14 February 2008

Enterprise decision making

By Jim Pinto

Consider how managers made decisions just a few decades ago. Information was gathered through a variety of mechanisms in isolated sections and departments. This was passed up a human chain, usually in a financial reporting timeframe: Monthly, quarterly, and annual information, to facilitate analysis and adjustment. So, the job of a manager was to shuffle old, and in many cases outdated, reports.

With the advent of the computer, some of the calculating and reporting mechanisms improved, but the time frame remained centered primarily on financial control mechanisms. Decision making was always a slow and laborious process, involving many people sitting around a table discussing printed reports. The reports were filtered and summarized through local decision making successively at each level, culminating in top-level review and approval in the executive suite. There were little or no inter-relationships between different activities and locations. All multi-location, enterprise-level decisions came in the same slow, lumbering timeframe—usually completely out of sync with the real needs. But, faster action seemed impossible; perhaps even undesirable.

You can compare these control mechanisms to feedback in a control loop. When measurements and control mechanisms change too slowly, the result is a series of oscillations that worsen the control rather than improve it. Using this PID-control model, conventional management theorists developed feed forward and other mechanisms to improve performance. But there were few results; real-world problems were too complex.

Today, Web Services, Supply Chain Management, Customer Relationship Management, Enterprise Application Integration, and a plethora of other software tools and services are available for enterprise integration. Measurements can be made in real time, and information from all segments of the enterprise, including multi-location plants and factories, can integrate to facilitate optimal decision-making at all levels.

The next, natural step is to make production and business systems work together for factory automation and process manufacturing. The same connectivity and interoperability benefits users expect from selectable printer drivers for word-processing have become available for measurement and automation systems. There are many standard tools available to enable connectivity and interoperability of plant floor information between disparate fieldbus networks, programmable controllers, distributed control systems, plant assets, and production management systems.

Integrated applications can access data the same way, whether it comes from a server connected to a PLC network, an industrial network such as Foundation Fieldbus, Profibus or DeviceNet, a SCADA system, or a production management system. This provides access to real-time plant-floor information, production, and business applications across the manufacturing enterprise in a consistent manner. Multi-vendor and multi-location interoperability is a relatively simple extension with web services.

Users can now have a seamless exchange of production and equipment status information across plant floors and multi-plant locations, directly up to the management level. And real-time decisions can be made in response to real-time changes—in some cases in as little as a few seconds to an hour.

Related links:

• InTech, December 2001, “Thar’s Gold in That Plant-Floor Data”
  www.isa.org/link/GoldData2001
• Enterprise Decision Management
  http://en.wikipedia.org/wiki/Enterprise_Decision_Management
• Are Web services the future of supply chain computing?
  http://www.mmh.com/article/CA512153.html?nid=2409&rid=1123551107

Behind the byline

Jim Pinto is an industry analyst and founder of Action Instruments. You can e-mail him at jim@jimpinto.com or view his writings at www.JimPinto.com. Read the Table of Contents of his book, Pinto’s Points, at www.jimpinto.com/writings/points.html.