Truth, Justice and the American Way -What happened to CONFIRM?
by Alan Sadowsky
With the exception of a tenacious legal staff seemingly preparing for what may be one of the landmark legal actions of the century, the dust has all but settled in Carrollton, Texas, former site of the CONFIRM RS development effort. Touted as the leading edge hotel and car rental reservations system by the International Reservations and Information Consortium (Intrico), CONFIRM was suspended in July, after more than four years of work and over $125 million in expenditures.
While there are still questions surrounding specific issues and actions relevant to the overall project, the main question being asked is, What really happened? What the data processing community is primarily interested in, are the technical matters directly attributed to the demise of the project. If in fact there are valid deficiencies in the products chosen for the backbone of CONFIRM, there is certainly a desire on the part of current and potential users of these products to be informed about them. This article will attempt to provide some of the answers to the questions, and at the same time clear up a few of the many misconceptions about what really took place in Carrollton, Texas.
To begin with, Confirm was to be the reservation system of the future. While there are any number of capable reservation systems in existence today for the airline, car rental, and hotel industries, Confirm was to be the first to integrate all three disciplines into a state-of-the-art complex providing functionality never before available to it's customers. Participation in Confirm involved the partnership of four major players in the travel field: AMR Information Services (AMRIS), Budget Rent-A-Car, Hilton Hotels, and Marriott Hotels. It was the responsibility of AMRIS to develop and operate (DMN)the Confirm system for the partnership.
With a technical staff headcount of more than 200(DMN) programmers, systems analysts and engineers and an additional 200+(IW) support and administrative people on the Confirm project, one would think that sufficient resources had been gathered. In terms of skills and abilities, that was in fact the case. The main pieces of the project were clearly defined in 47,000 pages of documentation,(DMN) and a physical configuration was established utilizing two IBM 3090 processors: one running TPF(CW) for reservations processing and one running MVS-based DB2(CW) for decision support. By all accounts, everything should have proceeded along as planned. So what went wrong?
The many newspaper and trade journal articles written about the Confirm demise have all maintained a common pretext for the technological problems. Focusing on the multi-platform configuration, the problem being emphasized is the integration and communication between the two different operating system platforms, TPF and DB2. In addition, rumors about the use of IBM's C/370 language and the use of IBM's TPFDF feature having caused major problems on the project began to surface throughout the industry. Let's address these issues before going any further.
Were there technical problems integrating the multi-platform configuration? Certainly, but it should be noted that TPF and DB2 were communicating with each other, albeit inefficiently. Moreover, with the announced availability of IBM's TPFDF Distributed Data Access Feature, and the TPF Application Requester, the magnitude of these problems might have been significantly reduced or even eliminated. The rumors regarding C/370 and TPFDF are nothing more than rumors. The problems experienced with these features were not specific to the products, but rather with the way the products were used. AMRIS Travel Services did have the distinct advantage of having an excellent group of technicians on the project, with a broad depth of C/370 and TPFDF skills. C/370 and TPFDF have both proven to be valuable tools in a TPF development context, and honestly had no role in the suspension of the Confirm project. Without the use of C/370 and TPFDF, it is conceivable that the degree of accomplishment on the TPF side of the project would have been significantly diminished.
Before we get completely mired in the myriad of technological problems and concerns attributed to the Confirm project, I want to make it very clear that this technical argument of rationalization is unfortunately only the effect of a more disturbing cause. Yes, there were technical problems. Yes, portions of the application were not performant. Yes, independently developed pieces of the "puzzle" did not fit together very well. However these shortcomings were not the root of the problem at Confirm. Sadly, the problem at Confirm was management.
When looking at business, there are five basic elements that are essential for the success of that business:
1. A quality or unique product -- There isn't any argument that Confirm was a unique product. When you consider the degree of functionality planned and the genuine need for a fully integrated reservations system, this first element for success meets the criteria.
2. Proper timing -- One more item in the plus column. The marketing analysis and the numerous studies conducted by the Confirm partnership all indicated that the timing was perfect for the introduction of the system. In fact, initial estimates on achievable worldwide market-share were near the 40% mark.
3. Adequate capital -- Money while always an issue was not an obstacle. The financial positions of AMR, Hilton, Budget, and Marriott were all healthy at the inception of the undertaking, and $125 million was jointly earmarked for the project. So far we're three-for-three!
4. People resources -- Whether you want to consider sheer headcount numbers or skills and abilities, Confirm had both bases well covered. From a numeric as well as technical perspective, the project boasted well over 600 employees and consultants with an enviable range of talents. Four down; one to go.
5. Effective management -- The influence that effective management has on the first four prerequisites is unyielding. Without effective management, the first four elements are lost. Quoting from the book 13 Fatal Errors Managers Make And How to Avoid Them by W. Steven Brown: "Without effective management, correct decisions cannot be made about the product's features and the proper time for its introduction into the marketplace. The company lacking proper management cannot acquire, much less sustain adequate capital. Above all, it takes good management to attract the best people and to coach and develop them. Every forward-looking manager recognizes that the greatest untapped resource within any company is its people's potential."
Originally scheduled for implementation in 1991, the target date for bringing Confirm online slipped into 1992. Intrico had announced that the first phase of the implementation would occur June 28. User acceptance testing by Hilton Hotels began in January of this year, and by April it was obvious that there were serious problems with the system.
According to a recent article in Hotel Business News magazine, insiders reported that AMR sent in a "SWAT team" of experts from their SABRE Computer Services division (SCS), to evaluate the Confirm situation. They uncovered serious flaws in the system's design. This was corroborated by a statement issued by AMR spokesman Al Becker, saying that AMR "believed that the project was on schedule and on track through the end of March." However, after problems came to light in April, "AMR launched a full-scale inquiry to investigate all areas of the system. A team of SABRE Computer Services experts was brought in to better evaluate the situation and assist in initiating corrective actions."
The SCS assessment determined that the problems uncovered would require 18 months to correct, delaying the project well into 1993. "For management to suddenly learn -- years into the project -- that they're still 18 months off, can only mean one thing: They didn't know where they were in the project" in the first place! "Somewhere in there, you've got a management problem(DMN)" said Donald Tatzin, director of Arthur D. Little's travel consulting practice.
Allegations of project status being misrepresented, of timelines being falsified, and of work not being completed rose quickly out of the confusion. Quoting one computer consultant close to the project, "This was more vaporware than anything else. People claimed there was progress on many things, and there was none."(DMN) AMRIS promptly fired eight members of AMR Travel Services management, including division president John Mott, saying it had "determined that information about the true status of the project appears to have been suppressed by certain management personnel."(DMN)
In spite of SCS efforts to attack the Confirm problems and salvage the project, both Hilton and Budget announced in late June that they were dropping out of the partnership and abandoning the project. By the end of July, Marriott made it unanimous, and Confirm ceased all development.
Putting the rumors aside, ignoring the allegations, and looking beyond what may or may not have been intelligent technical decisions, who is really accountable for the failure of Confirm? W. Steven Brown states: "When Harry Truman was president of the United States, he had a sign in the Oval Office: THE BUCK STOPS HERE. Each manager should adopt the same dictum. The buck must stop with management. Essentially there are two actions in life: performance and excuses. Two distinct and entirely attitudinal approaches exist based on these actions, and only one manages successfully. Internalists, those who are performance oriented, accept personal accountability for their actions, successes, and failures. They know that if they feel unhappy with their results, they have only to look in the mirror to stare the culprit right in the eye. Others refuse to accept their responsibility for their position in life and hide behind excuses. Because they constantly blame some external source, condition, or other people for their personal failures, we'll call them externalists. We'd rather not call them managers."
Does accountability lie solely with AMR Travel Services management, or does it climb higher up the corporate ladder? Just where does the buck really stop? The technical failures of Confirm are the outcome of management failures. It is difficult to understand how AMR senior management, pacesetters in travel technology, did not know what was going on at Travel Services. While I'm not in any way implying that AMR senior management did know what was transpiring, it is certainly inexcusable that they didn't. That's where the buck stops!
What was to be the reservation system of the century may become nothing more than the lawsuit of the century. Talk of legal action by the former partners is a threatening possibility. Suspension of Confirm has already translated into a $165 million pre-tax charge against AMR's earnings. Any additional expenses incurred by AMR (legal or otherwise) will only add to the magnitude of Confirm's demise. In conclusion, if asked to choose a single statement to wrap up the Confirm saga I would have to borrow from Aldous Huxley. "Facts do not cease to exist because they are ignored." Neither does responsibility.