Cost (and Quality and Value) of Information Technology Support in Large Research Universities
by Christopher S. Peebles and Laurie Antolovic
Sufficient data can be cited to support the following proposition: "Information technology (IT) adds value to the work of most enterprises." Equally abundant data, which Paul Strassmann offers in his book The Squandered Computer,1 support the opposite conclusion: "IT destroys value in most of the enterprises where it has been deployed." Poorly designed applications, uninformative World Wide Web (WWW) pages with their inevitable broken links, opaque decision-support systems, and poor-quality hardware not only can frustrate users but can destroy value and increase costs by consuming large, unproductive blocks of users' time. One of the crucial defenses against such destruction of value by IT is a strong organization for the education and support of those who use computers and who gather information from various network resources.
The cost and the perceived quality of user support services are another matter. Emily Paulsen's article in Support Management in 1997 asked the question: "What Does [IT] Support Really Cost?"2 The answer is, in large measure: We do not know. A similar question can be posed about the quality of IT support, and the answer is similar. User perceptions of the quality of IT services have not been measured in most enterprises, including higher education.
To measure organizational performance, Robert Kaplan and David Norton developed the Balanced Scorecard as an antidote to the narrow focus and outright "fadism" of many management methods and tools popular in the 1970s and 1980s. Their approach, backed by case studies, was published in a series of articles in the Harvard Business Review, and a full presentation was made in 1996 in their book The Balanced Scorecard.3 A subsequent book by Kaplan and Robin Cooper, Cost and Effect, amplified the strategic and operational uses of financial measures within the broader suite of the Balanced Scorecard.4
As its name suggests, the Balanced Scorecard measures the performance of organizations in terms of their strategic plans and operational goals along several dimensions. It assesses the internal health of the organization and measures such items as the length of the product development cycle, the ability to reach order-of-magnitude changes in the effectiveness of processes, the rate of improvement in the quality of products and services, and benchmark comparisons against competitors. It looks at the renewal of intellectual capital in the organization, the effectiveness with which these abilities are allocated, and the rate at which new products and services are developed and deployed in comparison with the rates of the competition. The Balanced Scorecard, like the Baldrige Award criteria, takes special note of customers and their perception of quality (in Joseph Juran's terms,5 the quality of a product or service equals its fitness for use and its freedom from defect). One of the most relevant measures, beyond customer satisfaction, is the number of successful cooperative efforts -- customer partnerships -- created in a year and sustained over the years. The final dimension, the financial perspective, takes Activity Based Costing (ABC) and Activity Based Management (ABM) as central to strategic and operational measurement as well as to control of the production of IT services.
Traditional accounting systems reveal the financial state of an enterprise at some point in time so that the enterprise fulfills its legal requirements and so that it can be compared with other firms. These accounting systems serve neither operational nor strategic roles at levels below those of major divisions and product lines within an enterprise. They can value inventory and distribute the costs of plant, equipment, and management over major products, but they cannot specify the costs required to run a process that produces discrete products and services. Moreover, such accounting systems cannot provide the information through which processes can be reengineered to reduce cost and increase quality. ABC and ABM -- the former measures the real costs associated with a product or service, the latter uses these costs to catalyze quality improvement and cost reduction -- were designed to do just that: to identify processes and products that add value and those that destroy value.
ABC measures and allocates all the costs that are incurred in the production, sale, and after-market needs (warranty, for example) of a product or service. Once costs have been determined, the processes that produce these goods and services can be redesigned for greater efficiency. Moreover, products and services can be designed to minimize production costs and maximize quality when they are in the planning stages (for example, consider the changes in the North American automobile industry over the last twenty years, or better yet, look at the design for Amazon.com). Relationships with customers can be established to minimize the cost of serving customers while at the same time increasing the quality of the service they receive; relationships with suppliers can be changed from adversarial to cooperative, and joint engineering and just-in-time delivery can be instituted (again Amazon.com can serve as a positive example). Without knowledge of the "real" costs incurred throughout the life cycle of these goods and services, such efforts would not be effective.
In the case of a college or university IT organization, all of the "products" are services. Moreover, as most of the readers of this article will know all too well, incremental increases in demand for services are not met with proportional increases in revenues. Higher education institutions all deal with "flat" budgets as a matter of course. Our organizations might receive some increase in resources from time to time; however, whatever additional funds are offered usually are not tied to levels and quality of service but are based on some arcane formula or new initiative that appeals to legislatures, trustees, and deans. In effect, although tied to the traditional missions of the university, these financial decisions are based on traditional accounting systems that serve the same functions in higher education as they do in the business world. They offer a financial condition report that is good for accountants and is within legal requirements but is not much use for strategic planning or operational management of the educational institution. Thus, given effectively static funding, the only way that college or university IT organizations can offer either increased levels of services or new services is to eliminate some current services or to reduce the unit costs of those services provided currently. Measurement of the cost and the quality of these services is a necessary first step.
University Information Technology Services (UITS) has measured the quality of IT services offered on the Bloomington campus of Indiana University (IU) for the last decade. This quality survey was designed by the UITS Center for Statistical and Mathematical Computing in collaboration with the IU Institute for Survey Research. Administration of the survey itself is done exclusively by the IU Institute for Survey Research, guaranteeing objectivity in and anonymity of respondents. ABC began for IT services offered on the Bloomington campus in 1995.6 The quality survey, which draws from a sample of 2,000 students, faculty, and staff on the Bloomington campus and which gets more than a 40 percent return of the questionnaire, asks users to rate each service offered by UITS on a scale of 1 to 5. On this scale, 1 = very bad and 5 = very good. The summary measure used here, user satisfaction, is the percentage who used the service and who rated it either 3 = acceptable, 4 = good, or 5 = very good.
The components of the ABC measures are sketched in Figure 1. The methods used in these calculations are those advocated by Kaplan and his colleagues, described by John Shank and Vijay Govindarajanin their book Strategic Cost Management, and ably summarized by Jeremy Hope and Tony Hope in chapter 6 of their book Competing in the Third Wave.7 Wages and benefits are fully allocated to the design, deployment, and operation of each service throughout its life cycle. If individuals work on more than one service, they are asked to estimate the percentage of time they spend on each. These estimates are usually accurate within a few percent and are sufficiently accurate for ABC calculations. The costs of training are allocated to appropriate individuals and services and are registered as a cost in the year they are completed.
The hardware that delivers the service and the hardware that supports those who deliver the service are depreciated over a three-year period, and one-third of the total expense is allocated to the services the hardware supports each year. Operating systems, system management software, and applications software are taken as an expense in the year they are purchased. Likewise, software license fees and maintenance contracts for hardware and software are distributed among the costs of services they support each year.
Where apposite, charges for telephones and voice networks are allocated to particular services. The costs of local area networks are distributed among the services they support. Costs for the campus networks and for Internetworking, on the other hand, are calculated on a per-packet basis and taken no further. For example, an e-mail message "costs" less than $0.0017 per message (almost 300 million messages were delivered on the Bloomington campus in the last fiscal year), but that figure does not include Internet and campus network costs (together less than $0.02 per thousand packets). At the moment, the intercampus video network stands on its own, and the ABC figures are given in a cost per hour for its operation (approximately $100 per hour).
In general, the direct costs for producing these services have been covered in the categories discussed above. However, some costs span the entire organization, and some costs are not included in the UITS budget but are borne by the campus and the university. The former we have designated "Organization Sustaining Activities," and in the choice of this term we follow general ABC practice. Included among these activities are the costs of the senior leadership of UITS -- the vice-president and the CIO, the associate vice-presidents, and the deans -- plus the Human Resources Office and the Business Office. The costs for these offices and their internal services are distributed proportionally among the several external services. If a service, such as e-mail, receives 10 percent of the total budget, then it is allocated 10 percent of the total costs for the Organization Sustaining Activities. The ABC figures in Table 1 are inclusive of these sustaining costs.
These ABC figures do not include the cost of facilities and utilities. IU does not bill UITS for the buildings it occupies, the maintenance of these buildings and their grounds, and the cost of electricity, heat, and chilled water. If the ABC figures given here are compared with ABC figures for private-sector companies (few of which offer public benchmarks), then 13 percent must be added to the UITS costs. This percentage is not arbitrary but has been derived from an IT industry average for facilities and utilities costs.
Measures of cost and quality for central IT support on the Bloomington campus can be apportioned among three major areas: (a) education and (b) support for those who provide support for departments, institutes, and schools; (a) education and (b) support to assist users in supporting themselves; and a General Support Center, which serves as a foundation for the first two elements. This trio is the foundation for the "Leveraged Support Model" developed by IU.8
The costs for these services can be presented in a manner that is consistent with traditional accounting practice but that is generally useless for strategic or operational management. For example, in the last fiscal year, UITS allocated a bit more than $500,000 for education and $3.5 million for support on the Bloomington campus. Put another way, UITS spent approximately $12 per capita for education and $86 per capita for support (Table 1). These aggregate and per capita amounts describe neither what services were produced nor what their quality might have been. They are fit neither for assessment nor accountability by those who ultimately paid these bills: the citizens of Indiana and the students and their parents, who paid through tuition and taxes. The changes we advocate focus on services rather than technology and on measures of the cost, quality, and value of these services to our scholarly community.
The customer base for IT services on the Bloomington campus includes 35,000 FTE students and 6,500 FTE faculty and staff. The number of individuals, rather than FTEs, is greater than 48,000. Of this total, 99 percent use computers and network services routinely, and 77 percent have computers in their place of residence. Although entering undergraduate students are not required to own a computer (campus Student Technology Centers have more than 1,400 machines, most of which are available around the clock), almost all will have to use e-mail and the WWW in their first term. The STEPS classes offered by UITS are designed to introduce students to the major applications -- e-mail, word-processing, spreadsheets -- that they will be required to use as part of their curriculum. These courses, which last between ninety minutes and three hours, are offered either in UITS facilities or, at the request of the instructor, as a part of the regular academic course. For the last several years there have been more than 10,000 separate registrations for these courses each year. These courses are very cost-effective (about $21.00 per student) and are highly rated by those who take them (with a satisfaction score of almost 94 percent). They are doubly effective because they introduce the students to the UITS Knowledge Base -- a collection of more than 10,000 answers, solutions, and links to common questions and problems that can be searched in terms of key words via a very friendly Web front-end -- so that they can further help themselves. Similar courses are offered to the campus community as a whole, with similar cost and quality measures (Table 1). Seminars on current topics in IT are offered for the campus professional computing community, and they too cost about $20.00 per participant.
The most effective components of the education program are the courses designed for Local Support Providers (LSPs). Now some 200 strong on the Bloomington campus, these support personnel, who are paid by the units they serve, work with faculty, staff, and students in departments, centers, and schools. Courses in Intel (Windows and NT), Apple (MacOS), and UNIX system administration and hardware configuration are given on a regular basis. These rigorous courses last one week and, on the participant's successful completion of the exam at the end of the course, provide certification of knowledge in platform and operating system fundamentals. Although the UNIX and MacOS courses are locally developed and their certifications have local significance, the Microsoft courses are part of the Microsoft Certified Support Engineer (MCSE) program and their certification is recognized internationally. Most departments now require such certification before they will hire a local support person. The average cost for this training is $278 per student (UNIX costs a bit more, MacOS a bit less), and 88 percent of those who have taken a course rate it as adequate or better. In turn, users' satisfaction with their LSPs is above 85 percent. The support of LSPs by the central IT organization goes far beyond this training and greatly amplifies the effectiveness of both local and central IT support personnel.
In addition to LSP training, a bit more than $900,000 per year is dedicated to LSP support. The Departmental Support Laboratory offers expert advice, maintains servers with common software, operating system, and system management tools, conducts software and hardware evaluations, and performs technology audits and IT plans for departments and schools. This figure comes to approximately $4,600 per LSP and $3,800 per department, center, institute, or school. These services are all designed to make the LSPs more effective and to reduce support costs overall. Moreover, the collective choice of recommended computers, peripherals, and applications covered by site licenses goes a long way to reduce support and training costs.
There are five specialized user-support organizations within UITS. Some are partnerships with other units, in which case the ABCs are those incurred by UITS activities, not those costs borne by the partners. Other support services are wholly within the UITS organization.
- The Data Management Support Group offers its services to faculty, staff, and research students. Services include the support of local data managers who employ such data-management tools as large relational databases (e.g., Sybase and Oracle), midlevel and desktop database engines (e.g., Microsoft SQL Server, Access), and desktop data-extraction tools (e.g., spreadsheets, client/server development, Web applications). This group supports a growing number of research and pedagogical users of Geographic Information Systems (GIS) and is the home of second-level support for a variety of Web-authoring tools (e.g., FrontPage). Although the group's primary function is support of the users of these tools, it also administers and supports the central pedagogical database environment, used by faculty in teaching, and the use of these tools and databases by students across the disciplines. This work with the wider IU community costs about $38 per contact, and 89 percent of the group's customers are satisfied with the service they receive.
- The UNIX Workstation Support Group provides specialized consulting with UITS and offers its expertise to the wider campus community. In this latter role it offers front-line consulting on the Bloomington campus. The cost for this service is about $33 per contact. Extended consulting costs over $100 per contact, and technical evaluation of hardware and applications costs almost $4,000 per project and report. The group's WWW server, which has UNIX support material, was accessed more than 7 million times in 1998 at a cost of $0.0055 per Web hit. More than 91 percent of those who use the UNIX Workstation Support Group are satisfied with the service they receive.
- The Electronic Text Resource Center (LTRS), a partnership with the IU Libraries, offers applications and operating system support as well as specialized hardware for humanities computing. Short-term consulting costs approximately $14 per contact; classes cost $93 per student; use of the center's hardware and applications costs nearly $14 per session; use of its collections of texts, bibliographies, and multimedia materials, which are available over the WWW, costs about $0.05 per use. Of those who used LTRS last year, 98 percent rated it good or better.
- The Center for Statistical and Mathematical Support, a partnership among the Graduate School, the Department of Mathematics, and UITS, offers general consulting on statistical and mathematical software and maintains the site licenses for software such as SPSS, Mathematica, and SYSSTAT. It does a brisk business in short-term consulting and works with faculty and graduate students on the design of courses and on the use of statistics and mathematical applications in research and dissertations. The short-term consulting costs approximately $35 per session; the work on courses and dissertations costs a bit more than $2,000 per project. Slightly more than 97 percent of those who use this center are satisfied with its services.
- The Teaching and Learning Technology Laboratory (TLTL), a partnership with the Dean of the Faculties Office on the Bloomington Campus, works with faculty and graduate students as they create and code course material for delivery via the WWW and in the classroom. An IT specialist, a specialist in teaching and learning, and a faculty member work together to create these multimedia course materials. During the last fiscal year, TLTL participated in 265 projects at a cost of more that $1,500 per project. More than 95 percent of those who use TLTL are satisfied with the service they receive.
The five support services discussed above are all "handcrafted." Each project is unique; each problem is recondite. Each UNIX problem, by the time the user is ready to ask for help (because UNIX users are a hearty and independent lot), is truly vexing. Each database question is complex because the consultant has to understand not only the application but the database schema as well. Each use of a new collection of texts and concordance programs is demanding and requires expert assistance. Thus the costs for solutions to even the most casual problem and for answers to even the smallest questions are on the order of tens of dollars. IT support costs for large projects in these areas run into the thousands of dollars.
Whereas the specialized support centers deal with a few people each week, the General Support Center serves hundreds of people each day, and its Knowledge Base (http://kb.indiana.edu/) offers online help to hundreds of people each hour. Whereas the specialized support centers employ a few IT professionals and a few part-time hourly student assistants, the General Support Center employs thirteen full-time professionals and up to thirty part-time professionals and students. Given limited funds, the General Support Center is open only from 8 a.m.to 5 p.m. Monday through Friday. A request for support for additional evening and weekend hours has been made and has some chance of being approved in fiscal year 2000.
Each year individuals initiate approximately 150,000 contacts with the General Support Center. Some walk into the center, and others telephone or send e-mail. The IT Knowledge Base received more than 3 million "hits" last year. In terms of costs (Table 1), the Knowledge Base is "dirt cheap." If one figures four clicks to an answer, then problems can be solved for approximately $0.44. Walk-in consulting costs $4.04 per contact, telephone contact costs $5.26 per call, and e-mail problems cost $7.04 per contact. The difference in cost between telephone and walk-in contacts is subsumed, in large measure, by the cost of the telephone. The root cause in the greater cost of e-mail consulting is the number of back-and-forth exchanges required to get a clear statement of the problem and a succinct solution for it.
The yearly quality measures for these four general support services are likewise instructive. The highest level of user satisfaction, 92 percent, goes to the walk-in service. Telephone consulting and the Knowledge Base produce approximately 86 percent satisfied customers. Consulting via e-mail satisfies only 79 percent of the customers. Because the same group of individuals with the same training and knowledge offer all these services, face-to-face communication with the walk-in customers is likely the cause of the additional 6 percent in customer satisfaction. When e-mail is taken as the benchmark, there is a 13 percent rise in satisfaction; this is a result of both the medium and the messenger placing the customer at greater social distance.
The current manager of the General Support Center, Sue Workman, has instituted a daily quality check to give real-time rather than once-a-year data. Each day thirty-five customers who ask the support center for help are chosen randomly and sent an e-mail that asks the following three questions: Did you get a solution to your problem? Did you receive this solution in a timely manner? Were you treated courteously? The response rate to this quality check is close to 100 percent. For the last question, that of courteous treatment, almost 100 percent answer "Yes." The question of a timely response gets approximately 97 percent "Yes" answers. The question about solutions averages near 90 percent "Yes," but the e-mail service receives only a 79 percent "Yes" response regarding solutions. The Knowledge Base also has a quality check. At the end of a session the user is asked, "Did you get an answer to your question?" Of 8 million "hits" and perhaps 2 million customers who used this service in the last twenty-four months, of those who took the time to answer the question, fewer than 3,300 answered "No." Each "No" answer is reviewed by General Support Center staff and leadership to discover if changes in the Knowledge Base are required by adding new topics to those covered already, by increasing the amount and accuracy of information available, or by refining the method by which information is delivered to the user.
These cost and quality data form the foundation on which UITS engineers and manages its organization and services. These concepts of service, quality, and value are at the center of the UITS culture. Distributed support and LSPs offer users familiar, friendly, and knowledgeable support services backed up by a fast path to the specialists and specialized software inside UITS. Although costly, the custom-made support for faculty and research students creates value far in excess of the costs. The general measures for the perceived quality of these services run from approximately 85 percent satisfied users to over 99 percent satisfaction. Though we would like all quality measurements to be in the mid-90s (actually we would like all customers to be 100 percent satisfied), we would be remiss if we did not stress that these measures are almost 20 percentage points higher than the average given, for similar services, in national satisfaction surveys. Although far from personal, the Knowledge Base, which is available twenty-four hours per day, 365 days per year, is nonetheless effective and generally user-friendly. It has received a number of national awards for excellence in online IT support.9 Clearly, e-mail is neither a friendly nor an effective way to deliver IT support. The center is trying to redirect those users seeking support from e-mail either to access the Knowledge Base -- which offers the option of e-mailing any question not answered -- or to make a telephone call to the General Support Center to speak directly with a consultant. E-mail use is a holdover from the days of inefficient response to telephone inquiries and of an abortive short-term strategy that sought to downsize the General Support Center in the late 1980s. It also reflects both the reliance of the user population on e-mail for communication and the lack of extended evening hours of telephone support. Though we want to eliminate e-mail support, rather than doing so by dictate or by use of negative incentives, we will do so by making the Web-based Knowledge Base even better than it is today and by expanding the hours of telephone support as we fold in support of students in residence facilities within the larger UITS support operation.
This mix of support services and the partnerships forged with departments and schools do add value to the work of students, faculty, and staff at IU. Without the focus on users helping themselves through education and through resources like the Knowledge Base, however, IU could not afford to offer timely and knowledgeable IT support. In the past fifteen years, the number of computer users on the Bloomington campus has grown from a few hundred to more than 50,000 and the number of computers from a few to more than 50,000 as well. Over the same time, the computing support budget, when adjusted for inflation, has remained more or less flat.
The cost and quality measures presented here can degenerate into mere "bean-counting" and serve as the means for invidious comparisons. Conversely, they can become central empirical elements in the adoption of the Balanced Scorecard and ABM. If the latter occurs, then they are measures of progress toward the strategic goals of the IT organization. They provide feedback and link individuals, processes, and services with the goals and the culture of the organization. In brief, they are part of the text for strategiclearning and management.
First, the foundation of all IT services is the intellectual capital that the staff brings to work each day (and that leaves with them when the staff goes home, usually late in the evening). Thus the learning and growth of the skills and knowledge of the staff is fundamental. Second, the ways that this knowledge is combined with networks, hardware, and applications to produce IT services is highly dependent on the ways that processes are organized and on the culture of the organization. What is valued and attended to internally is vitally important to what is offered externally. If processes are tracked for their ability to create value, their ability to design quality into their operations, and their capacity to respond to novel demands, then they will offer services that are fit for use and free from defect. Third, customers are always the best judge of the quality and value of the IT services they are offered. It is important to measure their satisfaction with current IT services and to use their desires and visions as catalysts for the design, development, and deployment of new services. Finally, we all work with limited resources and ever-increasing demands for IT services. Knowledge of the real costs for each IT service -- using the financial measures that are at the core of ABM -- makes the choices among services rational and the improvement of services mandatory and measurable. Recall the words spoken by the character of John Glenn in the movie The Right Stuff when he threatened to go to the media if changes were not made in the Mercury capsule so that the astronauts would be pilots rather than passengers: "No bucks -- no Buck Rogers." University and college IT organizations are in the same position: No resources -- no cycles, storage, and bandwidth. The measures we use to assess the services we offer, and the ways these facts are presented to our communities of customers and other stakeholders, are crucial to the future of IT organizations in higher education.
Endnotes
1. Paul A. Strassmann, The Squandered Computer: Evaluating the Business Alignment of Information Technologies (New Canaan, Conn.: Information Economics Press, 1997).
2. Emily Paulsen, "What Does Support Really Cost?" Support Management (May/June 1997), 14 - 22, 113.
3. Robert S. Kaplan and David P. Norton, The Balanced Scorecard: Translating Strategy into Action (Boston: Harvard Business School Press, 1996).
4. Robert S. Kaplan and Robin Cooper, Cost and Effect: Using Integrated Cost Systems to Drive Profitability and Performance (Boston: Harvard Business School Press, 1998).
5. Most recently, in sections 2.1 and 2.2 of Juran's Quality Handbook ,5th ed. (New York: McGraw Hill, 1999).
6. A full public accounting can be found at http://www.indiana.edu/~uits/business/ for the ABC and at http://www.indiana.edu/~uitssur/ survey/index.html for the quality survey.
7. John K. Shank and Vijay Govindarajan, Strategic Cost Management: The New Tool for Competitive Advantage (New York: Free Press, 1993); Jeremy Hope and Tony Hope, Competing in the Third Wave: The Ten Key Management Issues of the Information Age (Boston: Harvard Business School Press, 1997).
8. This trio is fully described in Brian D. Voss, "Coalition for Networked Information -- Institution-Wide Information Strategies Project 1997 - 98: The Leveraged Support Model," paper delivered at the CNI Spring Conference, April 1998, Washington, D.C. (http://www.indiana.edu/~ucs/tlit/cni/a.html). For more information, see also Christopher Spalding Peebles, Laurie G. Antolovic, Norma B. Holland, Karen Hoeve Adams, Debby Allmayer, and Phyllis H. Davidson, "Modeling and Managing the Cost and Quality of Information Technology Services at Indiana University: A Case Study," in Richard N. Katz and Julia A. Rudy, eds., Information Technology in Higher Education: Assessing Its Impact and Planning for the Future (San Francisco: Jossey-Bass Publishers, 1999).
9. Among other awards, the Knowledge Base was recognized by PC World Magazine as among those resources offering the "best free stuff online," was named "Distinguished" in the International Online Competition hosted by the Society for Technical Communication, was judged one of the top technical support sites by Windows Magazine, and was designated a "Cool Site" in the Netscape Open Directory Project Web site.
