Editor’s note: This is the latest in a series of "Executive Insight" columns, a regular feature for Local Tech Wire as part of its partnership with the and . Michael A. Eason leads the Information Services organization at the American Board of Anesthesiology, the certifying body for anesthesiologists in the United States and Canada.

By Michael A. Eason, special to Local Tech Wire

RESEARCH TRIANGLE PARK, N.C. – Information technology is a strategic asset. Companies rely on IT for technology infrastructure and the enterprise applications to conduct their core business. We invest substantially in IT, making it important to understand what benefits are received in return for this investment.

ROI is comprised of investment and return. A previous article addressed the question of investment (). The remaining question involves the return on that investment. Demonstrable performance should be commensurate with IT investment. The key question is which metrics are used to measure that performance and how an IT organization performs against those metrics.

Industry Metrics

This article will look at the two main types of work most IT organizations perform, infrastructure support and software development, and will also consider general metrics that do not fit into either category.

Infrastructure Support Metrics

  • Servers per support engineer. This is an efficiency measurement. According to a study commissioned by Computer Economics, in a company with a small data center, each support engineer should be able to support 5.0 servers.
  • OS instances per support engineer. This efficiency measurement takes into account physical and virtual servers, grouping them together by operating system instances. In a company with a small data center, each support engineer should be able to support 6.7 OS instances.
  • Personal computers per support engineer. This is similar to the previous two metrics, but applied to personal computers. There are many factors to consider, including the computer location, size of the IT organization, nature of the user population and so on. In a small- to medium-sized office, one support engineer should support around 75 PCs.

Software Development Metrics

  • Software engineers per application. COCOMO (Constructive Cost Model) is a formula that has been used for 30 years to compute the size of software project teams for development and maintenance. If a software development team can maintain a specific application with fewer people than the COCOMO index it is more efficient than the norm and vice-versa.
  • Post-release defects. This is an indicator of the development team’s process capability as it reflects the overall quality of the software product(s) delivered. In a perfect world, no Severity 1 (critical) defects should be reported in a system after it is released to the users. A reasonable expectation for legacy systems is one or two critical defects per product release.

General IT Metrics

  • System uptime. This reflects the availability of the company’s information resources, addressing both development and infrastructure capabilities. When no data exists to enable this measurement directly, measure the inverse (system downtime or outages). There are no “absolute” figures available as industry standards, but strategic systems should be up and running whenever needed – as close to 100 percent as possible.
  • Trends over time. This is a reflection of the IT function’s general health. If metrics such as response times and number of critical defects per period rise over an extended period of time, there is probably something wrong in the organization.
  • Customer satisfaction. An effective IT function is best defined as one that meets the needs of its customers, internal and external. There is no “absolute” standard for this metric; however, according to the American Customer Satisfaction Index (ACSI) companies that support software in the United States typically report customer satisfaction rates around 75 percent.

    IT Performance – Recommendations

Taking all this information into account, the following actions are recommended.

  • Create a metrics program. Establishing a metrics program that is reasonable, empirical and transparent is critical—success cannot be recognized if it is not defined. For infrastructure support, create an intelligent monitoring program and seamlessly integrate it into help desk software. For software development, create full traceability from requirements to software deployment. IT performance should be continuously improved.
  • Keep these metrics in perspective. Regardless of your IT group’s performance against these industry metrics, your primary mission is to meet the needs of your customers.
  • Congratulate your IT staff. Most IT teams perform admirably under less-than-ideal conditions. IT professionals want to succeed, and will do so if we give them what they need, be it hardware, software or resources. Create a “culture of success”.

In Summary

IT is a critical component of a company’s success and one that is created and maintained at significant expense. The metrics suggested by this article do not guarantee success. Success is most often achieved when those directing the IT function provide the necessary resources, guidance and encouragement.

About the author: Michael A. Eason has been an IT professional for almost 25 years. He has a broad range of industry experience, including software development, quality assurance, enterprise infrastructure, IT sourcing strategies and software process improvement. Eason has played a key role at global companies such as GlaxoSmithKline, Asea Brown Boveri and PeopleClick. He leads the Information Services organization at the American Board of Anesthesiology, the certifying body for anesthesiologists in the United States and Canada. Eason’s responsibilities include development and maintenance of the company’s enterprise software, as well as systems architecture, process improvement, and the operation of the ABA’s network infrastructure. His professional affiliations include the IEEE Computer Society and the Software Engineering Institute at Carnegie-Mellon University. Eason is a graduate of Augusta State University in Augusta, Ga.