Sensitivity Analysis Of A Proposed Capital Investment

The Proposed Investment in a New Data Center

Judi Ritts, senior project leader in the Corporate Information Services Group at Central Bank, sat at her desk with her head in her hands. The analysis she had just completed of Central Bank’s proposed investment in a new data center suggested that the project’s net present value (NPV) was negative—not by much, but by enough to be of concern. Scheduled to meet with her boss at 3:00 p.m. to review the results of her analysis, she contemplated what to do next. When she and her boss had discussed the economics of green practices in the context of the bank’s next data center, she had been sure she could justify the cost of the investment in green technologies; the cost of power and cooling infrastructure had become the primary cost drivers in a data center, and new technologies were promising substantial reductions in power and cooling costs.

Central Bank was considering a 60,000-square-foot data center, which would replace its site in Charlotte, North Carolina. Initially, it would contain 700 “racks,” each of which held 40 servers and their associated peripherals, memory, and disks. The cost of power for each rack averaged $45,600 per year (20 kilowatts × 24 hours × 365 days × 2.6 efficiency × $0.10 per kilowatt hour). The energy efficiency of the new data center was expected to generate substantial savings; Ritts thought Central Bank might be able to reduce its power usage effectiveness (PUE)— the total power usage of the data center divided by the power usage of the servers—from 2.60 to as low as 1.30 if everything went as planned.

Ritts’s boss had asked for a quick analysis based on some rough assumptions (see Exhibits 1 and 2). In addition, her boss had suggested that she use annual cash flows; assume that all investment in the new data center (as well as the sale of the old structure) occurred at year zero; and assume a simple tax treatment, PUE no lower than 1.35, and a 10-year project life with a residual value equal to the tax shield on the net book value at that time (see Exhibit 2).

The technology inherent in this proposed new data center was the way of the future. Central Bank needed to move in that direction, but if it was going to do so, it needed to be beneficial financially. Reflecting on all the assumptions she had to make in the analysis, Ritts wasn’t sure how reliable the projected NPV was. No doubt that projected NPV was sensitive to her choices about those assumptions. Not only that, she was certain her boss would expect her to make recommendations about what actions the bank could take to improve the financial attractiveness of the investment. Ritts realized that she still had a lot of work to do.