Transforming plant data into business intelligence |
 |
By Beth Stackpole, SearchManufacturingERP.com contributor
05 May 2009 | SearchManufacturingERP.com |
|
|
The advent of business intelligence (BI) software and enterprise performance management (EPM) frameworks changed how companies proactively monitor and maintain the financial health of their business. Now, an emerging category of software -- known as manufacturing intelligence (MI) or operational performance management (OPM) -- is helping manufacturers apply similar, actionable metrics to production processes and plant-floor operations, all in an effort to gain better supply chain visibility and drive supply chain success.
The discipline known as manufacturing intelligence or operational performance management integrates and analyzes data from a variety of plant sources, transforming raw data feeds into actionable information. The data is packaged in the form of highly accessible exception reports and key performance indicator (KPI) dashboards, so that everyone from quality managers to plant floor executives can view information in context, allowing them to make informed decisions in a timely fashion.
By leveraging the visibility of such an operations-oriented decision support system, manufacturers can establish links between operations KPIs and critical business metrics. As a result, they gain insight into everything from asset utilization to machine uptime and plant-floor productivity while also monitoring energy usage, uncovering the cause of quality problems, and ensuring consistent production across multiple lines.
|
|
|
|
|
"This is what people do in daily meetings on spreadsheets or on crude fishbone diagrams on the walls in break rooms at manufacturing plants."
Julie Fraser
President, Cambashi Inc.
|
|
|
|
|
|
|
|
|
|
Several forces are coalescing to make this type of intelligence invaluable for manufacturers: shrinking product lifecycles, demand for build-to-order production, increasing emphasis on quality, and a growing complexity in the product mix. Factor in an aging manufacturing workforce poised for retirement, and companies can no longer get by without automating and codifying these core metrics.
"It's a matter of 'I don't know what to improve if I don't know what's out there,'" said Simon Jacobson, research director at AMR Research Inc. "Visibility is the big thing to get at, versus blindly thinking we need to improve everything."
Operational performance management or manufacturing intelligence software can create the visibility into business metrics that is either lacking or occurring haphazardly today. In a 2006 study conducted by the Manufacturing Enterprise Solutions Association (MESA) and Industry Directions, only 3% of the respondents reported effective links between operations KPIs and business metrics. Even two years later, this still means that most manufacturing managers lack the views that accurately represent plant contributions.
"This is what people do in daily meetings on spreadsheets or on crude fishbone diagrams on the walls in break rooms at manufacturing plants," said Julie Fraser, who was with Industry Directions when the study was conducted but is now president and principal industry analyst at management consulting firm Cambashi. "It's happening, but in time-intensive, error-prone, manual ways."
One reason why manufacturing intelligence has lagged behind BI in the enterprise has to do with the complex nature of plant-floor data. For one thing, while plant-floor resources such as Materials Handling Systems, SCADA, Programmable Logic Controls (PLCs) and Manufacturing Execution Systems (MES) generate tons of data, incompatibilities among all these sources have made it difficult to correlate data between systems, let alone establish the trends needed to create a complete picture of what's occurred.
Furthermore, while enterprise BI software systems primarily tap into transactional data collected and managed by a data warehouse, plant-floor intelligence is real-time in nature and requires a much closer connection to the actual data sources. Finally, while BI is conducted offline in batch mode, manufacturing intelligence requires a direct connection to the actual production process.
"Traditional BI you might do once a week or once a month," Jacobson said, "whereas manufacturing intelligence you do in real time on the line." He added that enterprise BI tools can't meet the challenge presented by manufacturing intelligence because they are not built for pulling data from multiple machines, whereas most manufacturing intelligence applications are configured with pre-built integrations for widely used PLCs and data historians.
Another big difference between BI and manufacturing intelligence (and other manufacturing applications) is that companies deploying operational performance management solutions should see quantifiable benefits in months or even weeks. Such companies are advised to look beyond tapping manufacturing intelligence for tactical initiatives, such as quick visibility into asset performance, and instead leverage the technology as part of a broader campaign such as lean manufacturing, total quality, continuous improvement or Six Sigma.
"You really do need to understand how you're doing in order to make the next steps, so [operational performance management] really needs to be part and parcel of these types of programs," Fraser said. "[Manufacturers] all grew up with manual systems -- the next logical step to get more productivity is to automate."
About the author: Beth Stackpole is a freelance writer who has covered manufacturing techniques and manufacturing technology extensively.
|
|
|
|
