Data Center Benchmarking Analysis

Project Background

The client was the senior data center management of a major insurance company with five distinct operating companies. The data center was part of the shared services organization supporting all five companies. The primary operating environment was MVS, but there was also a significant, declining VM environment. In addition to supporting purchased software for these environments, the data center also supported a large variety of “home grown” utilities. Over time, the operating business had changed substantially, acquiring a substantial amount of new (but pre-existing) business.

Problem Summary

This company traditionally benchmarked its data center performance annually, employing a TBI competitor. Results were used to assure management that shared data center service costs and productivity were within average industry ranges. In most cases, no follow-up analysis to the annual benchmarking study was conducted because results were generally favorable. For three years in a row, however, the benchmarking study found that both mainframe software costs and technical support staffing were significantly higher than those of comparable data centers. The benchmarking vendor was asked to explain the root cause of this finding, but was unable to do so, so TBI was asked for assistance in this area.

TBI’s Approach

TBI conducted a benchmarking analysis of technical services costs and staffing in the client’s data center. The earlier findings were validated – mainframe software costs and technical support staffing were seen to be much higher than average. TBI then conducted more in depth analysis of data center software and staff work activities.

Software products licensed by the data center were identified and analyzed for functionality and usage with an eye towards identifying opportunities to eliminate software costs where alternate products existed with similar functionality and to “push back” the charges for software that was not in general usage, but rather used by one business unit, to that business unit instead of including it in the data center overhead for all customers. To examine the staffing further, TBI constructed a task inventory which was then completed by all Technical Services employees and which yielded data on the specific activities in which these personnel spent their time.

Successful Business Solutions

Several sets of software products with duplicate functionality were identified for which TBI recommended product standardization across users which would allow the client to cut their current cost “overage” (compared to the benchmark average) by two thirds ($2 million). Efforts to reduce the number of “home grown” utilities supported were also recommended.

The proliferation of systems software in this shared services data center and the lack of effective configuration management processes (i.e., to control hardware and software inventory components, contracts and invoices) were found to contribute to the much higher than average staffing in Technical Services (twice the benchmark average). Since the environment was so complex, it needed more people to support it. However, a significant number of personnel were also found to have responsibilities which are not typical of most data centers. TBI found that several of them performed user (application programmer) education functions and several of them provided regular consulting services to application programming teams. Further, these functions were found to have high business value. TBI recommended establishment of a “pay for service” programmer education/consulting unit instead of carrying these costs in the shared services data center overhead.

The customer learned why data center software costs and technical services staffing were so high and how to reduce both, without sacrificing customer service.

Share Button