Computer models allow managers to analyze multiple complex business process scenariosBY ERIK SHERMAN
Spreadsheets may be a marvelous time-saver to most, but there are times when a thorough business analysis simply cannot rely on them solely , especially when the business is complex.
Fletcher Challenge Canada Ltd. in Vancouver is a pulp and paper manufacturer focusing on the international communications paper market, including newsprint and newspaper inserts. With 1995 revenue of C$2.15 billion and a variety of shipping methods, Fletcher Challenge has quite a complex supply chain extending worldwide.
In an attempt to increase its competitiveness, the company embarked on an analysis of its entire infrastructure, focusing on three particular factors: lead times, inventories and costs. While the company felt that the analyses were sound, there was no practical way to examine the complex interactions of these factors.
"We had done some basic analysis on our infrastructure. And the analysis we had done on the most part was good analysis, but in the format of (Microsoft) Excel spreadsheets. We really couldn't look at the interdependency of those factors," said Gino Crisanti, director of distribution at Fletcher. "We pretty much had a feeling for where we wanted to go. When we did the analysis in isolation, it said, 'Go ahead, go forward with this.'"
However, when decisions are "fairly significant and going against the grain of long-standing tradition," as Crisanti put it, those responsible for proposing solutions want to be sure of their recommendations. To compound this, not only were the suggested changes significant, but so were the amounts of money involved.
Almost serendipitously, Pramod Jain, president of Chicago logistics consulting firm Indent Inc., was working with Fletcher Challenge on another project. He happened to hear one of Crisanti's presentations and, afterwards, suggested that Crisanti consider building a simulation of the system.
Simulations are mathematical approaches to modeling and analyzing the dynamic behavior of processes and systems, and not just historic, static behavior. The popular computer game SimCity is actually a simulation of urban operations.
Jain was well acquainted with simulation, having used it extensively in the past. Fletcher Challenge was in an ideal position to try the approach, because of both the previous analysis and the availability of extensive, unprocessed data.
"Most of the scenarios had been identified, i.e., 'these are the possible changes we would like to make.' What I proposed was to look at the problem from these three performance measures and evaluate these scenarios," examining the model behavior over a trial period of four years as the business changed, said Jain. "There was an extensive data-collection process. It was raw data, but clean."
For a simulation package, Jain chose Extend by Imagine That Inc. of San Jose, Calif. The product, available in both Windows and Macintosh versions, uses a graphical approach to modeling. Users choose pre-existing elements, or blocks, fill out dialogs to tailor block characteristics, and connect the blocks to create a simulation. In addition to the base set of blocks, users can order libraries for manufacturing or for business process reengineering. They can also develop their own blocks with a programming language similar to the C programming language widely used with Windows and other applications.
Jain said, "I've found [Extend] to be basically very good for what I want to do. The only downside, really, is an inability to do real animation, which some other products can do." Those other products, however, cost substantially more, according to Jain, and most of his clients decide to use Extend as a less expensive alternative.
Crisanti said he has been pleased with the use of simulation. "It was taking the analysis we had done and giving us another level of sophistication, and level of comfort, on what we had done," he said, "It verified our initial thinking, and that was extremely beneficial." Had the two analyses disagreed, Fletcher Challenge would have been forced into highly detailed reviews to learn why they were getting different answers.
This is not to say that all the results were as expected. As Jain performed the simulation, lead times rose by 2 percent to 5 percent as inventory decreased, which he found a bit surprising. "In some cases, an inventory drop means that your lead time will also drop," he explained, because the faster movement of goods can lower inventory levels.
"Obviously, we had to go back and rationalize [that result] and say why it was going on," said Crisanti. As it turned out, the lead time increases were due to some rerouting in some of the scenarios modeled. "At the end of the day, it netted out to be insignificant in the system," he said. "We got a lot more detail, and by detail I mean by region, by account, given these scenarios, what the cost was going to be, what the lead time was going to be," said Jain.
It was important to Fletcher Challenge to see how specific customers were affected. Extend's ability to perform discrete, as well as continuous, simulation was a key. The nature of discrete models allowed the company to track to a very detailed level, including individual shipments as an example.
Simulation seems to have found a home at Fletcher Challenge. "I think that there's power in the tool and its presentation. Never before in our organization have we been able to look at those factors on an interrelated basis," said Crisanti. "It brings power to the credibility story, and also demonstrates that you are trying to use leading technology to stay current in an increasingly competitive environment. It helps build confidence in the decision-making process."
But it is clear that simulation, while useful, is an additional tool, not a substitute for others. Not all decisions warrant the associated efforts and expenditures of this virtual business modeling. Fletcher Challenge is moving forward with the infrastructure recommendations, but still plans on a peer review. Crisanti felt that a "common-sense, business-intuitive point of view" was important to weigh the intangibles that cannot be easily measured.
-- Sherman is a freelance writer based in Marshfield, Mass. His e-mail address is 75201.1404@Compuserve.com
Appeared in Traffic World, September 2, 1996