Enterprise System Implementation and Use

At Bryant Manufacturing: An Analysis of ERP Fits and Misfits

Refereed Research Paper

Submitted to the Enterprise Systems and Applications Track

DSI Conference, San Diego 2002.

March 6, 2002

ABSTRACT

This paper is a case study and analysis of the implementation and use of an enterprise system (ERP) at a manufacturing company. The paper describes issues of fit between the ERP system and business strategy and business processes, and it applies a theoretical framework for analyzing these issues.

submission number is ERP-737

1. Introduction

This paper is a case study of the implementation and use of an enterprise resource planning system (ERP) at Bryant manufacturing company.[1] The focus is on the relationship between the company’s competitive strategy, several of its key business processes, and the implementation of the ERP in the Harrisburg plant.

Companies have experienced mixed results with Enterprise Resource Planning (ERP) systems: Along with some inspiring success stories come a number of conspicuous failures. In between these two extremes are many ERP implementations that are partly successful and partly problematic. The case at hand can be best described as occupying this middle ground. According to many of its employees, one reason why the Bryant ERP implementation is not producing more satisfactory results is due to a number of ERP-mandated changes to its pre-ERP business processes.

Problems (perceived and real) due to ERP-driven process changes have received some attention in other papers. However, this paper attempts to extend and add to the academic community’s understanding of this issue. It is fairly well-documented that, as packaged systems, “off the shelf” ERP systems, such as SAP and JDEdwards, limit the variety of business processes that can be modeled (and thus carried out) to the process options available within the software (McAffe, 1997; Somers & Nelson, 2001). For example, Scott and Kaindl (2000) estimate that the average organization finds that 20 percent of its existing processes cannot be modeled by its ERP system “off the shelf.” However, it is also important to recognize that decisions made by the firm during the ERP configuration process further limit (and often standardize) the manner in which particular business processes can be carried out within individual plants and other ”sub-units” that make up the organization. Often this type of intra-firm process standardization is considered a benefit. In fact, it is often cited as an objective of installing ERP (Cooke & Peterson, 1998; Mabert, Soni, & Venkataramanan, 2000). However, in a limited number of cases, this type of company-level process standardization creates problems for one or more sub-units that are forced to change processes (Gattiker & Goodhue, 2000; Stedman, 1999).

Many academic and practitioner authorities recommend that when firms are faced with misfits between existing business processes and ERP-driven ones, the business process should be changed (e.g. Connoly, 1999; Wilder & Davis, 1998; Curran & Keller, 1998)).. However, like most prescriptions, this one is likely not universally valid. Certainly it is possible to modify ERP code—albeit with some cost and considerable risk. Furthermore, when ERP-mandated intra-company process standards are the issue, it may be possible to design sophisticated configurations that allow individual sub-units to conduct processes in idiosyncratic ways, but still maintain the benefits of an integrated information system (Goodhue, Wybo, & Kirsch, 1992). However this too may be costly. Thus one important research question is, when existing process cannot be modeled under a companies chosen ERP configuration, under what circumstances should one the business process be changed and under what conditions should it be maintained (perhaps with the consequence of modifying the ERP system)?

This paper provides a rich description of ERP-mandated process-change issues in the engineer-to-order/make-to-order context. It analyzes several affected processes individually, paying particular attention to the business issues involved. The case suggests at least one answer to the research question: Some processes impact directly on a plant’s ability to deliver order winners. They are directly related to business strategy. Changing these processes is likely to be counterproductive if the changes decrease the plant’s ability to deliver order winners. By contrast, profitable ERP-driven process changes are those that increase the alignment between business strategy and the process.

2. Methodology

During late 1999 and early 2000, the authors interviewed (face to face) 13 employees from Bryant Company’s Harrisburg Plant and company headquarters (table 1). An initial telephone interview and follow-up interviews with the plant manager were also conducted. All interviews were recorded and transcribed. We also reviewed some documents.

Table 1: Employees Interviewed
Assistant Plant Manager / Rear Frame Department Supervisor
Production & Sales Engineering Supervisor / HR Manager
Sales Engineering Supervisor / HR Administrative Person
Manufacturing Engineer / Head of MIS
Materials Manager / Vice President, Engineering
Senior Buyer / Senior Vice President, Finance
Inventory Control Supervisor

3. Company Background

Based in Harrisburg, Pennsylvania, Bryant makes and sells electric motors used mostly in industrial applications. With five plants in four states, it has revenues of approximately three quarters of a billion dollars. It is the industry leader in terms of market share. We focused on the corporate headquarters and on one manufacturing plant, which is also located in Harrisburg.

In many ways Bryant’s five plants are fairly autonomous – they accept and schedule their own orders, build and ship their own motors, and they purchase their own materials, without any involvement from divisional headquarter in Harrisburg. There is one group centrally that negotiates raw materials contracts with vendors, but those contracts are then released down to the plants for them to purchase from as needed. With one exception, the plants have their own sales and production engineering groups, which report directly to the plant manager, and dotted line to the VP of Engineering at division headquarters. Division Engineering has about 50 people who do major designs and major revisions for the model lines, but typically don’t get involved with the processing of orders at the plants.

3.1 Business Strategy

The marketplace for industrial motors is very competitive. Large customers sometimes buy 100 or more at one time, which can give them quite a lot of leverage. "Selling motors is like selling horses. It’s a negotiation. There's no set price. You get competitors coming in, and they're going to talk about price, and it depends on how low somebody's going to go." (Senior VP – Finance).

One of the ways Bryant has competed has been to allow customers to customize the design to exactly fit their needs. Two of Bryant’s plants each produce a single type of motor with relatively little variation in its specifications. The three remaining plants allow customers to specify significant design alterations. The Harrisburg plant, does the most customization.

Customization is a key order winner for the Harrisburg plant’s market segments; however, delivery time and price are important as well. "Our sales people will go out and bring in this motor that's a real doozie, and all of a sudden somebody's got to spend forever trying to re-configure the whole thing, and they'll price that motor at exactly the same rates as they do a cream puff going down the line." (Senior VP – Finance). Short delivery lead times tend to be especially important when the market is soft.

The high customization focus has many implications for the manufacturing environment. Design engineering is a key first-step in the order management process. Engineering often consumes a great deal of allowable lead time. Most of the fabrication and assembly is done to order. Similarly, many materials are purchased to order. Manufacturing systems must be flexible in terms of the plant’s ability to produce a wide variety output--and also in terms of allowing design changes to open orders.

4. ERP Decision and History

Around 1991 Bryant started looking at bringing in a “packaged” manufacturing system to improve efficiency in the manufacturing process and reduce spending on information systems. The then Director of MIS pushed the initiative very hard, but in about 1993, just as the initiative was coming to a decision, he left the organization. The next MIS director, a long time employee of Bryant, took over and carried the initiative forward. A major component of the benefits that sold the system was the expected reduction of headcount (primarily in engineering and MIS) by about 70 people. Though Y2K compliance of their legacy systems was a consideration, it was not the primary one, and in fact they eventually made their legacy systems Y2K compliant anyway because the conversion of all plants to the ERP would not be complete in time.

Bryant had just decided on a system to be implemented by Arthur Anderson, when they learned that within a year Anderson would no long support that system. After some restudy, Bryant began implementing J.D. Edwards manufacturing system in 1997, just about the time that they were purchased by CC Industries. Bryant's new owner supported the move to J.D. Edwards.

The new ERP system would have common software (identical code) across all the plants. It would also have a common central item master.

The goal was to keep customization of the J.D. Edwards programs to a minimum. Bryant picked the Harrisburg plant as the best place to first implement the new system, because Harrisburg handles the most variation in the specifications for its motors. The thought was that if they could get the system to work there, it would be relatively straightforward to roll it out to the other plants. Also a major factor was that two key players in the effort -- the divisional MIS group and the divisional engineering group – were just down the street from the plant. The assumption was that any customization for Harrisburg would be useful to all plants.

Bryant spent a significant amount of time working though their business processes using a “conference room pilot” approach (i.e. bringing in knowledgeable managers and working through individual business processes in depth to see how the system would handle them.) Conference room pilots are intended to identify any problems or gaps between the required business processes and what the system can support. Decisions can then be made about changing business processes or reconfiguring or customizing the system. The conference room pilot was staffed by a 50-50 mix of IS and “business-side” employees. The MIS people were most senior IS staff. Participants on the user-side were middle managers.

5. Partial Overview of Order Management and Related Processes

Salespeople solicit orders. Pricing for standard motors comes from a price book with possible input from the centralized discount authorization group. A central pricing staff does pricing for non-standard motors. Sales and potential sales receive a booking number from sales engineering. Once there is an actual order, Sales Engineering reviews the specifications to make sure pricing is correct, to catch any engineering issues, and to do preliminary scheduling. then moves on to Production Engineering. Production engineering checks orders for problems from a design point of view, and they perform another pricing check. Managers from Production Engineering suggest all orders receive at least some modifications in production engineering due to the complexity generated by the diverse models and combinations. Corrections in production engineering often require changes in other areas for example in accounting or in materials management if parts have already been relieved. When production engineering releases orders (once a week), they are processed by the MRP system. Purchased materials are ordered by buyer-planners. Production orders are generated and released to the shop floor.

6. Issues and Impacts

The Harrisburg plant serves a niche where high customization and flexibility are order winners. Delivery lead time is an order qualifier when the market is strong, and it sometimes is an order winner when demand softens. This positioning has many implications for the manufacturing environment. Design engineering is a key first-step in the order management process. Similarly, many materials are purchased to order. Engineering and procurement can consumes a disproportionate amount of allowable lead time. Most of the fabrication and assembly is done to order.

Over time, the plant had developed many capabilities and practices for coping with the most challenging aspects of this business environment. Many of these practices center on compressing lead-times for design engineering and procurement work so that adequate manufacturing lead-time is preserved. Maintaining flexibility in materials and order management systems is another key characteristic. The JD Edwards system imposed some requirements on the plant that conflicted with some of these practices. On the other hand, the software increased the alignment between some business processes and the plant’s competitive environment. The following sections describe several examples of these issues.

6.1 Informal Systems

Two informal systems that the plant had evolved to facilitate the deliver of high customization with short lead times are its “per print” system and a culture and procedures that facilitated utilizing its employees’ formidable knowledge-base.

6.1.1 “Per Print” Parts

With the degree of customization the plant builds into its products, many of the components are either purchased or fabricated specifically for a given order. One way Bryant had traditionally responded to the need for customization is through “per print” parts.

To implement a new part formally requires developing engineering drawings, assigning a number to the part, putting it on the item master file, generating the required routings, and sending all that material through costing to get the price. This involves considerable lead time (as much as 10 days), all before the order can be released to manufacturing. Instead of that, for many customized parts, the sales engineering group has taken a copy of a drawing for an existing part that is similar and marked it up with correction tape and pen and ink, assigned it a special number, and sent it directly to manufacturing. These parts are called “per print” parts. The “per print” drawings are kept on file (or ultimately on microfilm), and the “per print” number is associated with the order. The Harrisburg plant typically had 200 to 300 “per print” parts per week.

These “per print” parts are invisible to the manufacturing information systems, and cause headaches in pricing the motors, but have allowed Bryant to rapidly manufacture what the customer wants.