Implementing an enterprise resource planning system can resemble a long ride on a pot-holed street. But with the model of pioneers like Analog Devices, the path to success can be smoother.
| Non-IT costs drove ADI's ERP implementation Hardware and software costs, including the initial R/3 license for more than 300 users and the Oracle database license, were lower than the cost of people, both internal staff as well as consultants ($ in millions). |
Source:Analog Devices Inc.
Fear, panic, and alarm were raging through the halls of Analog Devices Inc. back in 1993. The Norwood, Mass.-based chip maker was facing a bleak future.
In the early 1990s Analog Devices (ADI) was dependent on military projects for a large part of its revenues. Following the end of the Cold War, however, the company faced a decline in military markets. On top of that, its processes, systems, and staff were not prepared for the stiff competition, price sensitivity, and rapid changes in the obvious new markets it wanted to enter, such as cellular telephones, personal computers, and other electronic systems.
Company officials discovered, during a major rethinking of their business, that ADI had high costs, lackluster customer service, and inefficient systems and processes relative to the competition in these new markets. With less than $1 billion in revenue at the time, the companys operations were spread around the world, with manufacturing, warehousing, distribution, sales, and customer-service facilities in North America, Europe, and Asia. And ADIs sales, general, and administrative (SG&A) costs were 29% of revenue, while its competitors costs hovered in the low 20% range, according to company officials.
To avert disaster, ADI turned to enterprise resource planning (ERP) as part of a massive revamping of the company.
At that time, ERP implementations were not pretty. Between 1993 and 1995, a number of large organizations began ERP projects that consumed people, money, and time in ways IT managers had only seen in nightmares. Boeing, Alcoa, Chevron, and other organizations endured projects that in some cases cost more than $200 million and are still not complete, due in part to significant IT challenges. Yet some early ERP implementations were not only completed within a reasonable amount of time and money (given the immaturity of the technology), but also provided a model for successful IT management of an ERP implementation.
In fact, ADI's implementation of the R/3 ERP suite from Walldorf, Germany-based SAP AG beginning in 1993 and continuing through 1996 proved--at the time--to be a singular example of best practices in IT restructuring and management. Since then the company has enjoyed the substantial financial benefits of an ERP system, including a shortened order-fulfillment process, reduced inventory costs, and increased productivity, according to ADI officials. While ADI spent $21 million on the implementation--way beyond what it would cost these days--its accomplishment shows the importance of strong managerial direction and support, the selection of a global implementation team for a global project, and the value of centralizing an IT organization within a decentralized company.
Sizing things up
Business consolidation was the first order of the day at ADI--a major and traumatic undertaking. Numerous logistical inefficiencies had to be changed if the company was to successfully compete in commercial markets. Customers there demanded fast and certain delivery; ADI sales reps, however, could only guess when products would be shipped. Extra shipping steps required by the companys existing processes, along with the number of finished goods caught up in the lengthy shipping process, added to the cost as well as contributed to the delay.
Radical reengineering also was required within the companys financial organization. Its costs, as a percentage of revenue, were substantially higher than the competition, recalls Joseph McDonough, chief financial officer of ADI.
A single, integrated, order-management system providing real-time information around the world offered a way to improve customer service. So in addition to reengineering and restructuring the business, the company focused on its information systems. "With a single, integrated system, factories can ship directly to customers rather than to intermediate warehouses," says Gerry Dundon, now ADIs director of logistics.
This focus on information systems turned a hot spotlight on a classic host-centric IT organization. PCs, LANs, and client/server computing were new to the company. IT itself deployed a decentralized staff for each operating unit. Each corporate entity, for example, ran its own general-ledger system. There were eight host computers running 11 accounts-payable databases, three accounts-receivable systems, five purchasing systems, and six fixed-asset systems; ADIs three-building headquarters outside of Boston had five different accounts-payable systems. Reconciliation of the various systems was done on a PC using a spreadsheet, with numbers entered manually. The warehousing and logistics systems were equally fragmented.
"ADI had an incredibly ugly collection of systems with little commonality," recalls Jim Shepherd, vice president/research at AMR Research Inc., in Boston.
The splintered approach to systems deployment throughout ADI made implementation of the global business vision extremely challenging. In fact, company officials acknowledge that there was little coordination among the decentralized IT staffs. And ADI managers considered the IS organization to be unresponsive.
Not surprisingly, ADIs spending on IT was significant, about 4% of revenue, including IS department and business-unit spending. "Management didnt think it was getting the right amount of IT bang for its bucks," notes ADIs McDonough, "due to the lack of a strategic roadmap."
ADIs battle plan
As senior executives mapped out their strategy to remake the company, they decided that a better managed IT approach was necessary. ADIs executives elevated the status of the IS department to reflect the new importance placed on systems as the enabler of the new Analog Devices, and went in search of a seasoned professional to head that division.
"Information systems mission is to help create the new Analog," explains Larry Loh, who was hired as chief information officer in 1993. Loh replaced the old, decentralized IT structure with a more centralized approach, enabled by having all the operating units systems in one place on one system.
The transition to the new R/3 system was handled in two major pieces: order management and financials. Responsibility for implementing the order-management project fell to ADIs Dundon, then a program manager. The project team took the name GOLD, for global order management, logistics, and distribution. At its peak, the GOLD implementation group consisted of 42 employees and consultants, half of whom were full-time. Seamus Brennan, currently ADIs director of finance, led the financials project.
Gold Project Timeline
|Requirements and vendor analysis||May92-Dec.92|
|Vendor selection||Jan-April 1993|
|SAP selected||March 1993|
|Design||January -March 1994|
|Switch to R/3 2.1||March 1994-October 1994|
|Configure||May 1994 - December 1994|
|Testing||September 1993 - January 1995|
|Roll-out in US||February 1995|
|Roll-out in Europe||May 1995-February 1996|
|Roll-out in Asia||Early 1997|
Supporting each program manager was an IT manager. This person was selected from among the senior staff of the IT department. The responsibilities of the IT manager included the day-to-day management of the functional development teams assigned to each program. In addition, a program support council was established, comprised of the head of each functional team and the IT managers; a senior manager was added from the key software, hardware, and consulting service vendors after they were selected.
After an extensive review of available applications, in early 1993 the GOLD team chose R/3 from SAP as the platform for its new global order-management system. The product was still in its infancy, a version 1.x release, but ADI officials chose it because it came closest to meeting the companys needs. Later that year, ADIs finance reengineering team selected R/3 to support its efforts, mostly due to the obvious benefits of having a fully integrated logistics and financial system from one vendor. After evaluating several implementation partners, ADI hired Price Waterhouse, now a part of PricewaterhouseCoopers with U.S. headquarters in New York City, as a subcontractor, due to its knowledge of SAP software and UNIX.
Finance Project Timeline
|Project phase ||Timeframe|
|Vendor analysis ||March 1993 - September 1993 |
|Sap selected ||October - 1993 |
|Requirements analysis ||November 1993 - March 1994 |
|Prototype configuration ||Apil -May 1994 |
|Est, document, train ||May -September 1994 |
|Pilot system operational, debugged ||October - December 1994 |
|Roll-out in Wilmington, MA plant ||Jan-March, 1995 |
|Roll-out eastern US offices and factories ||April-September 1995 |
|Rollout in Europe (nine countries) ||Oct-Dec.1995 |
|Roll-out in US west coast ||Jan-March, 1996 |
|Roll-out in Asia (Japan, Taiwan, Philippines) ||April-September, 1996 |
|Worldwide consolidation ||October-December 1996 |
Among the companys early technology decisions, several were crucial. First, finance and GOLD wouldnt change R/3 source code, to avoid compatibility and upgrade problems down the road. However, some basic logistics needs were not supported by early R/3 releases. So the GOLD team wrote a series of subroutines that linked to existing R/3 modules to incorporate vital functions. Second, the GOLD team established an aggressive 18-month schedule to design, configure, test, and rollout version 1.x of R/3 in the United States and to ADIs factories around the world. That was to be the first phase of a three-phased rollout plan, with the European offices as phase two, and Asian offices in phase three.
Unfortunately, the plans were foiled within months, and GOLD began to tarnish almost as soon as it began implementation. The team struggled through a series of interim software releases that were not ready for production use. So when SAP released version 2.1D in 1994, ADI officials felt that it was enough of an improvement for the team to switch their development efforts from the prior release.
"This would never happen today. SAP releases are solid," notes Joshua Greenbaum, principal with Enterprise Applications Consulting, in Berkeley, Calif.
The GOLD rollout of version 2.1D for the U.S. factories and the overseas warehouses occurred in February 1995, followed by Europe in February 1996. The rollout in Japan followed in 1997 after the system was upgraded to version 3.0.
In contrast, the finance project team delivered its implementation of R/3 version 2.1D within its 12-month schedule, beginning with a pilot rollout in the Ireland factory in December 1994. Rollout of the finance project in the United States and Europe began shortly after that pilot began operation and was completed by late 1995. Financial modules were running at several ADI sites in Asia, including Taiwan and the Philippines, in 1996. The remainder of ADIs Far East locations followed the next year.
Not an inexpensive undertaking
Implementing a suite of enterprisewide applications around the world is not an inexpensive undertaking. ADI's estimated cost of $21 million to implement R/3 was 15% higher than the initial cost target, according to company officials, due to delays in development related to the shift to a newer version of R/3. Software and consulting fees, including Price Waterhouse as the primary implementation subcontractor, totaling $9.5 million amounted to almost half the budget, notes ADI. The R/3 software license also included the cost of the database software from Oracle Corp., of Redwood Shores, Calif., used by the applications.
While no analyst today questions ADIs expenditures or considers it extravagant, "we do not usually see price tags like this anymore," says Steve Bonadio, senior analyst at Hurwitz Group Inc., of Framingham, Mass. Streamlined implementation methodologies generally keep the costs down, except in the case of runaway disasters.
Part of ADIs cost was due to the expense of shifting legacy data to the new system. While AMRs Shepherd isn't shocked at the $21 million price tag in this situation, "it is not representative of what companies would spend today," he notes. A new crop of enterprise-application-integration tools, for example, can reduce the cost and time to interface or port legacy data and systems.(For more information on the current integration tool offerings fine-tuned for ERP, click here.)
Analysts from the now defunct firm Scalea and Co., of Boston, at the time estimated indirect costs at $8 million. This figure includes the value of the time spent by the ADI executives and operation staff to implement R/3, as well as contractor costs. The estimate also includes the time top managers spent on the project. "We did not hire specifically for the SAP project but used existing staff," adds CFO McDonough.
ADI handled the training very inexpensively. Only members of the implementation team were trained by SAP or its contractors. These members then trained ADI staff who were considered the local experts. They, in turn, trained other users. ADI also developed its own user and training manuals.
Success and its lessons
The impact of R/3 is inextricably woven together with business process reengineering (BPR). ADIs return on investment (ROI) in R/3 is a combination of the benefits accrued from the integrated suite of applications, along with its BPR efforts. Scalea analysts determined that the companys return on the R/3 investment over the first four years (mid-1995 to mid-1999) would be around $212 million.
Selling, marketing, and logistics functions were the largest contributor to the return. Over this four-year time frame, analysts at Scalea estimated a $162.5 million gain from reduced selling, marketing, and logistics costs. ADIs SG&A costs plummeted from a high of 29% in the early 1990s to around 15% in the 1998 fiscal year.
Other tangible gains included approximately $6.7 million in inventory savings over four years, due to improved inventory management. The Scalea analysts also projected a $38 million return from 1995 to 1999, due to streamlined financial processes, including chopping one week out of the monthly closing process. Soft, or indirect, benefits were not considered in the analyst projection, but they are nonetheless real and valuable: for instance, ADI now has many happy customers.
Looking back over the implementation, ADI managers took note of the IT-related decisions that contributed to achieving the high ROI. These include:
- Software selection. Choosing a package with functions that most closely resembled what ADI wanted to do turned out to be important. It reduced the amount of time required to implement the software and caused the least amount of disruption. As it turned out, SAP has emerged as the clear ERP leader.
- Planning and preparation. ADI prepared detailed planning documents addressing almost every IT aspect of the project. Although some went unused, the planning exercise itself was valuable, IT managers attest. Many organizations today still dont do effective upfront planning, which hurts the entire implementation effort, notes Hurwitz Groups Bonadio.
- Team management. ADI presented the R/3 project as a plum assignment, attracting top-notch people. "You cannot afford not to have your best people on this," says ADIs Dundon. Liberal incentives and on-going team-building efforts kept turnover at zero during the key rollout period.
- Communications. ADI embarked on an extensive communications program, including a monthly newsletter, to manage expectations and overcome the natural resistance to change. Major decisions and potential delays were flagged and explained in advance.
- Training. ADI adopted a hierarchical training model. Once core team members had been trained by SAP and the consultants, they trained local trainers who, in turn, trained their own groups. This ensured local SAP expertise where and when it was needed. A help desk, just a telephone call away, provided backup support, and the installation of a CD-ROM drive on the network allowed online searches of all documentation.
On the other hand, there certainly are places where, knowing what managers do now about SAP implementations, ADI could improve. Specifically, analysts today question ADI's extensive BPR prior to the implementation:
"Companies now are more likely to do the business reengineering as part of the implementation," notes AMRs Shepherd. The combined BPR-implementation approach helps to collapse the time.
In general, "companies do less business process reengineering now," adds Greenbaum of Enterprise Applications Consulting.
Then and now
ADI considers its SAP implementation a great success and, at the time, it surely was. But the ERP market and SAP have evolved considerably since then. If ADI were to undertake the project today, it would progress differently.
For example, the instability of ADI's initial version 1.x release of SAP delayed the project. Today, SAP releases are rock solid, notes Greenbaum. SAP's entire approach to releases is different, ensuring that organizations get a production-ready release out of the box.
Similarly, the implementation methodologies ADI employed were rudimentary. Today, "methodologies are much better," says AMRs Shepherd. And the whole issue of expertise is different. ADI had few choices when it came to finding knowledgeable consultants. "There is enormous choice now," Shepherd continues, although this expertise still is not cheap.
Back in 1993, ADI's adoption of SAP included the widespread conversion to PCs and client/server computing, which further burdened the projects financials and training efforts. Organizations implementing SAP today already have converted to PCs, LANs, and client/server computing. If anything, they are ready to move on to thin-client, Web-based computing.
Companies still shortchange training, though. When SAP implementations fail to deliver the expected benefits, notes Greenbaum, it usually involves user rejection, user frustration, and angry users undermining the system. Early training and education can mitigate such reactions. ADI could have done more with training, but the company got away with its minimally sufficient and relatively inexpensive approach.
But one thing remains the same. ADI's 156% ROI, determined by Scalea analysts, looks good anytime, and that is what really counts. With the advances SAP has made in stability and faster deployment, along with the widespread availability of SAP expertise, organizations should be able to achieve the kind of gains ADI experienced faster and cheaper.
Alan Radding, based in Newton, Mass., is a freelance writer specializing in business and technology. He can be reached at firstname.lastname@example.org.