickly collected and aggregated, and actionable information delivered in interactive reports and dashboards. Using this information, manufacturers can recognize opportunities and problems earlier, react swiftly, easily communicate the issues and make decisions that will enhance bottom line profitability.
BI software systems can also simplify the presentation and analysis of information without users needing to learn new tools and skills. The systems can produce intelligent reports that offer alert detection through live data connections and can automatically drive user attention towards exception conditions, missed targets and plan deviations. All of this information arrives via the applications that are already very familiar to end users, such as Microsoft Excel and Internet Explorer (or any other standard browser.Security is also important for a system that can connect to so many different data sources. A user should be able to impose security on any item in the model and specify exactly who is allowed to read, write, access or change any of the data coming in from the different data sources.
A complete business intelligence system for the manufacturing environment should allow the user to access all of the data from various data sources, put it in better context to turn the data into real information, and allow viewing and interaction with the information to make better decisions that can help manufacturers be more productive and profitable.