Growing the Customer-IS Partnership in System Development Copyright CAUSE 1994. This paper was presented at the 1994 CAUSE Annual Conference held in Orlando, FL, November 29- December 2, and is part of the conference proceedings published by CAUSE. Permission to copy or disseminate all or part of this material is granted provided that the copies are not made or distributed for commercial advantage, that the CAUSE copyright notice and the title and authors of the publication and its date appear, and that notice is given that copying is by permission of CAUSE, the association for managing and using information resources in higher education. To copy or disseminate otherwise, or to republish in any form, requires written permission from CAUSE. For further information: CAUSE, 4840 Pearl East Circle, Suite 302E, Boulder, CO 80301; 303-449-4430; e-mail info@cause.colorado.edu GROWING THE CUSTOMER-IS PARTNERSHIP IN SYSTEM DEVELOPMENT Susan R. Griffith Assistant Vice President for Planning and Administration Mary E. Cauble Assistant Manager, Information Systems and Services Donald D. Volz Manager, Information Systems and Services Southwest Texas State University San Marcos, Texas 78666 ABSTRACT The System Development Life Cycle (SDLC) process has been adopted at a major public university in the southwest to direct the development of automated and non-automated systems on campus. This paper discusses the problems and opportunities that spurred the consideration of SDLC, and the customer's involvement in the development of policies and procedures, and testing and improving the process, as well as the customer's critical role in implementing the process. The results of early projects and lessons learned are shared with the reader. INTRODUCTION It is now a generally accepted fact that the customer's role is vital to the success of any system development project. However, the customer's comprehension of what is involved in system development is often sketchy at best. Rare is the customer who understands the systems approach to solving problems - to first define the problem, to identify requirements of a solution, to explore the feasibility of alternatives, etc. For too long, information services organizations have sought, assumed, or been assigned too much responsibility for these activities. The lack of customer satisfaction with the results should not be a surprise to anyone. Southwest Texas State University is implementing a Systems Development Life Cycle (SDLC) methodology that strives to address these issues. SDLC identifies tasks to be accomplished and outlines mutually interdependent responsibilities of customers and system designers during the development process. It provides customers an understanding of the work and participation expected of their units in obtaining a desired service or product. It is a single, uniform approach applied consistently to system development projects. The use of SDLC should result in a product which is right for its intended purpose, which meets customer expectations, and which is delivered on time and at a reasonable cost to the University. This paper describes the problems and opportunities that provided impetus for action, the process used in developing the methodology, the policy and procedural documents that evolved, and the results and conclusions to date. BACKGROUND The Institution Founded as a Normal School in 1899, Southwest Texas State University has evolved into a comprehensive public institution with an enrollment of 21,000 students, the 7th largest institution in the state of Texas and the 8th largest in the United States without a doctorate. SWT is located in San Marcos, a rural area halfway between Austin and San Antonio. Internally, SWT is governed by President's Cabinet, comprised of the President, the Executive Vice President, and the Vice Presidents of Academic Affairs, Finance and Support Services, Student Affairs, and University Advancement. The information services function at SWT is known as Computing Services and reports to the Executive Vice President. Computing Services (CS) is comprised of three (3) service areas: Computer and Network Support Services, Information Systems and Services (ISS), and Systems Software Services. The Technology Infrastructure Computing Services administers two (2) computer centers equipped primarily with Digital Equipment Corporation (DEC) hardware. The academic center houses a DEC VAX 8820 and a DEC ALPHA-AXP 7000-640. The administrative center is equipped with a DEC VAX 7000-630 and a DEC VAX 6000-620. The two (2) centers form a mixed mode cluster, with OpenVMS 6.[1] as the operating system. Several MicroVAX, ALPHA-AXP, and other servers and workstations are also housed in these centers. In addition, SWT has implemented a multi-protocol Campus Wide Network comprised of FDDI, extended 802.[3] Ethernets, Novell and AppleTalk LANs. Students, faculty and staff utilize a diverse collection of over 2,500 microcomputers, most of which are connected to the Campus Wide Network. Another 500 VT series terminals are connected to the OpenVMS cluster in various ways. PROBLEMS AND OPPORTUNITIES: IMPETUS FOR ACTION Migration to VAXcluster Between 1976 and 1989, the SWT ISS department had developed applications to address information processing needs in virtually every campus department. Applications ranged from the more typical financial, student, and human resource information systems, to police, housing, library, and physical plant systems. All of these applications were developed in-house on DECsystem-10's, and utilized a common, highly integrated, non-redundant database. With the imminent demise of the DEC-10 architecture, all administrative applications and the entire administrative computing environment had to be replaced. In January 1989, a special Migration Steering Committee, with representation from each university division, was established to monitor the migration, enforce development restrictions, and establish priorities for allocation of scarce resources. Three (3) years later, SWT completed the migration of administrative computing applications from the DECsystem-10's to a VAXcluster platform. Administrative applications were replaced through in-house re-engineering, in-house conversion, or purchased software. Some of the purchased software required extensive customization. The migration was successfully completed in three (3) years due in part to support from the administration. With help of the Migration Steering Committee, enhancements and new development were severely restricted, limited almost entirely to externally imposed mandates. To expedite the migration and support subsequent operations, the administration also funded additional positions in CS and customer departments. There was a down side to the development restrictions: by the end of the migration, pent- up customer demand was at an all-time high. Ineffective Priority-Setting Processes With all this pent-up demand, the issue of setting project priorities quickly became critical. Everyone recognized that SWT could not go back to the pre-migration method of prioritizing requests. Prior to the migration, ISS time had been allocated among the university divisions in fixed hour amounts. Each division had its own priority committee responsible for maintaining a wish list of "things to be done". There were no criteria for measuring the importance of a request, just some generally accepted guidelines like: * If it is Payroll, it must be important. * If it is Registration, it must be important. * If the Vice President said to do it, just do it. * If the Legislature wants it, do it first! These priority committees met irregularly to determine what requests had been completed, why ISS had not completed more of the requests on the list, what new "top" priority items should be added to the list, and how many of the items were now a "number one" priority. With the exception of a few university initiatives, ISS decided which "number one" priorities would be addressed. In 1976, this method of prioritization actually worked well. Automated systems were few in number and relatively simple, therefore maintenance requirements were low and resources for new development were adequate. Over time, the number and complexity of systems increased while resources remained static. As a result, maintenance increased, development decreased and completion of requests became more and more infrequent. Divisions began to jealously guard their time allocations and an informal process developed that benefitted privileged customers. The same customers always enjoyed the lion's share of the resources. No division was willing to dedicate any of its allocation to university-wide projects; however, they still expected these projects to be completed. No division was willing to allocate time for the changes necessary to keep current systems viable. Inconsistent Level of Customer Satisfaction Access to an official priority list or the informal "privileged" list was still no guarantee of customer satisfaction. Many requests for new development or system enhancements were verbal, often a result of a hallway conversation. While customers regarded these verbal requests as the end of their responsibility, ISS expected (and waited for) a little more detail. Whenever these requests fell through the cracks the result was frustration for everyone. Even when a project did get off the ground, customer satisfaction was often low because the accepted methodology for system development did not include customers. Responsibility for all phases of system development rested primarily with ISS. At best, when partnerships were strong and responsibilities were shared between ISS and the customer, good system design practices and a successful project with a satisfied customer resulted. At worst, the customer abdicated responsibility for system functionality and workflow, projects were not successful and the customer was not satisfied. There was definitely a pattern: customer satisfaction was directly proportional to the extent of customer involvement in the process. Inadequate Customer Training and Process Documentation Many customers assumed that if a procedure was automated, then ISS was responsible for both workflow documentation and all training. While ISS provided procedures for operating the software, these procedures were rarely included in any overall office workflow documentation, because no such document existed. By default the procedures for operating the software became the office workflow documentation. The software began to dictate office procedure, customers effectively lost control of their office operations, and the ISS analyst responsible for that application became the expert in customer office operation. While it was generally understood that ISS provide initial training on newly implemented systems, customers expected that ISS continue to provide training whenever necessary. This expectation and the lack of workflow documentation meant that current office staff were not prepared to train new employees without ISS assistance. ISS was frustrated; they did not have the resources to provide offices with procedure manuals and training for all new employees. Customer managers were frustrated, realizing that they had become too dependent upon ISS. Internal Mandates Internal mandates were another problem. "Internal mandates" have implementation deadlines set without adequate data, planning, or forethought. They commit the institution to action without regard to the impact upon stakeholders, projects already underway, and resource commitments already made. A project name and a due date is all there is. A typical scenario might be as follows: The President, Provost, or Registrar promises delivery of a long anticipated Telephone Registration system within a year. Trouble is, the Information Systems Manager and the Telephone Services Manager find out about the "commitment" and the associated implementation calendar in the campus newspaper. Every person on campus has a different definition and expectation of telephone registration, but they all share the same date of delivery. Islands, Orphans, and Cannons SWT is like many campuses in that departments enjoy considerable autonomy with investments in technology. Even the most up-to-date and clearly articulated standards will be ignored in the face of gifts-in-kind and grant requirements. FREE = GOOD, right? Just as night follows day, the result has been the inevitable "Islands of Technology" where hardware and software incompatibilities abound. But not to worry, "The experts in CS can make everything work together, or else they must not be very good." Orphan systems are another common problem -- outsourcing at its worst. A department hires a consultant (translation, Computer Science graduate student) to develop an application they need because ISS is too backlogged. The application is written using obscure languages or tools that the graduate student wanted to learn, but which were otherwise new to the campus. The graduate student documents nothing and cares little about maintainability. Eventually he/she graduates and leaves for parts unknown with the source code. But not to worry, "The experts in ISS can fix it and support it in the future, or else they must not be very good." And let's not forget "loose cannons", the self-appointed experts who know how to do everything they've never done. They can be easily recognized by the way they begin their sentences: "All CS has to do is ..." or "Why can't CS just ..." or "At Utopia U., we did it this way ..." are favorites. But not to worry, if it's been done anywhere else, then "The experts in CS and ISS can and should make it happen here, or else they must not be very good." Internal Audit Findings An Internal Audit finding in February 1991 was critical of SWT's system development methodology. The major issues were that customer involvement was minimal, customer requirements were not documented, cost management was nonexistent, and project schedules had no credibility. The Auditor recommended that the University adopt a formal system development methodology. All these problems led to the inescapable conclusion that SWT needed something significantly better. DEVELOPING THE SDLC METHODOLOGY Forming the SDLC Team To begin addressing the problems identified, the System Development Life Cycle (SDLC) Team was appointed by the President in June 1991. The Team was comprised of at least one member from each of the five (5) divisions at the University, two (2) members of CS, and a representative of Internal Audit ex officio. Faculty, staff, and administrative constituencies were represented. The Team was charged with drafting a policy and procedure document that outlined mutually interdependent responsibilities of customers and systems designers during the system development process, and offering an appeals process for conflict resolution. Adopting a Philosophy The first action of the SDLC Team was to develop a philosophy to guide its work. That philosophy emphasized partnership between customers and ISS. It stressed communication between stakeholders and designers of the system to insure the development of a product which was right for its intended purpose, and which was delivered on time and at a reasonable cost to the University. Adapting a Model The Team began with a model, "Understanding the Systems Development Process" (Long, 1983), that defined system development as a five (5) phase process. In addition, the Team referred to the IEEE software requirements specifications prototype outline (IEEE, 1984), Boehm's book Software Engineering Economics (1981), several articles from the CAUSE exchange library, etc. One of the early challenges was taking a model developed for the corporate world and adapting it to higher education in general and SWT in particular. Some of the differences include: the motive of the institution is not profit, but expansion of services within existing resources; ISS can't politically refuse to work on a project; customers are not accustomed to budgeting for technology. To the Team, the most important elements to be added to the model were well- defined deliverables and responsibilities for those deliverables. With these things in mind and guided by their philosophy, the SDLC team adapted its model to the SWT culture. Expanding the Team's Charge The SDLC model was documented in the form of a university policy/procedure statement. The initial optimistic deadline for a draft document was five (5) months after the SDLC Team was formed, but it was not that easy. In that time, the Team developed their first draft, but circulating it to several customer groups and incorporating their feedback added eight (8) additional months before the revised draft was delivered to the President. In addition to delivering the policy and procedure document, the Team recognized the need for some companion activities to insure success. These included: * Completion of a resource impact study. * Completion of a pilot study. * Identification of SDLC's impact on current policies and organization. * Education of the university community to the process, and gathering feedback from potential stakeholder groups to improve the process. * Development of a separate, detailed procedure manual to allow more flexibility in the policy statement. While the charge to the SDLC Team was initially limited to the drafting of a policy and procedure statement, they knew that successful implementation of SDLC would require that upper management address certain issues. The Team recommended to President's Cabinet. First, the membership and structure of priority setting committees had to be strengthened. Second, an oversight body with a university- wide perspective on resource allocation had to be established. Third, President's Cabinet had to be dissuaded from establishing "internal mandates" without regard to resource availability. Fourth, improved communication and tracking mechanisms had to be developed. The Team was given authority to present the draft policy and procedure to various groups of customers across campus. True to a commitment to customer involvement, over a period of two (2) months, members visited all the division councils, academic support and school councils, plus Faculty Senate, Staff Council, Computing Services, and the area functional analysts. With the comments received from these groups and the results of the first phase of the pilot study, the Team revised the SDLC policy and procedure document, which was officially disseminated in July 1993, two (2) years after the Team was appointed. The Team's charge was expanded to include oversight of the pilot study, development of training curricula, development of the procedure manual, and review of the process's impact on other SWT policies and organizational structures. The SDLC Document The purpose of the SDLC document is threefold: to define the six (6) phases that all SWT system development projects will follow, to identify and assign the attendant tasks and responsibilities, and to specify the responsibilities of the oversight body, the System Development Council. The document also directs the reader to the procedure manual for greater detail. The six (6) SDLC phases, familiar to IS professionals but not to their customers, include: Phase I - System Initiation and Feasibility Study Phase II - System Analysis Phase III - System Design Phase IV - Programming Phase V - Conversion and Implementation Phase VI - Post-Implementation Evaluation Perhaps the greatest aspect of the SDLC document is the delineation of responsibilities to five (5) major groups of players: customers, feasibility study and project teams, area functional analysts (AFA), ISS, and upper management. These groups and the responsibility assignments are described here. _Customers_. The SDLC process is designed to return control to customers and stakeholders by requiring their involvement. They are assigned responsibility for documenting the current system, documenting requirements of the new system, examining alternatives, providing human resources, monitoring progress, reviewing formal requirements and system design, testing and formally accepting the system, completing procedure documentation, and monitoring the system for enhancements and corrections. This drastic increase in the level of customer involvement is the greatest change to prior practice and the greatest risk to success: SDLC will fail if it does not demonstrate that the quality benefits justify the intensive customer time commitment. _Feasibility Study and Project Teams_. Customers are the primary members of both the feasibility study team and the project team, and it is strongly suggested that a customer lead both these teams. These leaders do not need computing expertise. They need to be good at time management, people management, detail work, task coordination and project tracking, and most importantly they need to be champions of the projects. Other members of these teams include ISS and CS representatives, AFA's and other stakeholders identified. _Area Functional Analysts_. While a customer is or should be the team leader, a great deal of responsibility falls on the "Area Functional Analyst" (AFA). This position was created during the VAX migration to act as a liaison between customers and Computing Services. The AFA reports to the customer's department or division, which gives the customer control over more resources specifically dedicated to system development. The AFA assists the customer in preparation, submission, and communication of service requests and plays a leading role on the feasibility study team. The AFA also participates heavily on the project team in detail requirements definition, design review, acceptance testing, and training of customers. _ISS_. ISS now concentrates on screening and categorizing service requests, and in the system design, integration, and programming phases. Basically, the responsibility for project initiation, feasibility study, functional analysis, product implementation, and post- implementation evaluation shifted from ISS staff to the customer. _Upper Management_. Upper management's responsibilities include understanding and judging the relative value of projects, committing the resources at their disposal, monitoring progress, and providing timely policy decisions which affect the process itself. Developing the SDLC Procedure Manual As a result of the feedback received from the community, the policy and procedure document was condensed. It became apparent that the process was complicated enough to warrant an operational manual to help people implement all six (6) phases of the SDLC process. In the manual, each of the six (6) phases of the process is detailed in operational steps. Each step includes a brief narrative of the associated procedure, identification of the parties responsible for and involved in that step, and the deliverable outcomes expected. Due to the rather "laissez faire" attitude of customers prior to the SDLC process, the Team was very concerned about clarifying responsibilities so that ownership of the process indeed rested with the customer. In addition to narratives, flow charts of each phase and examples of deliverables are included in the manual to support those customers for whom "a picture is worth a thousand words". The customers have had the manual for SDLC Phases I and II since December 1993. In Summer 1994, feedback from customers was incorporated and revisions were distributed. The first draft for Phases III-VI is currently in the hands of several readers, whose feedback will be incorporated in the version distributed in December 1994. New ISS Policies and Procedures Early on, the SDLC Team recognized that ISS policies and procedures were poorly communicated and generally misunderstood. Customers simply did not know how ISS resources were allocated or how to get ISS resources focused on their needs. The migration had imposed restrictions, structures and processes that were no longer needed, and the pre-migration methodology had been happily forgotten by all. Customers needed new structures and processes to fill the void, to help ISS serve them. To address that need, and to act as a companion with the SDLC policy and procedure statement, ISS drafted a document entitled "Obtaining Administrative Information Systems and Services". In this policy/ procedure statement, ISS provided a detailed explanation of: * how ISS staff resources are allocated, * how ISS services are requested, and * how projects are initiated and their scope is assessed. ISS Resource Allocations. ISS staff resources are allocated in two (2) ways: by activity type and by activity area. The policy statement defines five (5) activity types and establishes annual targets for the percentage of total ISS staff hours to be invested in each activity type. These types are burden (30%), maintenance (20%), enhancement (20%), development (20%) and research (10%). Articulating these targets has given the customers a more realistic perspective of what ISS does, and how ISS spends its time. The second method of allocation is by activity area. This method identifies the ISS FTE's devoted to specific applications and services. It provides customers with an idea of how ISS resources are spread across the customer base. It tells them how many ISS staff are involved in the support of general database administration, student records applications, human resource applications, internal ISS training, etc. By publishing the ISS resource allocations by both activity type and by activity area, ISS has eliminated some of the "secrecy" and customer suspicion about ISS activities. Customers now have an increased understanding of the ISS resources available, where and how they are invested, and the need for customer prioritization of requests. The ISS Service Request. Past audit findings were critical of the fact that significant modifications were often made to applications without written authorization, and the modifications were often made without input from others who might be impacted by the change. ISS also found that customers were dissatisfied because applications sometimes treated symptoms without solving the root problem, and/or did not meet the customers' expectations. To address these issues, the ISS policy statement defines a process for requesting services that focuses on problem and/or opportunity definition and communication of the request to all stakeholders. The policy specifies that service requests include a title, a goal statement and supporting objectives (by which satisfaction of the request will be measured), a description of the problem or opportunity to be addressed, and a list of known stakeholders. Implementation methods and solutions are specifically EXCLUDED from the problem/opportunity description, and the request must be published to solicit feedback from stakeholders. Area functional analysts and ISS staff readily and regularly assist customers in completing service requests to assure that the problem is stated, not just a solution. In addition, helping customers with requests provides an early idea of the scope of the request and sometimes even eliminates the request if the solution already exists. Assessing the Scope of the Request. A well-written request and feedback from stakeholders provides the information necessary for the customer, the area functional analyst, and the ISS analyst to determine the size, complexity and general scope of the project. This information allows them to get general answers to questions like: * Is this an enhancement, development, or research activity? * How might other customers and applications be impacted? * What needs and opportunities exist for integration? * What are the major constraints? * What customer resources are available to satisfy the request? * Are current hardware/software/network capacities adequate? * To what depth should alternative approaches be pursued? And most importantly, * What SDLC activities and prioritization level(s) are appropriate for this request? RESPONDING TO RECOMMENDATIONS OF THE SDLC TEAM The SDLC Team recommended to upper management several policy, procedure, and structural changes to assure that SDLC worked well. The Team submitted recommendations related to priority setting, an oversight body, mandates, and improved communication and tracking mechanisms. Improving Customer Prioritization and Commitment of Resources SWT is no exception to the shrinking of financial resources afforded to higher education. SWT's student:faculty ratio is 20:1 and its student:staff ratio is 17.5:1, so human resources are extremely limited as well. One of SDLC's goals is to apply these resources where they provide the greatest return on investment in pursuit of the strategic plans of departments, divisions, and the University as a whole. To do this, customers must decide where to invest, not only their own resources, but ISS resources as well. They must actively participate in leveraging these investments to provide the greatest benefit at the least cost to the University. Priority-setting bodies at SWT have improved. They are following the letter and the spirit of SDLC, and are consciously making priority decisions. ISS is no longer in the position of having to choose from among dozens of top priorities. Under SDLC, the appropriate level (department, division, and university) of prioritization is determined by the size, scope and complexity of the project. _Department_. Departmental prioritization is all that is necessary if the customer or requestor, ISS, and the stakeholders all agree that the scope of the request is limited enough to address it at that level. Generally, this means that the request can be satisfied using resources that are readily available to the customer, and with little or no impact upon other departments. A department must set the initial priority of each request it originates relative to its other pending requests. _Division_. If the customer/requestor, ISS, or a majority of stakeholders believes that the request is large enough, complex enough, or has significant impact outside of the requesting department, then divisional prioritization is required. Requests receive divisional prioritization via whatever mechanism is established by the division vice president; this mechanism is usually a division priority committee. _University_. If a request requires commitment of resources normally allocated to offices outside of the requesting division, or has a high risk and/or visibility, or is considered of strategic importance to the University, it must obtain a university-level priority. University-level priorities are established by the System Development Council. Addressing Mandates An issue related to priority-setting which was improved by the SDLC process was the handling of mandates, both internal and external. Internal mandates were set by President's Cabinet (PC); external mandates came from state and federal agencies. _Internal Mandates_. Perhaps the most sensitive recommendation of the SDLC Team dealt with PC's habit of endorsing project implementation dates without regard to the impact of that decision on available resources and current project calendars. PC readily agreed to restrict its endorsements to support for ideas, and to rely on the feasibility study to identify method, resources, and timeline. _External Mandates_. Prior to the establishment of SDLC, customers tended to abdicate much of their responsibility for defining how state and federal mandates would be addressed. Whenever such a mandate was received, the memorandum from the mandating agency was simply forwarded to ISS, along with a directive to "make sure we can comply with this." It was not at all uncommon for ISS staff to be working directly with personnel from the mandating agency, rather than from specifications provided by university staff. SDLC has changed all that. External mandates are now handled in the same way as any other request. A service request that defines the problem, goals and objectives is submitted and must be prioritized like any other request. This change has had many positive results. Customers are aware of the impact of external mandates on other projects and scarce university resources. They are more involved in determining the scope of the mandated project and methods for compliance, and have begun to see the opportunities in external mandates. Where before, SWT might derive no direct benefit from complying with an external mandate, customers now take the opportunity to address local needs at the same time. Establishing an Oversight Body The administration endorsed the establishment of an oversight body for the SDLC process. This body, the System Development Council, oversees system development, prioritizes and tracks university-wide projects, makes policy and procedure revisions necessary as the SDLC process evolves, resolves conflict among divisions, and sponsors the development of SDLC training resources. This Council is chaired by the Executive Vice President, and includes one representative from each division, as well as two ex officio members, the Director of Computing Services and a representative of the Academic Computing Committee. The divisional representatives lead their divisions' priority- setting bodies and have the knowledge and authority to commit divisional resources to projects. Improving Communication and Tracking Systems As one might expect, involving an expanding number of customers created additional communication problems. Priority committees and project teams needed a better way to communicate service requests and track the status of projects. To meet this need, an electronic conference is being pilot tested. The conference serves as a repository for service requests, feasibility study results, and other SDLC deliverables. As projects move through the six (6) phases of SDLC, those with access to the conference will be able to follow the progress of the projects and comment on developments. The conference opens up the process and indirectly increases the accountability of those involved. INITIAL PROJECT EXPERIENCES Below are snapshots of a few early projects which have used the SDLC methodology. These experiences helped reinforce the belief that imposing structure on the decision- making process (SDLC) would result in a product which better satisfies the customer. Pilot Project SWT's pilot project was the emergency tuition loan program management system. The members of the feasibility study and project teams were initially resentful of the amount of time they had to invest in the project by following SDLC guidelines. However, when the project was complete they acknowledged that they had a better product than they would have had using previous methods. They were forced to talk across departments and divisions and solved problems that previously would not have been discovered until much later. Student Health Center To paraphrase a Health Center staff member, "I thought SDLC was only a bureaucratic hoop. But after two (2) previous attempts at automating our system, the feasibility study experience forced us to define and document the different but related needs of doctors, nurses, receptionists, pharmacists, and others on the Health Center staff. And for the first time we had a comprehensive document to use in evaluating solutions." Budget Development The SDLC process has helped to control some rogues who would advance hidden agendas at the expense of solving the problem. Feasibility study team members used the SDLC process to maintain focus on the problem and kept the University from making an expensive mistake. Payroll The feasibility study team for the electronic payroll deposit project involved university employees in the design of a new earnings statement. SDLC helped the Payroll Office recognize that they are not the user of the earnings statement. For the first time employees had an opportunity to define what information they needed from their payroll system. CHANGES IN ATTITUDE Several major changes in attitude and expectations have been among the most important outcomes of the SDLC process. First, there is broader campus understanding that customer satisfaction (quality) is directly proportional to: * The level of customer involvement. * Communication of stakeholder and customer expectations. * The mix (Stakeholder types) of involved customers. * The leadership of one or more customers as "Champions". Second, the implementation of any campus-wide methodology of this magnitude is a long-term commitment of resources. As a result: * SDLC policies and procedures are recognized as continually evolving. * Attention to customer support needs are addressed through continuous investment in the procedures manual and training activities. * Priority committees have increased in number, size and representation and are becoming leaders in making SDLC work. CONCLUSIONS: SOME RULES FOR SUCCESS "CUSTOMER FOCUS", "SATISFYING THE CUSTOMER", "EMPOWERING THE USER". These are some of the buzz phrases of the '90's, especially within TQM. These are worthy goals, but they can only be accomplished in systems development when management grants and the customer accepts responsibility as a full partner. SDLC is not a perfect solution. Neither was MBO, nor is TQM. It is up to the people to make any of these work. As a result of the SDLC experience, SWT has accepted certain truths as basic to the understanding and operation of the process. Reviewing the following "pearls of wisdom" before developing a similar SDLC process may help others save time, organize debate, and limit unrealistic expectations: * SDLC does not create new resources; it just uses them more effectively and more efficiently. * Prioritization is often a determination of what won't get done. * Some worthwhile projects will never get done; customers will still hear "no", but now it will be "no, because ...". * Customers will have to spend more time planning and managing projects. * Priority committees, and especially their chairs, are critical to making the process work, and these individuals will develop necessary expertise at different rates. * It will take time and training to turn customers into project managers. * Estimates will be awful at first but will improve with experience. * Focus on the "WHAT", i.e., delineate WHAT deliverables are needed. * Focus on the "WHO", i.e., delineate roles to be filled - WHO does WHAT * Allow flexibility with the "HOW", i.e., formats and methods. At SWT, the SDLC process in its early implementation has proven successful due to the involvement of the customer, the reliance on a tested decision-making process, the development of simplified procedures for guidance, and a degree of flexibility in choosing from among the suggested SDLC steps and the ordering of those steps. In addition, the institution is better able to establish realistic expectations for project development within limited resources. At SWT, it looks like SDLC is working. REFERENCES Boehm, Barry W. (1981). Software Engineering Economics. Englewood Cliffs, N.J.: Prentice-Hall, Inc. Institute of Electrical and Electronics., Inc. (1984). IEEE Guide for Software Requirements Specifications pp. 18-24. Long, Larry E. (1983). "Understanding the Systems Development Process", in Manager's Guide to Computers and Information Systems, pp. 257-296. Englewood Cliffs, N. J.: Prentice-Hall, Inc.