Ertl's Decision Support Journey
by D. J. Power and R. M. Roth, University of Northern Iowa
Ertl Company, Inc. was a medium-sized toy manufacturing company located in Dyersville, Iowa in the Midwestern United States. In Fall 1991, the company embarked on a data warehousing project to improve executive information access and to eliminate manual preparation of reports. This case describes the decision support projects at Ertl from 1992 through 1998. The case illustrates the challenges of implementing decision support technologies in a traditional legacy computing system environment. Also, it showcases the decision support journey that occurs in many companies and the role of Information Systems department leadership in successfully completing and maintaining decision support projects. The chronicle of activities highlights the implementation issues associated with data warehousing and decision support.
During the Fall of 1991, top managers at the Ertl Company were pushing for better management information. Also, cost pressures were impacting all aspects of the company, and a centralized data warehouse and client/server computing system were viewed as one way of meeting these conflicting pressures. Under these circumstances, the Ertl Company, Inc. of Dyersville, Iowa, USA, began a data warehousing project. Ertl manufactured replica toys. One of its best-known products, replica John Deere tractors, is a classic favorite of Midwestern USA children and their parents.
In the mid-90s, the successes of Ertl's data warehousing project were widely publicized. The Data Warehousing Institute Web site had a testimonial, a case example was published by the primary software vendor, Software AG, and an article about the project appeared in the popular business computing magazine, PC Week. Also, Software AG's Open Data Warehouse Program won DB/EXPO's RealWare Award for the Ertl Company data warehouse project.
This case is a retrospective account and analysis of the various decision support projects at Ertl from 1991-1998. The case draws on published materials, email correspondence and a field interview. The evidence suggests Ertl's IS/IT staff faced many challenges including dealing with the publicity on the projects and sustaining momentum in the use of advanced information technologies for decision support. The ongoing decision support "journey" was especially difficult at Ertl because of changing data warehousing technologies, constrained financial and IS/IT personnel resources, senior management expectations, and changes in IS/IT leadership.
Brief Company History and Background
At the end of the summer of 1945, Fred Ertl Sr.'s union struck the Dubuque, Iowa foundry where he worked as a journeyman molder. So, Fred started a business in his basement. Fred melted down war-surplus aluminum and poured the molten metal into sand molds. His sons helped bolt the miniature parts into scale models.
In 1959, Ertl built a manufacturing plant just west of Dubuque in Dyersville, Iowa. Dyersville became "The Farm Toy Capital of the World" and a "Farm Toy Hall of Fame" was established. In 1967, the Ertl family sold the firm because of problems with their estate planning (see Jeffrey, 1997).
The Ertl family sold the company to Victor Comptometer. In 1977, Victor Comptometer sold the business to Kidde & Co. In 1987, Hanson PLC, a leveraged buyout firm managed by Lord James Hanson and Lord James White purchased Kidde. Hanson PLC incurred a significant corporate debt. So, in 1995, it ‘demerged’ its debt-laden U.S. holdings into a separate corporation called U.S. Industries.
Hanson PLC had opened a toy factory in Tijuana, Mexico in 1983 that was used only half the year. In the fall of 1995, with the recently passed NAFTA trade agreement guaranteeing they could bring products into the U.S. tariff-free, U.S. Industries announced it was moving 300 jobs, more than a quarter of its Ertl payroll, from Dyersville, IA to Tijuana, Mexico. So, 1995 was a difficult year for Ertl and the Dyersville community: the company lost $6 million and the community lost 300 jobs. In 1996 however, Ertl contributed the biggest improvement to earnings at U.S. Industries, generating a $3 million profit.
In April of 1997, the Ertl Company strengthened its presence in the European market by acquiring Petite, Ltd. and its associated brands. This 150-year-old British company, based in Nottingham, England, was a leading manufacturer of military soldier collectibles, plastic and die cast metal farm toys, figures, animals, buildings and plastic preschool toys.
In February 1998, Ertl brought online its Web site, www.Ertltoys.com. Also, George B. Volanakis resigned as CEO of Ertl. On April 14, 1999, Racing Champions Corp. acquired Ertl for $105 million in cash. The Racing Champions acquisition marks the end of the Ertl decision support journey and the beginning of a new RC Ertl (rcertl.com) decision support journey.
Beginning the Journey
In 1987, Hanson PLC had purchased Ertl's parent company. This ownership change marked the beginning of a period of turmoil for the Ertl Information Systems organization. Consultants convinced senior management that an IBM AS400 should be acquired to replace the existing mainframe system. One year and 10 months was spent trying to convert the legacy software and move data from the existing mainframe to the AS400. At that point, the Ertl IS staff demonstrated that the AS400 was inadequate for Ertl's needs, and the conversion effort was abandoned. Ertl eventually acquired a more powerful IBM 4381 mainframe computer.
There was considerable instability in IS leadership during this time. The company went through three IS directors in two years. The hiring of Bill Bruttaniti as Director of Information Systems in October 1991 provided much needed stability for the IS organization. His leadership over the five-year period ending in July 1996 brought considerable growth and improvement in both technology and applications. For example, only a few PCs could be found at the time that Bruttaniti joined Ertl. By the time he left, there were over 650 PCs throughout the company. In 1997, the standard desktop PC was a Compaq Deskpro 4000 with a 200 MHz Pentium processor, 32 MB of RAM, a 2.1 GB hard disk, and the Windows 3.1 operating system.
Staci Theisen was promoted internally to succeed Bruttaniti in October 1996. At that time, the Ertl IS organization consisted of 23 employees, including six full time developers, three contract programmers, three managers of systems development, one database administrator, one business analyst, one manager of end user computing, two help desk specialists, and one network specialist. At the peak of data warehouse development efforts 21 contract programmers had been employed. During 1998, the firm was actively trying to recruit additional IS staff, but salary constraints were making it dificult to attract candidates.
What was the IS/IT problem? and what was the goal?
Bill Bruttaniti explained the implementation of Ertl's data warehousing project in a 1994 Data Warehousing Institute interview. "Our overall goal" was "to develop a global network and a corporate application suite that runs on any hardware platform," explained Bruttaniti. "With such a network," he continued, "all departments can share information so that we can deliver the quality products our customers expect. At the same time, we can select the most cost effective hardware while minimizing application maintenance overhead".
Ertl began updating its information systems by building a data warehouse. According to Bruttaniti, "The data warehouse was our first step in replacing legacy systems with newer applications. It has enhanced both data accessibility, as well as data availability." A data warehouse is often on a dedicated server, but Ertl kept both its transaction processing systems and data warehouse on the same mainframe platform.
Bruttaniti noted that "Ertl was a typical mainframe legacy site, suffering from performance problems and backlogs on a machine that couldn't cope with increasing information demands. The only way we saw out of this situation was a radical overhaul of our information strategy. But we had a big investment in mainframe-based technology and legacy information. We wanted a way to progress without writing off that investment."
Despite the promises from some software vendors, reinforced by articles in trade publications, major shifts to open systems were generally difficult in the early 1990s. Only a few organizations were able to eliminate all of their legacy applications and move completely to client/server systems. Other firms, like Ertl, chose a coexistence strategy that allowed legacy systems to continue to operate while new applications were developed. At Ertl, the data warehouse project provided the means to implement that strategy.
Ertl's data warehouse was a step along the migration path to an overall goal of improved decision support. So in choosing tools for data warehousing, Ertl IS managers identified conversion as a critical task. Application conversion cost was used as an important criterion in selecting the database and tools that would be used for the data warehouse.
Justifying the Data Warehouse Project
According to the Data Warehousing Institute case study (1994), Ertl's justification for data warehousing was based on direct and measurable costs. Users needed information and were already incurring large costs to get it. Ertl's IS planners identified all of the places where extensive manual efforts were being used to generate decision support information. Identifying the current tasks gave the planners two key benefits. Bruttaniti noted first, it gave them a dollar value for their project – "18,000 hours per year translated into hundreds of thousands of dollars per year in savings. Second, it showed them targets of opportunity for the data warehouse - subjects and formats for information that were clearly needed by the organization."
Once the inventory of current reporting efforts was complete, Ertl re-engineered information access. "We divided the company into three core competencies, and then we mapped out the functional workflow across all the processes and identified areas where we could improve productivity by adding new applications that would act as enablers," related Bruttaniti. One of the easiest places to see the value of this cross-process information modeling effort is in the ability of Ertl's users to track information from forecasts to orders to shipments to production status. Bruttaniti claimed "The information is available easily and instantly for a range of decision support applications."
The Data Warehousing DSS Project
In January 1992, the Ertl IS staff began the analysis of various software tools to use in constructing its data warehouse. The analysis included products from Oracle, FOCUS, Informix, SYBASE, and Software AG. The following timeline of the data warehousing project is based on conversation and email with Jim Singsank, manager of systems development, and published comments by Bill Bruttaniti.
Both Bruttaniti and Singsank noted you can’t just buy your data warehouse. To build its warehouse and provide tools for future application development, Ertl decided on a set of criteria and then tested tools to see how well they met those criteria (cf., Bruttaniti, 1994). The most important criteria were:
Ertl's tests showed that most software vendors could offer two or three of these benefits, but according to Bruttaniti, "virtually all products required a major conversion effort, with the notable exception of those from Software AG. Its VSAM and DL/1 integration was a key differentiator because it enabled us to complete the entire data warehousing transition without impacting end users." VSAM is an acronym for ‘virtual sequential access method,’ and DL/1 stands for Data Language 1, an early database from IBM for its DOS/VSE mainframes (TechEncyclopedia, 1998).
To gain Ertl's business, Software AG had "to prove that its ADABAS data management software could outperform other database management systems and offer subsecond response time. Software AG also had to prove that the data could be easily accessed from all the popular query and reporting tools, including its own ESPERANT. Finally, Software AG had to prove that its NATURAL fourth-generation language could improve application development productivity four or five-to-one over COBOL, and that both ADABAS and NATURAL ran well on the mainframe and could run on all other popular operating system platforms."
By September 1992, Ertl had made the decision to partner with Software AG. According to Bruttaniti, the lower migration costs, the mainframe support along with real portability, and the performance all combined to make the product selection decision easy. Software AG was the 10th largest independent software firm in the world, with dual headquarters in Darmstadt, Germany and Reston, VA. Software AG reported revenues of $533 million (US) in 1994.
In October 1992, Ertl staff developed the Request for Funds from the Corporate office to purchase new application software and to install a new PC network with standardized desktop software. Ertl standardized on Compaq Deskpro PC’s with Windows 3.1, Lotus SmartSuite, Amipro, and CC:Mail as standard desktop applications. Replacement of old CRTs with the new networked standard hardware and software began in May 1993.
Installation of the Software AG products occurred in September 1993. Training for the systems development staff on these products began in October 1993. By January 1994, the systems development staff began the baseline Data Warehouse development in NATURAL and CONSTRUCT, using extracted data from the old CICS/DL/1/COBOL systems.
User training and testing of Data Warehouse applications began in June 1994. These applications utilized sales and order information from past and present transactions with nightly updates of daily activity. "The [query] backlog for accessing data from the [4381 IBM] mainframe was tremendous," said Doug Dalrymple, WAN/LAN Network Administrator in 1995. "The company needs all kinds of information at [its] fingertips." Late in 1994, Ertl opened its mainframe to PC users by installing EDA/SQL middleware from Information Builders and Software AG's ESQ for Open Database Connectivity (DiCarlo, 1995).
In 1994, the success of Ertl's data warehouse project allowed the IT department to begin the enhancements necessary to implement Software AG's financial and manufacturing software which were also purchased in 1993. This is an integrated suite of tools that allows all departments to share information. To serve the needs of the executive staff, in December 1994, a decision support application called SOP (Sales and Operating Plan) was developed in NATURAL.
Figure 1. Ertl Network Diagram
In March 1995, the company linked its remote sites to the mainframe via an AT&T InterSpan frame-relay WAN. Ertl's operations consisted of two buildings in Dyersville; two buildings in Tijuana, Mexico; one in Toronto; one in Exeter, England; one in Nottingham, England; and an office to coordinate contract manufacturing in Hong Kong (see Figure 1 for network diagram). Ertl’s world wide network was constructed using a combination of Frame Relay, fractional and full T1 lines, and Virtual Private Networks (VPNs) over the Internet. Frame Relay is a high-speed packet switching protocol used in wide area networks (WANs). It has become popular for LAN to LAN connections across remote distances, and frame relay services are provided by all the major carriers. T1 lines are point-to-point dedicated lines provided by telephone companies that carry data at 1.544 Mbps.
In June 1995, a Visual Basic 4 interface to the Sales and Operating Plan (SOP) was developed accessing data directly from ADABAS on the mainframe for drilldowns from the Company level to the Sales Item level, allowing selection of data to compare and presenting it using applicable graphs and charts. "We developed the [decision-support system] for our executives so they can get quick graphical pictures of the state of the business. They can also run through 'what-if' scenarios for the planning process," said Dalrymple (in DiCarlo).
We "continued developing data warehouse queries through 1996, rolling them out to the users as testing and user acceptance was completed. During this same time period we installed networks in Tijuana, Mexico, Toronto, Canada and Exeter, England. We also had a staff reviewing current business processes looking for ways to improve efficiency and effectiveness, developing and enhancing the application software, training staff at all locations and helping people to accept CHANGE! We also experimented with an EIS in FOCUS from Information Builders and EIS from SAS. We determined there was too much overhead required at the desktop and on the mainframe so we will develop our own," according to Jim Singsank.
In 1995, the Ertl IS staff investigated Lotus Development Corporation's Notes as a document repository and discussion database for marketing and traveling sales people. That investigation determined that it would be too costly to deploy and was still undergoing too many changes.
In June 1996, Bill Bruttaniti left Ertl. In an InformationWeek article that same month, Emily Kay reported "100 Ertl workers use Microsoft's Access and Software AG's Esperante front-end query tools on Windows 3.1 PCs to query ... Software AG's Adabas D relational database management systems. Ertl's three-tiered implementation has reduced the time needed for query responses to one minute. That's plenty fast for both senior executives analyzing trend reports and midlevel product and purchasing managers and customer service representatives running more detailed scheduling and forecasting comparisons."
Staci Theisen was promoted to Ertl IS Director in October 1996. In January 1997, Ertl implemented new application software including additional Data Warehouse functionality to reflect Shop Floor Metrics for measuring effectiveness in the factory.
Jim Singsank explained, "Ertl developed a scaled down or baseline Executive Information System with Visual Basic 4 to access information stored on the host". Development of a more elaborate EIS system began in October 1997, utilizing Acucobol. This system showed daily, monthly, and year-to-date activity with drill down capabilities from the Company level through the customer level to the Sales Item level.
"Throughout our development efforts on the PC side we have utilized Microsoft Access as well as direct calls to ADABAS from Visual Basic and Acucobol. We have performed testing with Visual Basic and Acucobol and have determined that Acucobol can produce the same results in 75% less time," explained Singsank. Ertl is constrained to development in a Windows 3.1 environment. In 1995, Ertl was experiencing financial problems and the decision was made not to upgrade to Windows 95.
In November 1997, Jim Singsank noted, "We will be adding Europe to the mainframe software applications during the next 16 months .... We have definitely proved that there is a need for mainframes in the correct situations. Not every business can run on a PC network or on an AS/400 or a mainframe. One complements the other in certain cases."
Results of Decision Support Projects
In 1994, Bill Bruttaniti, Director of Information Systems at Ertl said, "We used to spend more than 18,000 hours every year creating reports. That task has simply been eliminated, and it's a repetitive saving. I expect these gains to accelerate as we grow in expertise, with as much as 50% improvement in the productivity of our information users." Both Bruttaniti and Singsank asserted that a new data warehouse and new applications that enhanced information accessibility, allowed Ertl to reduce annual overhead costs by $750,000 and increase information access productivity 15%. The managers reported achieving these savings in just eighteen months, almost a year and a half ahead of schedule.
In November 1997, Jim Singsank at Ertl concluded, "The biggest hurdle we have had over the past five years is getting people to accept change and get them to think outside of the box. We have accomplished our goal that was set in the beginning: ‘Get the data at the users fingertips’ (but they always want and need more). Our work will never be finished, due to changing business requirements, new technologies, and new ideas. The IS world still needs better query tools that operate more efficiently."
In February 1998, Ertl IS managers were reassessing their information systems and information needs. The use of Windows 3.1 as the desktop operating system continued to constrain development choices, as evidenced by the decision to discontinue Visual Basic development because of Visual Basic 4's slow performance. Ertl's shortage of staff was perhaps exacerbated by their location in a small Iowa town. Ertl was also experiencing some network problems, and was plagued by slow access to mainframe data. A site visit by the authors in February 1998 was cordial, but the Information Systems staff members seemed stressed and disillusioned. At the time, U.S. Industries was trying to spin off Ertl to another owner and there was a pervasive atmosphere of uncertainty.
The highly publicized data warehousing project at Ertl achieved a number of successes. The system was able to substantially reduce clerical costs. Ertl did successfully move from solely a mainframe/dumb terminal environment to a PC/Novell/Mainframe LAN environment. A data warehouse was created that significantly improved access to managerial information. Ertl was also able to separate operational from decision support data. The total project cost exceeded $5 million.
However, despite the achievements of Ertl's IS transformation project, the timeline above illustrates that Ertl was slow in its move to a client/server environment. Ertl's mainframe environment was extended through this project, but the mainframe may still be approaching obsolescence. The future role of mainframes remains a debatable question. By 1998, new decision support technologies were emerging (OLAP and web-based DSS) and Ertl staff were stretched so thin they needed to outsource deevelopment of a company website, and IS staff made the decision to abandon Lotus Notes. The partnership with key vendor, Software AG, deteriorated as demand for its products declined. The vendor was slow to enhance products and some products were not being maintained. The Executive Information System on the Windows 3.1 platform built using Visual Basic 4.0 was not successful. Keeping the momentum for innovative decision support going in the face of budget difficulties is always difficult and decision support innovation at Ertl seemed to become a low priority in 1997 and 1998.
About Racing Champions Ertl
Racing Champions Ertl (www.rcErtl.com) is a leading producer and marketer of innovative collectibles and toys. The Company's diverse product offerings include scaled die-cast replicas of John Deere agricultural equipment and NASCAR stock cars, other licensed vehicle replicas, pre-teen toys, sports trading cards, racing apparel and souvenirs, and collectible figures. These products are sold under the Company's market-focused brand names, including Racing Champions®, Ertl®, Ertl Collectibles®, American Muscle™, AMT®, W. Britain®, Press Pass®, and JoyRide Studios™. The Company reinforces its brands and enhances the authenticity of its products by linking them with highly recognized licensed properties including NASCAR, NHRA, John Deere, Harley-Davidson, Case LLC, Polaris, Honda, Caterpillar, Ford, GM, Chrysler, Texaco, Universal Studios, Warner Brothers, Gullane Entertainment, Nintendo, Sega, and Electronic Arts. The Company's products are marketed through multiple channels of distribution, including mass retailers and specialty and hobby wholesalers and retailers, OEM dealers, and to corporate accounts for promotional purposes. The Company sells through more than 20,000 retail outlets located in North America, Europe, and Asia Pacific.
Email Comment and Update from Neil McCarty:Subject: Ertl's Decision Support Journey
Date: Fri, 31 Jan 2003 11:46:22 -0600
From: "McCarty, Neil"
I read your "Ertl's Decision Support Journey" article today. I think your article is good and accurate. I was the DBA at Ertl from January 1994 to June 1997. You mention me by title in your article.
I left Ertl to join Gateway in May of 1997. At Gateway I have helped develop and deploy a couple of data warehouses including a CRM. I am now a Manager in IT and frequently do project manager work.
I don't think I could have written this article from a technical perspective as you did. As I remember - there was more work and time spent with personnel conflicts than there was on engineering. You see, Bill Bruttaniti was from Boston. He was a very smart man with a desire to make a difference. He had a great idea with the data warehouse, but he came to Iowa and with his thick New England accent told the people in Dyersville IA they were going to change and they were going to change in a hurry. "Change in a hurry" almost never happens.
The chemistry in the shop when I was there didn't work well at all. Bill was great at keeping the conflicts behind closed doors. We were all co-located but there were too many outsiders. I was from Minneapolis, a few were from Canada and a few from South Africa. Maybe if we weren't co-located things might have worked better.
It took a little while to get it straightened out. I knew by 1995 that Bill had to leave and Jim had to be put in charge. I could have worked for Jim and probably liked it but when Chet replaced Bill with Stacy I had to leave. Stacy was a professional trainer and an administrative assistant promoted to CIO.
I have taken up a lot of your time. I just wanted to add my little bit to a project I worked very hard at. If you have any questions please feel free to contact me.
Please Note: Neil McCarty gave permission to post the above comments with this case on Friday, February 28, 2003.
Jim Singsank is currently the Senior Director of IS at Racing Champions Ertl. According to a customer testimonial at the CrossAccess website (http://www.crossaccess.com), the CrossAccess eXadas Data Integrator product was used to access mainframe data for the Racing Champions Ertl B2B Web site. Singsank is quoted as saying "We saved 30% compared to the cost of moving to a client/server model."
A Racing Champions Ertl press release dated November 4, 2002 described the success of RCE On-Line Advantage, the company's B2B Web site (www.rceb2b.com). According to the release, "This site serves small to medium-sized hobby and specialty gift shops as their primary ordering and information tool for Racing Champions Ertl products. User response to the Web site has been extremely enthusiastic, with 82.7% ranking their overall experience as "very good-excellent" in a recent survey of active buyers. Customers ranked the top advantages of the Web site as:
The ability to conduct business at their convenience: 55% of site transactions occur during non-traditional business hours and 24/7 accessibility was ranked as the #2 reason to use the site, just after ordering terms which ranked #1
The amount of detailed information on-line: 86.2% placed high value on the digital photography and another 70.7% on the detailed product descriptions
Questions for Analysis and Discussion
Bruttaniti, B., "Maximizing Profitability, Gaining Share, and Minimizing Overhead: The Ertl Story", Commentary of the Director of Information Technology at the Ertl Company, originally from URL http://www.dw-institute.com/cases/ertl.htm, 1994, see local copy.
"Data Warehousing Helps Ertl Slash $750,000", from URL http://www.sagamericas.ca/customers/Ertl.html
DiCarlo, Lisa, "Toy supplier plays to win at client server game (Ertl Toys Inc builds a network)", PC Week, Feb 6, 1995, v12, n5, p1(2), see local copy.
Jeffrey, T. P., "A Phony Class War: Ertl Toy Story", Human Events, January 31, 1997.
Kay, Emily, "Decisions", InformationWeek, June 17, 1996, http://www.informationweek.com/584/84mtdec.htm .
Singsank, Jim, a Manager of Systems Development at Ertl, Email correspondence, November 18, 1997.
Software AG Press release, Reston, Va., November 27, 1995.
About the Authors
Daniel J. Power is a Professor of Information Systems at the University of Northern Iowa, Cedar Falls, IA and Editor of DSSResources.COM. E-mail: Daniel.Power@UNI.edu.
Roberta M. Roth is an Associate Professor of Information Systems at the University of Northern Iowa, Cedar Falls, IA. E-mail: Roberta.Roth@UNI.edu.
Please cite as:
Power, D. J. and R. M. Roth, "Ertl's Decision Support Journey", 2003, at URL DSSResources.COM.
Please note: Initial contact on this case was made in Fall 1997. Both Jim Singsank and Staci Theisen reviewed various early versions of this case in late Spring 1998. Jim Singsank reviewed the most recent version before it was posted at DSSResources.COM. We especially thank Jim for his help and comments, but the responsibility for any errors or misinterpretations is solely the authors.
This case study was posted at DSSResources.COM on January 31, 2003.
"Ertl's Decision Support Journey" was written to stimulate discussion and for informational purposes only. This case study is not intended to demonstrate "best" or "worst" practice in developing decision support systems. DSSResources.COM makes no warranties, express or implied, about the practices in this case study. All trademarks are the respective property of their owners.