Pareto Chart Analysis

 

Chapter 2 - Optimal Purchase Decision Criteria for Information Technology
Managing IT in Government, Business & Communities
by Gerry Gingrich (ed)
Idea Group Publishing 2003

	Brought to you by Team-Fly

A Pareto diagram is a special type of vertical bar chart in which categorized responses are plotted in descending rank order of their frequencies and combined with a cumulative polygon on the same scale. The main principle supporting this graphical device is the ability to separate the "vital few" from the "trivial many," enabling us to focus on the most important responses (Chen, Chong, and Tong, 1994). The Pareto diagram is widely used in the statistical control of process and product quality (Berenson, 1999). Studies on the Pareto Principle (80% of wealth is owned by 20% of people) use a modified Pareto Curve that aggregates items of the same number of usage into subgroups and then shows only the subtotal of that group (Chen, Chong, and Tong, 1994). For example, using the data from http://www.xtechnology.com/, it shows the pricing of AMD's Athlon CPU chips, ranging from 800 MHz to 1.1 GHz. Using Excel to generate an x-y-chart based on these data and we can visualize the Pareto curve in Figure 1. It is easy to see that the beginning section of the curve indicates that the benefit rises faster than the cost, supporting the notion that 80% of benefit comes from the first 20% of expenditures. A similar chart created about two months earlier showed a much shorter linear section, implying that chips with higher speed demanded a premium back then. However, this elongated linear section implies that perhaps due to intensive competition, at present time consumers may choose a faster system without having to pay undue higher prices for performance they may not need.

Figure 1: Pareto Curve for Athlon Chip Pricing

Cost-benefit analysis includes a systematic categorization of impacts as benefits and costs, valuing in dollars for assigning weights. Of course, the valuation of cost and benefit is only an operational definition in our assessment. Based on the result, we may make a final decision based on the net benefits of each alternative (Boardman, Greenberg, Vining, and Weimer, 1996). We may apply the Pareto Principle to the cost-benefit analysis and state that if the full price reflects the full benefit, then there is a small fraction of expenditure that represents the major portion of benefit. In other words, 80% of the total benefit is realized in the first 20% of cost. The mission then, is to identify this optimal point. Recall that in microeconomics, to maximize total profit, the optimal production level is where marginal revenue equals marginal cost (MR=MC) (Samuelson, 1976). Thus, we should choose the alternative that is at the point where marginal benefit equals marginal cost. From our earlier discussion, if we can construct a cost-benefit curve, that should be the point where slope equals 1.

Unfortunately, with our experience of using this system, the x-y chart never yielded a perfect Pareto diagram, partly because a vendor's pricing strategy does not perfectly match the consumer's perception of benefit. However, invariably some systems would bunch around the general area where the slope = 1. There is also an interesting observation that may have a deep implication. If we construct the x-y chart using (0,0) as the origin, it may surprise us that often the new diagram does show a pattern similar to a Pareto chart and indicates that the cheapest computer is the optimal choice today! Intuition tells us that indeed for most consumers today, even the cheapest computer has more capability than s/he needs in surfing the Internet and word processing. Given the low cost and the advancement of technology, it may be a good strategy to purchase a cheap computer with the intention of replacing it in two to three years. In other words, treat computer purchasing as an "expense" rather than a "capital investment."

Some may view this conclusion to be a "dangerous" one that may make consumers overlook new features that may improve their productivity. However, this possible conclusion also suggests that in order to fund the rapid IT manufacturers and providers, consumers may have paid (and continue to pay) an exorbitant amount for the excess capacity. In either case, for a variety of reasons, decision makers may feel uncomfortable choosing the "cheapest" computer on the market anyway. This is may be one of the overriding factors we mentioned earlier.

Instead of using the result as a method for finding the "best" or "optimal" product right off the bat, this x-y chart is a very effective tool for eliminating "worst" choices. In other words, the graphical analysis can be effectively used to provide a short list for decision makers. From this short list a purchasing then can apply other methods to come to a decision. However, an enormous amount of time can be saved for the decision maker, since s/he will not have to wade through a myriad of computer choices.

	Brought to you by Team-Fly