IS departments are increasingly organised as internal businesses that provide products and services to internal customers. But realistic executive customers can be a hard sell. They listen, they think and, like the educated consumers they are, they inevitably ask: "What does it all cost?"There are no easy answers.
Internal customers won't easily buy into esoteric or theoretical advantages of a technology but insist on knowing what benefits they will directly or indirectly reap from a particular technological change. Whether a technology change is proposed by the executive director of a line of business who vaguely understands the complexity of a request or by the director of an IS department, IS executives are expected to deliver a holistic real cost of ownership (RCO) model that uses business-based metrics to evaluate and communicate a technology's effect on the enterprise's bottom line.
Two kinds of costs and benefits must be evaluated in all cost of ownership models. "Steady state" costs and benefits are ongoing. While it is possible to quantify the steady state IS costs, it is often difficult to quantify the steady state benefits that IS produces. Instead, IS departments should attempt to quantify only the second type of costs and benefits, those that are incremental or brought about by a particular migration or change (for example to 32-bit Windows or fast Ethernet). In many cases, it is possible to take intangible benefits such as ease of use, faster response times and wider data access and transform them to quantifiable benefits such as head count reduction, increase in productivity or decrease in cost of goods sold. But, like it or not, IS managers must frequently make a subjective "gut feel guess" of what a project is worth by comparing qualitative benefits with quantitative projected costs.
Cost of ownership models often inappropriately mix hard and soft costs, use incompatible assumptions to compare alternative architectures and ignore the benefits that accrue with ownership. A complete value analysis supplies an IS manager with the tools to determine internal pricing for services, benchmark operations, improve efficiencies, evaluate outsourcing options and justify capital investment in both infrastructure and applications. To determine RCO, IS managers must separate those items that can be quantified easily from less tangible factors and consider them in separate models.
Quantifying the costs
Capital and operational costs are the major expenditures for which IS is most directly responsible and on which IS managers can have the most effect. In contrast, end-user costs do not appear on IS budgets and are the most difficult to measure and change. A popular practice lumps together both cost categories to identify total cost of ownership; however, this approach, which combines fundamentally different categories into a single model, is ill-advised for three reasons. First, so-called hidden user costs, such as shadow support, the loss of productivity when one employee trains another, are often deemed equal to "hard" costs, inflating the cost of technology in the minds of users. That inflation is unacceptable in an enterprise that makes the IS department the seller and users the buyer. Second, our analysis shows that many of these hidden costs are not actually reduced under any managerial scenario. The amount of time users spend initialising, maintaining and operating their systems remains relatively constant; the actual financial variable is the return or benefit derived by a constant investment of user time. Therefore, since user time with most systems remains invariable, those associated time costs are better calculated as part of productivity rather than as a technology cost issue. Third, we have found a gap between what IS managers say and what executive customers hear. An IS manager who is selling a system improvement argues that implementation of a particular technology will reduce overall costs by reducing hidden costs. Executive customers often hear only cost savings and expect a reduction in IS budgets. Never, heaven forbid, their own. (See "Using RCO", page 14) Gauging RCO To establish a baseline to measure costs, RCO analysis must consider the following: Standards. Multiple platforms are difficult to manage and can double operational costs while reducing user productivity because of the complexity of testing and integrating systems.
Application portfolio. Application mix includes the type, number and availability requirements of the application portfolio. Certain applications, such as airline reservation systems, have a much higher priority within the business and, because they are more critical, require a higher level of support.
User population profile. Because the economies of scale have an enormous impact for many operational costs, any RCO model must account for the number of supported users and their geographic dispersion. For geographically dispersed organisations, if no IT staff is present, operational tasks must be outsourced at premium rates. Organisations should negotiate warranty agreements to include on-site service for all locations.
Service-level requirements. Guaranteed response and resolution rates for help desk calls have to be included in calculating the RCO, but the cost of delivering actual systems needs to be included as well. Some lines of business will inevitably request more frequent application or system updates than other lines of business. Fortunately, this factor is the most negotiable with the internal IS customer base.
IS management infrastructure. Automated management tools, such as electronic software distribution, hardware and software inventory tools, remote control/diagnostics and help desk trouble ticketing, and knowledge-based tools are the most effective ways to reduce IS costs without eliminating the perceived value IS provides to business units. Organisations should look toward automated management processes that will provide a significant amount of centralised control yet offer enough user flexibility.
Successful IS organisations must move beyond the generic vendor-hyped cost of ownership models to develop a systematic methodology to measure the real costs and benefits of technology. In determining RCO, IS executives must focus on measurable costs that they can directly control, factor in the actual application requirements for the target organisation, compare various architectures with an eye toward the cost of migration and, finally, emphasise the less tangible benefits of the change.
David Cearley, senior vice-president and service director of Meta Group Inc's workgroup computing strategy service, can be reached at firstname.lastname@example.org Using RCOCompare the TCO of NCs with that of standard desktops Real cost of ownership (RCO) analysis should be used to compare architectural alternatives at the client, server and network levels. For example, this approach to cost model strategy is important when evaluating the potential impact of network computers (NCs) as a way to reduce overall cost of ownership.
Meta Group Inc has researched global companies that have a 10,000-user organisation distributed across multiple geographic locations with minimal standardisation and systems management infrastructure. Their expenditures on equipment (capital costs) and human resources (operational costs) make the cost of a typical distributed LAN environment nearly $US2800 per year per desktop.
Traditionally, capital costs seemed insignificant when compared with operational costs. However, our research shows that when costs of clients, servers, networks and applications are considered, capital costs virtually equal operational costs for the life of the system. Therefore, asset management strategies that minimise direct costs and seek to extend technology life cycles will have a significant impact on overall IS costs. An optimal strategy must find the balance between paying for the latest and greatest and maintaining and caring for dinosaurs. The operational costs (approximately $US1400 per PC per year for a typical PC/LAN environment) are measured on the full-time equivalent personnel needed to provide a given level of service. This RCO model shows these operational costs can be more than half of total costs (see "PC/LAN Costs").
It is no wonder that the promise of the "zero maintenance" NC has sparked IS interest. In fact, a pure NC model with thin terminal-like clients and heavily managed or controlled environments will save 30 per cent in overall capital costs and 26 per cent in total costs over a typical unmanaged PC environment, according to our research. At first glance, the NC would seem to be an obvious choice for those organisations wishing to reduce costs. However, using the principles of RCO, further analysis shows that some costs will increase in an NC environment. With greater demand placed on server and network resources, these costs will offset key parts of the NC savings. In addition, the NC model assumes that IS totally redesigns all its applications, a process that will likely prove costly and painful. Finally, the NC model assumes IS can tightly manage the user environment, including the "lock-down" of all client devices, disallowing any user installation of hardware and software. (See "NC System Costs") Given the tantalising potential of the NC and the significant risk and additional cost factors, IS organisations are more often looking at a middle ground. Using the RCO models and an existing LAN environment, IS can approach NC-like savings by moving toward a managed PC environment. User surveys show that this middle ground increases standardisation while limiting user installation options and can reduce overall costs by 12 per cent - approximately one-half of the target NC savings.
- D Cearley
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.