Business Analysis and Management of Change: Applying SSM to the CONFIRM Case

Executive Summary

The failure of CONFIRM, a state-of-the-art information system, was to be developed by AMRIS, the IS wing of AMR, the parent company of American Airlines. They were the managing partners in a consortium formed of four companies. The other three were Hilton Hotels, Marriot and Budget Rent-a-car. AMRIS was to design and develop the new IS which would serve as the single comprehensive reservation system for the three other companies and would enable them to improve their business and profitability while at the same time help them retain existing cost structures in terms of cost per reservation. But once the project started, it got delayed. In the end, AMRIS had to fire some employees, and the project was ultimately shelved. Intrico was then disbanded and even lawsuits between the erstwhile partners resulted. However, the companies settled these lawsuits out of court. Thus, although this case was similar to other IS development failures, yet, due to the huge costs and scale involved, this case was unique. Various systems thinking approaches exist, which could be applied to analyze the reasons for the failure and also to gather knowledge for future use in other, such projects. However, in the CONFIRM case, since human activity systems are basically predominant, and due to the undefined and complex nature of the problems, it is recommended that a soft or unstructured methodology like the Soft System Methodology or SSM due to Checkland be applied in analyzing the case. This report thus attempts to apply SSM principles to the case and evolves a conceptual model as different from the real world situation in the case. The same are then presented as recommendations to the steering committee as a solution to many of the problems facing the project.

Introduction to the CONFIRM Case

Organizations often feel the need to change over from old information systems (IS) to new ones. Depending on the size and nature of the organization, developing and implementing new information systems within any organization entail substantial cost, labour and time. But often, IS projects are never completed or even if finished much beyond scheduled time, require costs way beyond budgets initially estimated. A case in point is the failure of CONFIRM, an IS which was sought to be developed by Intrico, a consortium of Hilton Hotels Corporation, Budget Rent-A-Car Corporation and Marriott Corporation. This IS was envisaged as a modern, advanced and comprehensive reservation system for providing up to-date information on hotel lodging, car rental and flight booking. The IS project was actually sub-contracted to AMR Information Services Inc. (AMRIS), the IS arm of AMR, the parent company of American Airlines Corporation. The companies spanning three different industries and the IS sector felt that such a system would facilitate a state-of-the-art comprehensive new reservation system as part of a single computerized infrastructure that could be marketed globally, bring in enhanced profits at existing cost levels, and provide a strategic advantage to the company using the system. AMRIS, which was designated as the managing partner in Intrico, clearly felt that it could tap the existing demand for such a single reservation system for the car rental, hotel lodging and airline booking sectors for effectively enlarging and diversifying its products line and also hoped to bring in constant revenues while functioning as the data processing arm for the system.

Issues and Tasks

AMR already prided itself on its successful SABRE system for online flight reservations. It hoped to duplicate this system, albeit in a larger scale. Thus, as managing partner of the four member consortium of Intrico, AMRIS, the IS wing of AMR, started work on the CONFIRM project in May 1988. The entire project was to be completed in two stages (design and developmental) by end of Jun 1992 at a cost of USD 55.7 million with operational costs per reservation pegged at USD 1.05. But by the time the project was ultimately shelved and the consortium itself was disbanded in Jul 1992, almost USD 125 million had been spent on the same. The case also resulted in lawsuits, which ultimately were settled out of court by the companies.

The failure of the CONFIRM IS raised uncomfortable questions. One was that relating to ethics. The managements of the companies forming Intrico obviously were driven by profit considerations, for maximizing returns for their stockholders. However, the role of top management of the companies in the CONFIRM case evidently points to their deficient ethical conduct. Whatever it is, the employees at AMRIS, the clients for the system (the three partners of Intrico) and the management of AMRIS, all were remiss in complying with professional codes of conduct. Oz (Oct 1994, p. 34) observes that there appeared to be three reasons for the CONFIRM failure. One, there were unseen technical hitches that could not be resolved. Two, the estimates of costs and deadlines were not realistic. Three, the system developers failed to comprehend the user requirements and were also hampered by changed requirements post start-up.

Whatever the reasons for the CONFIRM failure, an examination of the case can yield concrete analytical results than can determine forecasting and prevention of future IS development failures. In particular, such huge losses from a failed project similar to CONFIRM can only weaken management’s abilities to minimize costs, improve profits, and consistently evolve and implement successful IS projects. While many systems methods are available to analyze problems faced by management within organizations, this analytical effort examined the CONFIRM case by applying the soft systems concepts called Soft System Methodology due to Checkland (1972, pp. 87-116)

Choosing the Analytical Model

Information Systems (IS) developments require substantial investment by way of time, labour and money. However, it has been found in the course of various research studies that many IS projects are cancelled prior to completion. Also, many projects overshoot the budgeted costs by huge amounts, even in case they are completed. One study has found that 31 per cent of new IS projects get cancelled prior to completion and that, of those projects that do get completed, a substantial 52.7 per cent overshoot the initial budget by a phenomenal 189 per cent (PC Week 16, Jan 1995, 68). Another survey has found that a massive 75 per cent of all information system development projects were not completed, or even used when completed (Gladden, 1982, pp. 35-39). The failure of CONFIRM, perhaps, is very similar to those other unfinished organizational IS efforts. But in terms of waste of funds, the failure of CONFIRM was a colossal failure, made the partners of Intrico suffer a huge setback and ultimately the consortium had to be disbanded. Obviously, managers of organizations are continually seized of the need to contain costs, adhere to project deadlines, and manage the changeover from old systems to new ones. A need is therefore felt to analyze failed IS projects so that human learning can be enhanced for predicting and preventing such failures.

While the approaches to analyzing information system failures are many, the systems thinking were traditionally observed as hard and could not solve many real world problems of organizations. The soft systems methods evolved in this backdrop. One of the most prominent soft systems approach was that due to Checkland (1981). Checkland envisaged Soft Systems Methodology or SSM as a holistic approach that conceptualizes rich pictures in depicting problem situations.

Lyytinen (1987, pp. 5-46) states that the SSM is a system engineering methodology applied in action research settings for resolving problems in the real world. While there are other systems engineering approaches as applied to business problems including the Viable System Model (VSM), Business Process Reengineering (BPR), and Business Process Transformation (BPT), these are ridden with inherent problems of their own. For one, these traditional methods could not be satisfactorily applied to organizational problems which were caused by purposeful human activity, were subject to the different views on the problem by different people and were also unstructured and difficult to define. Two, the application of such methods itself created further problems since the method of inquiry interfered with the outcome of the study. Third, when the problems were to be pinpointed too early, it was difficult to see more basic or different problems. Thus, in complex real world situations like the CONFIRM case, a soft model like the SSM is considered better adapted for proceeding on the line of inquiry as to the nature, reasons and solution to the problem situation. The CONFIRM case involved various human players or stakeholders. The problem concerned the human situation requiring purposeful action to be taken to resolve issues. Accordingly, SSM could be viewed as the most appropriate methodology for examining the CONFIRM case in as much as the same entail an enquiry of human activity system (HAS) and since the various perspectives of the different stakeholders need be considered in as broad a manner as possible.

Analyzing the Problem Situation in the Case, Key Findings

CONFIRM was the dream project of four leading companies each in their own field. The issue involved the development of an advanced multi-user, multi-industry reservation system. Once implemented, it would provide a single comprehensive reservation system for the clients of Hilton Hotels Corporation, Budget Rent-A-Car Corporation and Marriott Corporation. The project was conceived by AMRIS, the IS wing of AMR, the parent company of American Airlines Corporation. AMR itself hoped to take advantage of a perceived demand in the market for the product when it found that only 20 per cent of clients were found to book their reservations online for hotel lodging whereas their own airline reservation system SABRE handled 80 per cent of clients in that industry. The four companies formed a consortium called Intrico in which AMRIS was the Managing Partner. While the hotels and car rentals would use the IS for daily operations, AMRIS was to be aided by the other three partners to help market a tailor made version of the system to other potential businesses. AMRIS was also to be given the task of maintaining the data processing operations. The entire project was to be done in two phases. The first phase was the design phase and this would take seven months to be completed. The second phase would be the developmental phase, which would take another 45 months to be completed. Thus, the project would run from 1988 to 1992. While AMRIS would help design and develop the IS, implement the same and also help maintain processes, the other three would get a state-of-the-art reservation system which could help improve their bottom lines but keep the costs of operations at existing level. There was also the added incentive of returns from marketing the new IS system to prospective clients. The partners of the consortium would bear a total cost of USD 55.7 million for the project and the cost per reservation would be USD 1.05. The partners agreed to maintain a resident team each at AMRIS headquarters in Dallas where the project would be designed.

AMRIS consistently maintained that the new IS would be very superior to existing ones and that the costs entailed would also be less compared to the existing structure. But, although it maintained that the project would be completed on time, the project started getting delayed. Gradually, the cost of design started shooting up. The functional specifications also were not as desired by the individual companies. Marriott in particular expressed its objections to the specifications outlined in the initial development design. The AMR personnel who developed the hugely successful SABRE system substantiated this and stated that the specifications were too general and were insufficient for the developer to understand the user requirements. The first plan brought out by AMRIS thus had to be revised and this meant that the estimated cost of designing the plan was scaled up to USD 72.6 million with the cost per reservation expected to be a higher USD 1.30 in the first year of full operations while the same would be 0.72 and 0.40 in fourth and fifth years respectively. AMRIS gave the others an option to withdraw once the development was completed at a penalty of USD 1 million. In mid 1989, Marriott found from an inspection of pro-forma financial statements of AMRIS that the company had misrepresented its number of reservations, understated operational costs, and derived erroneous per reservation cost which was one of the most crucial information for viability of the project for each of the three other partners. In spite of the fact that the companies estimated the cost per reservation to be around USD 2, the companies carried on as part of the project. In Feb 1990, the project could not pass the BAA deadline (a phase of the project). The developers then redefined the part of unfinished phase as a part of the next phase. AMRIS was not forthcoming on project details to Marriott and maintained that the project was on schedule as revised. Even when its internal problems began to surface, AMRIS maintained to the other partners that the project was on schedule and also could be completed within the new budget. Even when it missed the deadline for launching the terminal screen in Jan 1990, it kept on stating that it could still make up for lost time. It was 13 months behind schedule by Mar 1990. Yet, the AMRIS top brass insisted that the project was on schedule. Even the AMRIS employees voiced their doubts on whether the project could be completed on time. They also expressed their objections to the way their managers handled things. When the company top brass understood that the managers had lied to them on the project details, they fired 8 top executives and later another 15 employees. This was in Apr 1992. In Jul 1992, the total costs crossed USD 125 million, much above the initial estimates and the project had to be shelved. In fact, the consortium Intrico itself was disbanded.

The CONFIRM episode also resulted in some legal cases between AMRIS and the other three companies. Each side maintained their version of the sequence of events that resulted in the failure of CONFIRM. In this respect it need be mentioned that the agreement between the partners had clearly mentioned as the stated objective of the project that:

1. The CONFIRM would be an advanced IS that would be designed, developed, operated and maintained as also marketed worldwide for profit.
2. The system would provide a single comprehensive reservation system for booking hotel lodging, car rental and flight tickets.
3. All interfaces would be developed by AMRIS, which would also help convert the old information systems of all the partners to new ones.

The entire responsibility for design and development vested in AMRIS. The teams from each company stationed at AMRIS headquarters would provide inputs for functional development and evaluate the system even as it was under design. AMRIS in return warranted to other partners that the costs would not be over budget and that the project would be completed by deadline.

In spite of false representations by AMRIS, Marriott still gave the AMRIS management further opportunity to complete the project. Its management tried to maintain two divergent perspectives. One was that as a user of the proposed system they did not believe that the project could be completed. But as a partner of the consortium, they hoped that AMRIS could complete the project. Two divergent roles of the same entity mutually in conflict was thus in evidence. Even Hilton followed the line followed by Marriott. In a communication the chairperson of AMRIS stated in Apr 1992 that its operational managers had been inept and had also concealed information on technical hitches and work performances of employees. They also stated that the technical staff had not been successful in designing the database systems and the complex interfaces between sub-systems. AMRIS top management then said that they would refund 100 per cent of partner investments at this time, if the partners so wanted. They squarely blamed some employees for suppressing vital information so that the developed system was neither viable nor could any progress be made even in 2 years time.

At the time of disbanding, the AMRIS president admitted that the data processing facility was incompatible with the decision support system of the project. Even the database was found irrecoverable in the event of a system crash. However, in course of legal proceedings that followed the disbanding of Intrico and filing of law suits and counter suits by the erstwhile partners, AMR maintained that the other partners had changed original plans for determining the specifications of the single reservation system and also stated that the companies had then asked it to provide three separate reservation systems under CONFIRM. The role of the various persons from AMRIS and other partner companies, particularly the management staff is questionable since they showed a lack of ethical and professional principles. Also, another failure could be concerning the timely communication and sharing of right information among stakeholders. Whoever was to blame for the project failure, the fact remains that substantial money and time was wasted without any result.

Some Relevant Systems Considered

The CONFIRM case, as stated before, can be suitably analyzed by applying the principles of the Soft Systems Methodology (SSM). Obviously, the case involves the interactions between the various human stakeholders and involves the question of ethics. Also, the project requires strong functional management communication to be maintained for it to be consistently evaluated for compliance with user requirements and improved upon. The structure of the partnership formed is also open to debate as each partner was not able to work in the best interests of itself and the consortium. Also, cost control assumes vital significance since the project had to be completed economically for all the partners to retain and improve profitability. Various systems can be visualized in relation to the problems. Simply put, a successful information system development project is based upon the effectiveness of various organizational systems as outlined below.

1. A management information gathering and two-way communication system for obtaining effective feedback on technical problem issues and work progress for adopting suitable measures to ensure adherence to product quality, cost constraints, and project deadline. This essentially means maintaining and sustaining human interactions consistently over extended period, monitoring the functioning of the entire communications system and improving the same for deriving optimal results.
2. A project cost management system for aligning the actual costs of the project along estimated lines throughout the life of the project, restricting cost escalations if any, taking clients into confidence on all cost issues relating to the project, and also ensuring that costs do not impact the project viability by appropriately factoring in clients’ basic cost requirements.
3. An effective human resources system for ensuring that the right people with requisite technical skills get selected and employed on the project, with up to date knowledge on software development for the stated client requirements. The system would also be required to address any employee grievances that could inhibit the project development process as also provide the right motivation to staff for successful project completion.
4. An organizational ethics framework or system for ensuring that all processes and human activities are based on the ideal and highest professional conduct standards and are geared towards protecting the interests of all stakeholders like client-users, developers, top management, partner company, etc.
5. A staff accountability system for ensuring that all employees from the bottom up to the highest positions in the organizational hierarchy are adequately accountable and can be held responsible, so that a quality project as per specific client needs can be developed within deadline, by containing project costs and ensuring optimum client-user satisfaction.
6. A marketing system for ensuring that the IS can be successfully sold to potential clients for mutual benefit of the user as well as the producer and marketer of the system whereby the user/client gets value for his money and the developer/owner gets profits from sales to satiate his basic motive-that of improving his profitability.

Choice of the Relevant System

The project being an IS development project, the need for a functional and effective management information system is essential for providing the AMRIS owners/top management the necessary and timely information on what the lower level employees working on the project are doing at any point of time, on whether the managers who are controlling the project are doing their job well, and also to know if the liaison persons who may be assigned to handle inter-partner and external issues are doing what they have been assigned to achieve so as to minimize negative fall-outs of probable failure. A strong MIS is also essential to control the various aspects of the project so that the higher management can get sufficient feedback from the bottom upwards as to the technical details relating to the project, evolve suitable strategies to resolve any technical obstacles, reorient employee attitudes and skills for the tasks, and also align project specifications, deadline and costs as per need of the clients and owners.

Rich Picture and its Annotation for the Chosen Relevant System

A Rich Picture for the CONFIRM situation was conceived and drawn (Appendix Fig. 1). It was attempted to depict pictorially the problem situation relating to the case. Thus AMRIS, the managing partner being vested with the technicalities and successful completion of the project was represented as responsible to the three other partner companies. A two way communication or exchange of information was envisaged between the managing partner and the other three partners so that each of them could represent their interests to the others and a successful collaboration could be attempted. As the managing partner, AMRIS could also steer the partners in a specific direction in case of dispute among the four partners. The project team was responsible to the managing partner bosses. They could communicate through an intermediary managerial staff only who in turn could link up with higher management. While AMRIS top bosses were driven by profit motive, the project team, in an ideal situation, would be driven by the need for developing an IS acceptable to the three partners based upon the specifications gathered from the consultant team who were stationed at the project site by the partner companies/clients. The risk management team does not appear to have been constituted by the partners of the consortium, if accounts gathered from secondary sources are to be believed. A vital part of any organizational changeover, the risk management team could exercise restraint on the top management in case the company bosses tried to follow profit interests even when these were conflicting with client interests. The risk management team could gather the required data from the project team itself, who would be fully responsible to their managers in the process. The ethical considerations would flow from the top down from the top bosses down to the risk and project teams. Wherever there may be chances of conflicting interests, there may be required some intense negotiations, which again means that the entire communications and information sharing infrastructure needs to be reinforced. Based on the simple rich picture representation, the Rich Picture Annotation and Summary were derived and provided in Appendix (Fig 1).

The Root Definition and CATWOE Analysis: Some Observations

The criteria usually followed for forming an exact root definition is also followed for defining the root definition of the chosen relevant system. This is as follows:

“DO X by Y in order to achieve Z”

Next the elements of CATWOE are also identified for the system as follows:

• C=Customers of the system i.e., the three partners of the IS viz Marriott, Hilton Hotels and Budget Rent-a-car
• A=Actors of the system, I.e., the top management of the three partners and AMRIS, the Project Team, the Risk Management Team, the intermediate Managers, and the Consultant Team of each partner company stationed at the project site.
• T=Transformation, i.e., Lack of proper and timely communication “transformed into” an effective and functional two-way communications and information sharing mechanism.
• W= ‘Weltanschauung’, or the world view shared by all partners of Intrico that the resulting IS CONIRM be the most advanced, comprehensive and single reservation system capable of handling reservations for three diverse industries-airlines, hotel lodging and car rental. Also, all companies were driven by profit maximization, need for facing market competition, and keeping costs constant in the process.
• O= Ownership, or in this case, obviously AMRIS, the company developing the IS for the other three partners.
• E= Environmental constraints, which, in this case is the lack of information sharing among inter-actors, lack of technical knowledge of all the stakeholders, the predetermined deadlines and cost restricted by estimated budgets, etc.

Next the Root Definition as derived from the elements of CATWOE for the chosen relevant system is as follows:

“A system owned by AMRIS to effect optimum information sharing and two way communications mechanism among immediate stakeholders by streamlining organizational information exchange processes and strengthening internal controls in order to achieve relevant and timely feedback for managerial decision making”.

In other words, comparing the statement “Do X by Y in order to achieve Z”, it follows:

• X= Affecting optimum information sharing and two way communications mechanism among immediate stakeholders
• Y= Streamlining organizational information exchange processes and strengthening internal controls
• Z= relevant and timely feedback for management decision making

The Conceptual Model: Deriving and Analyzing

The conceptual model for the chosen relevant system could be derived from listing the front line activities to be performed for achieving the stated objectives of the transformation. This simply meant finding out the key action words or verbs and then depicting these by means of diagrams, albeit in an informal way only since the concept of SSM stresses on unstructured and soft methods of enquiry rather than on formal or hard approach. The key verbs as obtained from the root definition were identified as follows:

1. Effecting (optimum) information sharing
2. Effecting (two way) communications (among immediate stakeholders)
3. Streamlining (organizational) information exchange processes
4. Strengthening internal controls

The conceptual model that then emerged is as provided in Appendix (Fig 4)

Comparing the Conceptual Model with the Real World

A comparison chart of the conceptual model for a system of two-way communications and information sharing is presented in Fig 5 in the Appendix. These form part of the key points for discussion in the management agenda.

Some Possible Agenda Items & Recommendations

The proposals to the steering committee of the consortium would need to include a proposal for transforming the communications and information sharing system in the AMRIS company which is the managing partner, into a more functional and effective one. In view of the issues in the case, the following would form the key points for discussion:

1. AMRIS have to take full responsibility and initiative as the managing partners and follow ethical and equitable practices for successfully developing and implementing the CONFIRM system. They need to do this by complying with their promised deadline for the project as also by adhering to estimated costs.
2. The consultants or technical experts stationed at the project site need to be more effective. They may need more skills development, or they may also need to be replaced by better and more effective overseers who can successfully provide their expertise to the development of the project as also convey their independent and timely feedback to respective company managements.
3. The consultants also need to highlight perceived shortcomings, so that suitable improvements may be incorporated in the project.
4. AMRIS also needs to build up its internal control mechanism as also reinforce its information sharing mechanism so that it can take timely and considered corrective actions.
5. The feedback from its consultants and its project team members could prove crucial in the design and development stages. Hence, the selection of the employees and consultants must be realistic and more skills based.
6. AMRIS should also not misinform its partners on the status of the project, and follow ethical standards even as it follows its profit maximization objectives.
7. AMRIS should set realistic timeframes to its team members and increase their incentives and develop their skills in order to achieve quality of design.
8. Right communications form the top management need to percolate down to the lowest employee and all should work as an effective team with each member of the team having his own role to play.
9. AMRIS should use the carrot and stick method to goad its employees and managers to perform so that better performing employees are not discriminated against while deficient workers are appropriately dealt with. Essentially, the management must do well and appear to do good by its sincere workers and only a sound communications channel can bridge the differences among the top and lower rungs of employees in the techno-major so that its health is fine and the project is completed to the satisfaction of all partners and their clients.

Conclusion

The SSM is one of the best and advanced systems thinking approaches for analyzing problems in the workplace where the problem issues cannot be defined. The method adopts a soft or unstructured method of study of the systems and subsystems present within organizations. In the real world, the organizational activities are too complex and inter-related as to prevent a satisfactory study by using the traditional hard approaches of systems thinking. Also, organizational problems are actually the result of human activity systems (HAS). An internal participant or stakeholder of the system best applies the method. The perspectives of the various stakeholders may vary and hence as many different systems, root definitions and conceptual models may be envisaged for a particular problem issue. The use of rich pictures makes the core issues seem pictorially clear to the analyzer and also the method is in keeping with the original methodology as outlined by Checkland in the 1960’s. In a case like the CONFIRM case, the application of the SSM is all the more relevant since an unbiased, undefined and unstructured analytical methodology is followed in arriving at the possible reasons for the problem and the proposed solutions of the same. While this particular analysis has had to rely on some secondary sources of information only, and also since Partner Company sources are not too forthcoming on what actually happened between them, only an approximate analysis by conceiving a conceptual model and comparing the same with the real world situation has been attempted.

References

Checkland P., 1972, “Towards a systems-based methodology for real-world problem solving”, Journal of Systems Engineering, 3(2): pp. 87-116

Checkland, P., and Scholes, J., 1990 “Soft Systems Methodology in Action”, Wiley, NY

Gladden, G.R., 1982, “Stop the life cycle, I want to get off”, Software Engineering Notes, 7, 2, pp. 35-39

Lyytinen, K., 1987, “Different Perspectives on Information Systems: Problems and Solutions”, ACM Computing Surveys, Vol.19, No.1, pp.5-46

PC Week 16, 1995, p. 68

Appendix

Fig 2: Rich Picture Annotation

Fig.5: Comparing the Conceptual Model and the Real World