[<-- Previous] [Home] [Contents] [Next -->]

Support quality deteriorates

In the absence of a significant change in approach, the mismatch between demand and capacity produces a death spiral of decreased quality of service and support. Continuing the University of Virginia example, contention for dial-in lines is so fierce that individuals set up attack dialers and then camp on in an attempt to assure continued access. During the first month of the academic year, the call-waiting queue at the help desk is longer than the line at a Grateful Dead concert. Information technology staff are frustrated because they cannot meet needs, despite the long hours they invest at work. Once viewed as heroes, these same people are now considered incompetent because they cannot handle the flood of requests. In an attempt to serve, help desk staff implement automated solutions to providing help that irritate customers who want the personal touch. We do not have time to plan successful implementation and support of new technologies or to communicate with customers about changes. Instead, we rush new services out the door. Because they are not production-quality services, they further increase pressure on the help desk, and the spiral continues.

It is easy to understand that if demand is drastically increasing and support resources are not, both the quantity and quality of support will diminish. Given the difference between demand and supply on most campuses, perhaps the most amazing observation is that our support mechanisms have not entirely collapsed-perhaps because some users have developed their own support mechanisms.

Some say we are approaching crisis, when the data indicate that we should actually be in meltdown. In addition to the simple mismatch between supply and demand, qualitative changes in the technology environment of our institutions also contribute to the deteriorating level of support. Exploring the mechanisms and characteristics of technology support will provide a better understanding of the problem and will ultimately lead us to a possible solution.

Centrally provided primary customer support does not scale

Primary support consists of answering user questions directly with the customer. This contrasts with secondary or expert support in which a group of staff provides information resources, tools, deep expertise, and backup for those who provide the front-line, primary response.

Historically, when users were fewer and more homogeneous, the computer center provided most of the primary support, often employing expert consultants and programmers to do so. As the population of users has changed, this practice has become both ineffective and inefficient. It is ineffective because no single consultant can possibly know all the answers that the wide spectrum of users requires. The variety of user information needs spans such a broad range of technology services and computing applications that users, especially new users, often have difficulty communicating the critical elements of their environments in terms that experts can use to answer their questions. As we scale up the number of these questions, it also becomes inefficient to use highly skilled experts to provide primary consultation, such as "Have you turned on the power switch?" Instead we need people on the front lines who are trained in diagnosis, more like primary care physicians, rather than specialists. A single pool of staff will be less good at answering front-line questions for a heterogeneous population of thousands of users than the same number of staff who are divided among the users in smaller sets, with the opportunity to learn the capabilities and environments of their individual subsets of customers. Thus, we believe that the current centralized primary support model in many institutions is doomed to collapse.

One exception we see is centrally provided, subject-specific consulting and service centers, such as the Center for Mathematical and Statistical Consulting at Indiana University, the Academic Computing-Health Sciences Center at the University of Virginia, and the High Performance Computing Center that is now getting started at the University of Houston. These centers provide primary and expert consulting and often system and network administration, and they seem to be very successful in meeting their users' needs. However, unlike general help desks or consulting pools, their domain is very focused in content and they serve a small subset of the entire range of users on campus. These centers seem to be the exceptions that prove our rule.

Assignment of support responsibility is ambiguous

As central resources fell behind in their ability to meet demand, users responded in two ways. Those who assumed it was the responsibility of the central technology organization to provide all of the answers and help they required took the position that it was the IT organization's job to teach, rather than their responsibility to learn. Departments and individuals purchased hardware and software that did not meet their requirements and then expected central support to make it all work. In many cases, the IT organization has promoted these misperceptions by attempting to convince the institution that it had complete responsibility. Some distributed units, however, responded by developing their own support mechanisms that functioned quite independently of the central services. Many of these have not been able to keep up with the increasing complexity of technology or to interconnect their idiosyncratic environments with the rest of the campus and world. The central organization's response to these units has often been to simply ignore them, relinquishing the ability to influence users' decisions and learn from their experiences (both bad and good). In either case, both central technology staff and the users perceived the quality of support to be significantly diminished.

Distributed systems need special support

Problems with local area networks, desktop client applications, and remote access are sometimes impossible to troubleshoot from a distance. With no one else to turn to, the user resorts to central technology support providers. This no-win situation causes frustration for the customer and serious demoralization for the support staff who try valiantly to find solutions from afar. Unable to do so, they eventually make office calls or house calls. This wonderfully personalized service typically alienates other customers unless there are enough support providers to handle everyone's needs this way.

Every machine in the institution is different

In the early 1980s, microcomputers were limited in processing capability. Software was likewise very limited in functionality. Under these conditions, the choices we had to make significantly determined the degree to which our machine actually met our needs. No machine could solve all of our problems, so we chose the one that came closest and that we could put on our desk soonest. In addition, we funded our equipment through donations, grants, begging, and every one-time method we knew. Compounded over a decade or so, the result of these practices-especially at large research universities-is an assemblage of equipment, software, and configurations that is nearly insupportable at any reasonable cost.

Central units are merging

There has been a trend over the past five years to consolidate academic and administrative computing organizations and, in many cases, also telecommunications, media services, and libraries into a single organization. These mergers are probably the right thing to do, but they require redefinition of identity and responsibilities for staff and users and major reorientation-both technical and cultural. Precisely how this affects the effectiveness and efficiency of the support staff and the user's ability to get needed support is not clear, but it may contribute to a sense of frustration. On the other hand, there are synergies and new knowledge that result from these organizational changes that promise, at least eventually, to outweigh the transient problems.

[<-- Previous] [Home] [Contents] [Next -->]