Reoccurring cost pressures made establishing a sound accounting and reporting structure at Chalmers University of Technology (Chalmers) instrumental in motivating the IT governance change needed to focus on the right cost-cutting measures. A successful CIO-CFO relationship laid the foundation for benchmarking, year-over-year reports and forecasts that demonstrated progress and kept the change momentum.

The basic accounting and reporting process enabled:

• A cost consciousness that affected daily decisions on buying IT equipment, and had a result on the bottom line
• A culture of comparison that facilitated an internal dialogue about the consolidation of IT service delivery
• A focused discussion about strategic cost-cutting measures for the whole institution, rather than blanket cuts of the central IT organization alone

• Work with your head of finance (CFO or similar) to establish:
  • A "chart of accounts" that captures at least the basic IT costs throughout the institution
  • A reporting procedure that enables fair internal benchmarking and year-over-year analysis
  • Continuous communication and education through established communities to keep focus and show progress
• Do not forget to establish other measures than cost, such as service levels, so that a meaningful price/performance dialogue can be upheld.

A proven CIO strategy for avoiding blanket cost-cutting demands from your fellow senior executives is to know not only the IT organization's numbers, but also the IT numbers of the whole institution. However, transforming from an anecdotally driven institution into a data-driven institution takes a lot of time and work. This research presents a case study where basic IT-cost accounting and reporting were built from the ground up and how that enabled the CIO to shift from being defensive about IT costs to having a more constructive dialogue about the strategic yield of IT.

In our contacts with clients, we continuously come across the need to move from anecdotally driven decisions to data-driven decisions, and yet we find that surprisingly few institutions have even basic reliable financial data at hand. Nowhere is this situation more true than regarding noncentralized IT spending, which continues to put higher education CIOs at a disadvantage when trying to motivate campuswide initiatives. During tough financial times, this problem is often amplified, because it is difficult to cut costs that you do not know about, while central IT gets charged with blanket cuts that are visibly followed-up on. This case study is about how Chalmers went about creating a fair, cost-effective and quality-assured reporting process for IT costs that had an impact on the bottom line, as well as keeping the momentum in strategic change.

Even though this case study has its beginnings in 1998 and, in IT terms, can be considered rather old, we have decided to publish it, because the problem it describes is still not fully solved in the majority of institutions with which we talk. Furthermore, the current recession cycle makes real-world experiences from the former recession cycle — the dot-com bust — even more relevant and timely. This research is part of Gartner's extensive research on cost optimization (see www.gartner.com/it/themes/economy/economy_100.jsp ).

Chalmers, founded in 1829, is a top research university in Sweden. Chalmers' turnover is more than €200 million — of which 70% is related to research, and the remainder is for education. Chalmers has about 2,400 employees (including Ph.D. candidates) and 7,500 full-time equivalent (FTE) students.

IT at Chalmers grew internally, with little central control until the mid-1990s, when the hype around IT, with phenomena such as the World Wide Web, indicated a strategic use for IT. Even with this insight of potential strategic value, the decentralized nature of the institution, and the common problem of academic freedom morphing into administrative freedom led to high resistance toward institutionally aligned and integrative IT solutions.

In 1998, the organizational structure was decentralized, with 10 schools as the main units for budgetary control. IT at Chalmers was delivered by more than 20 in-house IT organizations, with no coordination. Initial analysis of the financial data available showed IT costs to be substantial in absolute numbers (approximately €16 million) and as a percentage of turnover (approximately 9%), even though it was understood that all IT costs were not captured in that analysis. The trend was an increase year over year, and the costs were approaching the levels of brick-and-mortar facilities. At the same time, an IT risk assessment, performed by Arthur Andersen and based on BS-7799, showed, in general, that "a number of serious risks has been identified, that can impact the completeness and accuracy in information and usage of IT." More specifically, Chalmers scored 0% in the area of "organization and governance."

This caught the attention of the senior management team and triggered a series of events that would eventually lead to a working governance ecosystem — of which, IT-cost control was one of the five main pillars (see "Case Study: The Seven-Year Journey From Chaos to Order at Chalmers University of Technology").

The external circumstances in the beginning of this effort included an era of financial growth (dot-com boom) and a reactive, mostly unplanned investment frenzy in all that related to IT, even within the university. The main purpose of the governance efforts in this era was to get control of the massive investments and increasing cost levels, including spiraling wages. However, the financial recession of 2000 (dot-com bust) turned the focus to cost cutting and a need for tough prioritizations. But, irrespective of good or bad times, the challenge remained the same: Know the yield of your resources, which starts by understanding the full cost (see "Financial Slowdown Emphasizes Budget-Planning Effectiveness in Higher Education: But It's About Yield, Not Cost").

The first round of fact finding consisted simply in identifying what kind of data was already in the financial system. Fortunately, a number of accounts existed in the chart of accounts that related to IT costs. Unfortunately, even though there were more than 40 of them, including "6612 special computer paper," there was no account related to IT-staff costs. It soon became clear that other IT-related costs were missing, such as physical network and overhead costs, such as administration and facilities rent (e.g., data centers and offices). A conservative estimate of IT-personnel costs based on 60 FTEs showed that IT wages alone were 20% of the overall IT costs, which were at least 7% of the total turnover and rising.

To get as complete a picture as possible of the employees involved in supporting IT, the CIO teamed up with the director of HR and launched a combined survey of IT personnel and wage benchmarking activity, which included regional, public and private organizations. The survey instrument, benchmarking model and analysis were performed by an external company, Löneanalyser AB. The results showed that 151 people, corresponding to 110 FTEs, were working to support IT in some capacity. The wage costs alone (including taxes) were €4.3 million, almost exactly double the first estimate. As a result, the CIO, CFO and HR director all agreed to implement a new account (4034 IT staff) and include "IT staff" in the monthly institutionwide report on employee categories.

A side effect of the benchmark was a good understanding of the competitive landscape for IT talent in the region, where, surprisingly, Chalmers — as an organization that was essentially public in nature — competed well, contrary to the general perception among the IT staff.

The two previous actions, together with the disturbing IT-risk assessment from Arthur Andersen, convinced the senior management board to invest in an IT-controller consultant with previous experience of looking at IT costs in major decentralized organizations. A deeper dive into the books of, among others, the general ledger and the "book of assets" produced the next set of numbers, which showed IT costs as at least 9% of turnover and that wages comprised 30% of IT costs. Furthermore, it was clear from discrepancies in the inventory and interviews with administrative personnel that there was a lack of consistency in accounting due to lack of guiding principles and training. This resulted in questions about overall data quality.

These results finally convinced the senior management board to approve the "IT plan" project, which had the ultimate project objective to "create the foundation for high yield by making it transparent how IT services are ordered, implemented and maintained," which was to support the institutional objective of achieving "high yield from Chalmers' IT resources." The key metric at hand to which the latter was evaluated was total IT costs as a percentage of turnover. The measurable target was initially set to just breaking the trend of increasing IT costs.

The initial phase of the IT plan project spent a lot of time building a basic governance framework (the second of the governance ecosystem's five pillars), and educating decision makers about their roles and the impact of their decisions. At the same time, a major Y2K-related implementation of a new financial system drew a lot of resources. This resulted in an understandable lack of resources to focus on the data quality issue. However, soon, work on the third pillar of the governance ecosystem, services, demanded higher-quality data — financial and service-level data — to enable fair internal benchmarking, especially because there was a strategic intent to open up an internal market of IT services between the existing IT organizations.

With the implementation of a new financial system successfully accomplished, there was renewed interest in financial data quality overall (see "Drive Data Quality Improvement From a Foundation of Metrics"). During the revision of the chart of accounts, the CFO and CIO decided to reduce the number of IT-related accounts from more than 40 to the 13 shown in Table 1. Preceding this decision was a discussion about detail versus quality. The need for quality won and resulted in the approach to have as few accounts as possible. This change was communicated within the normal channels of financial support. The far-left column shows typical coding in a Swedish chart of accounts. Understanding the logic behind the numbering was crucial in gaining the trust of the financial administrative community. The far-right column shows a high-level grouping of the 13 accounts, which was used for senior executive reporting.

Included in this iteration of quality enhancement was the development of new standard reports and new reporting tools for ad hoc reporting. Among the results in this iteration was the awareness of how much of the IT costs were outside central control. In this case, it was 73% (2001 data).

The increasing awareness of the size of the local IT costs, together with further consolidation work toward higher yield driven by the IT plan, resulted in a high demand for accurate data to base decisions on. During 2001, this was also amplified by cost savings made necessary by the recession (which quickly reduced the institution's endowment). This resulted in intensified collaboration between the CIO and the CFO, with a threefold focus:

• New financial model: A new financial model for capturing all the costs of local IT organizations (including those omitted from the accounting chart), such as administration, employee training and facilities costs (e.g., data centers and offices), as well as electricity and cooling. This was developed in collaboration with two schools with the involvement of the deans and the local finance director. The model was jointly presented to all local finance directors.
• IT in the budget process: An increased focus on IT in the new budget process increased the visibility of IT costs and the awareness of alternatives for sourcing IT services.
• Continuous training: A document was developed that instructed other stakeholders about the use of IT accounts and financial reports, including information about the reasoning behind the new financial model for local IT. The CIO had personal dialogues with local finance directors, and a joint meeting was held with all the local IT managers and finance directors.

At this point, the relationship between the IT community and the finance community, in general, and the CIO and the CFO/local financial directors, in particular, had been sufficiently developed to facilitate collaboration. The collaboration was (and had to be) one of mutual respect and compromise to deal with the many iterations of improvements that followed.

There was a broad impact of this approach, which had a major part in the reduction of the IT costs that, at their peak in 1999, were 11.2% of the total turnover. The costs were reduced to 6.9% within five years. The impact can be categorized in three main areas:

• Cost consciousness: There was simply an increased awareness of IT costs that translated into caution in IT spending. A lot of local and individual decisions resulted in buying right-sized computer equipment for the task at hand, rather than oversized computers for potential "future" needs and introducing simple measures, such as "handing down" older computers to be used for simpler tasks for another year. The life span of computers and other IT equipment was reviewed and realigned with financial rules of depreciation. Extending the life of most desktops and network equipment, and shortening the life of laptops, displayed a more accurate asset value. The increased awareness also led to increased use of existing common procurement processes and contracts for IT equipment. This was probably the major factor in short-term turnaround in the IT-cost trend.
• Culture of comparison: The availability of the data and the sudden ease of (basic) reporting, together with a strategy of service orientation, at first triggered a spur of internal benchmarking. The first realization that less than 30% of the IT costs were under central control took considerable heat off the CIO and turned the focus to overall service delivery efficiency. Further dialogue about the cost, scope and yield of existing IT services fostered a collaborative approach among the IT staff. This, together with a strategy of an open internal market, led to several consolidations of internal IT organizations. The internal benchmarking was extended to external benchmarking within the higher education community and, eventually, with external service providers. This latter development included further refinement in IT cost reporting models to secure fair comparisons. But the external benchmarking activities also showed how much less effort Chalmers needed for obtaining data than most of its peers. This, together with the apparent usefulness in decision making, convinced senior management that the efforts had been worthwhile.
• Food for focus: From a CIO's point of view, the availability of accurate and comprehensive cost data was a tremendous asset for focusing the senior management on a strategic dialogue, rather than a detailed discussion of tactical measures. Data, such as the example shown in Figure 1, provided background for discussing specific strategies, such as improved procurement processes or the degree of outsourcing for cost reduction, rather than giving the IT department the task to cut costs by 5%.

These better-informed decisions resulted in higher compliance and better overall yield. Other examples include:

• Discussions about IT costs in research (personal productivity) versus administrative IT costs (organizational efficiency) led to standardization and cost reduction, where possible, to have flexibility and investments where necessary. This also led to further refinements in costing models for research, where containing direct research-related IT costs within the research project became a guiding principle (see "IT and OT Interaction Gives Rise to New Governance" and "IT and OT Convergence: How to Get Started on the Journey").
• Consolidation of e-mail services on campus but with retention of the service in-house, despite a 10% benchmarked cost-premium, kept a critical mass of skill within the IT organization.

Reporting IT costs gave the CIO a simple tool to show progress in the overarching work to improve governance and the yield of Chalmers IT resources, as well as to pinpoint areas of improvement. The availability of data gave the CIO the credibility needed to keep the momentum of change.

• Ensure that there is CIO-CFO collaboration: The key success factor was the collaboration of the CIO and the CFO, especially the willingness to try to understand each other's goals and having respect for each other's domain-specific knowledge in designing the path to those goals.
• Keep it simple: Realizing that IT is not the center of everyone's attention, and sacrificing detail for data quality and actual usage by simplifying the accounting and reporting process, was another critical success factor. Having reports in minutes rather than days meant that they were used for benchmarking and real decision making. Furthermore, quality increased because people used the data they put in and identified any mistakes quickly. Transparent data enabled trust and sustainable decisions.
• Spread the word: Building personal relationships not only with the CFO and the central reporting team, but also with the financial directors of the schools, was crucial. This also gave access to existing communities, such as that of the local financial directors, which greatly facilitated communication, training and acceptance of the new financial model. Coordinating and aligning changes with other major changes in financial accounting and reporting reduced the effort in achieving change.
• Think through the metrics and change them: A good chart of accounts greatly simplified the production of key performance indicators, but care needs to be taken to ensure that all major cost areas are represented. Start with a simple set to ensure year-over-year comparison. Acknowledge and discuss flaws in the model, and strive for continuous improvement when more-complicated comparisons (e.g., with services) need to be made.
• Have a vision: A final critical success factor is to have a vision of the end result. In this case, the goal of better "yield of Chalmers' IT resources" put the whole vision of the governance ecosystem in perspective, enabling iterative progress on all five pillars of governance: data, decision arenas, services, processes and accountability dynamics. Each pillar had to be built alongside the other, collaboratively supporting each other by balancing the maturity levels.

• Find out for yourself: It was realized that accounting and reporting processes serve many purposes, and visualizing IT cost is usually not one of them. This is especially true if you are a public university and use processes designed by central agencies. Its accounting and reporting focus lies more often in controlling utilization degree of funds and accuracy of tax payments than in any kind of "business outcome." Furthermore, you have to realize that not everyone welcomes transparent reporting. For example, a widespread use of Ph.D. candidates as IT support staff as a means to finance their positions was in direct contrast to the president's guidelines. This was unexposed until the new rules of IT accounting came into effect, thus exposing a hidden IT cost and a flaw in Ph.D. student financing. Both issues were reluctantly acknowledged by departmental heads. You have to ensure that accounting and reporting fulfill the intended goals and do not relax until they do. Introduce the role of an IT controller, or at least use external help from benchmarking experts in designing your accounting and reporting process. However, be careful not to make it too complex.
• Never stop training: There was always a fight against entropy. The tendency for academic leadership and organizations to be in constant flux means that you can never stop educating key individuals as they move into roles that affect the governance ecosystem. The need for data quality means that finance administrators need to be kept up to date.
• Bring home the benefits: Do not be too overconfident in the financial data. Ensure that you can bring home the benefits before you act on an apparent cost-saving opportunity. For example, can you really consolidate that quarter of an FTE or reuse that data center for something else?
• Determine what goes into the nominator and denominator: The sudden availability of some hard cost data oversimplified some benchmarking discussions and focused too much on cost. This was fine, as long as the benchmarking was internal and all uncounted costs, such as data center costs, were comparable. However, when benchmarking with external services was introduced, this became an issue that demanded refined accounting. A more serious lesson learned was about how to get from cost focus to yield and define the price/performance ratio of services. Even though standardized service-level agreements were introduced early in the process, they were not standardized enough to compare with externally available services. Furthermore, the most important indicator of service level — user satisfaction — was not measured until very late in the process.

"Case Study: The Seven-Year Journey From Chaos to Order at Chalmers University of Technology"

"Financial Slowdown Emphasizes Budget-Planning Effectiveness in Higher Education: But It's About Yield, Not Cost"

"Toolkit for IT Executives: Developing Higher Education Benchmarks"

"Drive Data Quality Improvement From a Foundation of Metrics"

"Cost Optimization: Classify Costs to Achieve Sustainable Savings"

"Establishing Financial Credibility"

"Financial Management Is Key to IS Organizations' Success"

"Managing the Finances of the IS Internal Service Company"

"How to Define IT Financial Management Competencies and Roles"

"IT and OT Interaction Gives Rise to New Governance"

"IT and OT Convergence: How to Get Started on the Journey"