Modeling the Optimal Life Cycle of Faculty Personal Computers at UNC
Student Consultant: Christopher M. Hawke
Date: April 1996
The Office of Information Technology (OIT) at the University of North Carolina at Chapel Hill oversees computer services for the entire university community. The client, OIT’s finance and planning director, is charged with promoting the use of information technology among faculty and other university employees in a cost-effective manner. In one potential future scenario, OIT would develop a plan for the University to supply personal computers (PCs) to large segments of the university population, such as faculty or support staff. To minimize the University’s costs of providing these PCs, OIT needs to determine:
- The replacement interval for a PC, determined by its economic life cycle.
- The factors that most influence the economic life cycle.
- The factors that have the greatest effect on the cost of providing a PC to a faculty member.
- The factors that are most significant in determining the benefit the faculty member receives from using the PC.
- The costs of three alternative university PC acquisition and management strategies:
- purchase all PCs and provide all attendant services, including maintenance, training, and help desk support.
- lease all PCs but provide maintenance, training and help desk support.
- lease all PCs and secure maintenance, training, and help desk support services from outside vendors.
- The factors that most influence the costs of these strategies.
To provide a quantitative basis for answering these questions, this project creates two models. The first determines the life cycle of a faculty PC that maximizes the discounted net return. The discounted net return is the total benefit of using a PC minus the costs of operating the computer, expressed in terms of today’s dollars. The model uses dynamic programming, an operations research optimization technique, to develop a replacement plan that maximizes the discounted net return over a 15-year planning horizon. The second model uses the life cycle determined by the first model to evaluate the costs of implementing the purchasing, leasing, and outsourcing strategies over a 10-year planning horizon. Both models are formulated as Microsoft Excel spreadsheets. The use of Excel allows OIT to update data estimates and parameter values, and to evaluate different scenarios as the computing environment changes. In addition to calculating life cycles and costs, these models also enhance one’s understanding of which factors affect the economic life cycle of a faculty PC and the costs of providing a large number of users with PCs.
Data were collected to estimate the values of parameters in the models. Based on these estimates, the PC life cycle model determines that a $2500 PC has a three-year optimal life cycle. This computer has a 90 megahertz Pentium processor, 8 megabytes of random access memory, 540 megabytes of hard drive storage, a 14.4 bits per second internal modem, and several software applications, including Microsoft Office. The second model determines that leasing is the least expensive of the three acquisition strategies. The net present value of a 10-year program that provides PCs to 2000 faculty is $18.7 million for the leasing plan and $21.2 million for purchasing. The leasing strategy has an average annual per user cost of $1230, compared to $1370 for purchasing.
Since these data estimates are subject to error, a sensitivity analysis was performed that allows for errors of + 20% of each estimated parameter value. Results were generated for 900 possible sets of parameter values. If the models tend to make the same choices, even while the parameter values vary, we have confidence that errors in estimates are not a serious concern in evaluating results. This was indeed true, as the models produced the same policies for a variety of parameter values. Sensitivity analysis revealed that:
- In 75% of the trials, replacing faculty PCs every three years maximized the discounted net return.
- PC acquisition price, PC market price, and the amount of training provided to the user were the most significant factors in determining the optimal life cycle of a PC.
- In 85% of the trials, leasing was the least expensive of the three acquisition strategies. In 45% of the trials, leasing induced an NPV that was at least $4.3 million less than the NPV for purchasing. In the 15% of trials where leasing was more expensive, the net loss never exceeded $430,000.
- Lease rates and salvage values were the most important factors in determining the optimal acquisition strategy. In all 122 trials where leasing was more expensive than purchasing, these two parameters were set significantly above their estimated values.
- For given acquisition price and replacement interval, hardware maintenance costs and support staff salaries were the factors that were most significant in determining the cost of the purchasing strategy. The lease and outsource rates were the factors with the greatest effect on the costs of the leasing and outsourcing strategies.
Based on these results of this study, we recommend that the University use a three-year replacement cycle for any large PC acquisition program. Before selecting an acquisition strategy, the University should determine if both lease rates and salvage values are significantly above the estimates used in this study. If they are, further analysis is required to determine if leasing is cheaper than purchasing; if they are not, the University should proceed with leasing.
For the impact of this study, see the article Some Colleges Find It More Practical to Lease Computers Than to Purchase Them in the November 8, 1996 issue of The Chronicle of Higher Education.