Computer Reservation Systems

Lorenzo was by no means through acquiring airlines. In 1985, just prior to Continental’s com­ing out of reorganization, Lorenzo made a play for TWA. Some said this was because he needed its computer reservation system to manage the traffic in his growing conglomerate of airline companies. Nobody had yet heard of the Internet. Computer reservation systems were proprietary with each of the Big Four, and it was realized that these systems gave huge advantages to those airlines by increasing their passenger market share, to the detriment of the smaller lines.

Travel reservations, the process of match­ing an available seat with a named passenger to occupy it, had always been a complex undertak­ing. Even with the railroads, where it was largely a matter of recognizing where passengers on the line of road were getting on and getting off, keeping up with the availability of seats was a daunting task. In the early airlines, as with the railroads, reservations were tracked manually, usually at a central location. Entries represent­ing reservations were made in pencil so that they could be erased if the reservation was cancelled.

When traffic picked up in the 1930s, ledgers became even more impractical, and the system was expanded to chalkboard displays in large rooms, also at a central location, on which entries and cancellations were noted. Clerks who took the reservation request from passengers handed off the information to runners who relayed the details to writers at the chalkboards. In time chalkboards were replaced by electric light displays, also in large rooms, and despite the advanced technology of electricity, the process was still manual, cumbersome, and inaccurate. Increased service to multiple cities in random directions, even on one airline, exponentially increased the difficulties of keeping track of res­ervations manually. Booking seats on multiple airlines made the job even harder.

By the 1940s, efforts were being made to automate the process. Makers of computational equipment, like adding machines, were the logi­cal choice to assist in solving these mathemati­cal complexities, but in turn these companies advised that they could not handle the number of variables presented in the problem.

C. R. Smith of American Airlines, himself an accountant and numbers man, was preoc­cupied with the reservations dilemma. Unable to secure assistance outside of the company, he authorized American’s technical people to come up with a solution. The result was a mas­sive mechanical monstrosity consisting of verti­cal cylinders, each one representing a different flight on a given day, which was filled with marbles representing available seats. When a seat was booked, an agent activated a switch that released one marble from the cylinder. A reciprocal arrangement at the top of the cylinder released a marble back into the cylinder for each reservation that was cancelled.

This arrangement was an improvement, but it was no match for the growing problem of reservations as traffic increased. With the beginning of the jet age in commercial air traffic, once again the problem was made exponentially more difficult. The process was not only marginally inaccurate, but also very costly for the company as personnel and terminals had to be added to the system.

IBM, through its primitive computer technol­ogy, during the 1950s was out front in developing solutions for the federal government related to the problems of monitoring the potential for incoming ballistic missiles. The acronym for the IBM pro­gram was SAGE, Semi-Automatic Ground Envi­ronment. SAGE was the first computer game in real time as strategic and tactical planners engaged each other in simulations of nuclear warfare.

Under contract with American, IBM began applying its SAGE technology to the reserva­tions problem, and for almost 10 years its best minds labored away. The project was originally known as SABER, Semi-Automatic Business Environment Research, and later as SABRE, and the first commercial activation of the system did not occur until 1962. At that time computer technology was truly rudimentary, and almost all commercial computers were engaged in solv­ing mathematical equations, or in streamlining the problems of accounting in corporate Amer­ica, like payrolls. And these applications were applied to dealing with numbers in a historical context, not real-time. With SABRE, real-time computing in business applications was born.

With this development American Airlines had a real commercial advantage over its compet­itors. In the 1960s, the Civil Aeronautics Board was still in control of all meaningful decisions related to the running of an airline, so SABRE’s function was expanded to not only solve Ameri­can’s reservations problems, and to assure con­sistent and accurate reservations for the very first time, but also to track every passenger’s name, address, personal information, and most details of that passenger’s travel preferences, such as hotel usage. Not only could SABRE track cus­tomer information, but the technology was soon expanded to begin to solve the company’s day – to-day operational problems. But management at American was slow to realize the full potential of the advantage given them by the computer sys­tem they had developed.

The other airlines began their own experi­mentations with computers, particularly as applied to the reservations system. The technol­ogy was still relatively primitive, and the cost was enormous. In 1966, TWA committed $75 million to solving the problem, hiring Burroughs Corporation to come up with a proprietary com­puter reservations system (CRS). By 1970, no workable system had been achieved, although in time TWA would perfect a system known as PARS. United began its own program, called APOLLO, and made reasonable progress. At the same time at American, SABRE was losing its advantage as management failed to upgrade equipment, and as uninstalled computers sat in storage, allowing its competitors to catch up. Still, all computer reservation systems at the time were considered works in progress.

In 1970, in spite of their individual efforts up to that time, the major airlines realized that, from a cost-effectiveness standpoint, it made a lot more sense to pool their resources to develop the ultimate computer reservations system than for each to go it alone, thereby duplicating effort and wasting untold sums of money. When pre­sented with the airlines’ plan, the Justice Depart­ment announced that it would consider such a combination between the major carriers to be a violation of the Sherman Antitrust Act, and would prosecute the airlines criminally if they proceeded. Many considered this an extreme example of government abuse of power, but there was little the airlines could do. The opportunity was thus lost to have a single, unbiased reserva­tions program developed for the benefit of all of the airlines and the public at large. The only course left for the airlines was for each of them to develop their own, proprietary system. Few in 1970 realized the commercial potential of the computer, or the great benefits that would inure to the owners of these proprietary systems. The joint plan proposed by the airlines would have allowed the unbiased computer reservations sys­tem to be used by all travel agents in servicing the flying public. Now the public would have to wait, as would the travel agents.

Around 1975, the travel agents got together to announce that they were planning to develop their own CRS. United had its APOLLO up and running, and by 1974 it was generally considered to be the best in the industry, having surpassed SABRE. It was, however, still in development. No one at the major airlines believed that it was in their interest to lose control of CRS, and be faced with a giant travel agent computer network where all flights of all airlines would be available to all travel agents everywhere. It was feared that such a system would require the airlines to pay a transaction fee for every reservation, in addition to the commission that they paid.

Another effort was made by the airlines to convince the government of the desirability of the joint approach. The CAB this time gave the airlines antitrust immunity, but only to permit the airlines to explore the possibilities of such a system—to talk, but not to proceed, with build­ing the program. It was at this stage, in July 1975, that United unilaterally declared that it would no longer participate in seeking government approval for the joint effort, and that it would go it alone. United, as the biggest bear in the woods, believed that it had a competitive advantage over the other airlines in its CRS, and it began to appreciate what favorable nuances could be incorporated into the program to heighten that advantage. United’s plan was to gain control of the travel agent business by supplying travel agents with its APOLLO program which would, of course, have built into it biases in favor of United.

The world of travel agents at the time was one of telephones and paper transactions. The Official Airline Guide (OAG) was a periodical publication containing all the world’s airline departures and arrivals, displayed in a city pair format. The procedure was for the travel agent, upon receiving a request from a traveler for flight information preparatory to booking a reservation, to go to the OAG, discern the flight information and the airline that most closely matched the traveler’s request, and then secure authority from the traveler to book the flight. The travel agent would then telephone the airline, confirm the res­ervation, secure the airline’s authority, and then telephone the traveler back with the confirmation. The agent would then write the ticket and ulti­mately transmit it to the traveler, usually by mail. The travel agent was paid a commission by the airline.

The United plan would simplify this proce­dure greatly. The plan was to install computer terminals in the travel agents’ offices for a fee, and then provide the agents with all of the flight information available in the OAG on an inter­active, real-time basis so that the travel agent would be able to confirm the reservation while the traveler was still on the phone, then the com­puter would issue the ticket. Unstated, but appre­ciated by some of United’s competitors like Bob Crandall at American, was the fact that APOLLO would contain preferences for United through outright biased presentations that would likely cause the travel agent to favor a United flight over any other.

Typical of the types of bias that the com­puter could generate was the positioning of the flight information on the computer screen. Amer­ican had conducted research that showed that 50 percent of the time, travel agents selected the flight that appeared on the first line of the computer screen. Ninety percent of the time, the travel agent picked a flight that appeared on the first page of a multi-page computer display. If the proprietary CRS program were configured to offer its own flights on the first line, or at least in a favorable position on the first page, there was an advantage to that airline. Various algorithms were developed to accomplish these ends.

Dick Ferris of United and Bob Crandall of American, with their companies in a nip and tuck race to lead the industry in the middle 1970s, were head-to-head competitors. Crandall resolved to bring SABRE back up to a com­petitive level, and to pitch SABRE to the travel agents as the best system for them. Crandall did his homework, made presentations at national travel agent conventions, conducted mail-out campaigns, and before long, American was out in front again.

The agents who signed up with American were provided with terminals, computers, moni­tors, and the essentials for using the system in their business, and they were charged a fee. Only the largest “commercial” agencies could afford to par­ticipate, but the hardware was getting cheaper by
the month. Then United struck back by providing some of the agencies with the equipment without a fee, and allegedly gave rebates (kickbacks) to the agencies for using United’s CRS.

By 1983, the proprietary computer reserva­tion systems included American’s SABRE, Unit­ed’s APOLLO, TWA’s PARS, Delta’s DATAS II, and Eastern’s SODA. All of these systems began as in-house reservation systems, but their databases were expanded and their access sys­tems were configured to allow distribution to travel agents under either lease or outright sale. The airlines’ mainframe computers were oper­ated by the airlines, telecommunications equip­ment connected the airlines’ computers with the travel agents, and the travel agents equipped their offices with computer terminals and printers.

Eastern was losing money in 1983, so much so that insolvency appeared to Frank Borman, the first American astronaut to orbit the earth and Eastern’s CEO, to be a distinct possibility. The rigors of deregulation, along with the serious eco­nomic situation that existed in the early 1980s, were taking a toll. Borman, like others in the industry, went to the rank and file with pleas for help in the form of “givebacks,” or voluntary wage cuts, in order to meet the emergency. Reluctantly, the pilots and the flight attendants cooperated, but the machinists did not. In fact, they demanded and got a 32 percent wage increase on threats of a strike, which created a rather incongruous situation among the respective crafts. The pilots were not happy, nor were the flight attendants, and morale plummeted. And Eastern continued to lose money.

In 1985, Eastern’s debt approached $2.5 billion and income was dwindling. But East­ern had a computer reservations system. Texas Air was still flush with cash but Lorenzo still did not have his own CRS. Lorenzo offered to supply the needed cash to Eastern through a straight buyout. Because of Lorenzo’s reputa­tion, neither management nor the employees favored this idea. Borman desperately sought ways to right the ship, appealing to the work­ing crafts for even more concessions, but it was clear that without the support of Charlie Bryan and the machinists, there was little hope. With all options exhausted, Borman and the Eastern board of directors, at a midnight meeting, reluctantly agreed to the sale to Texas Air. While awaiting government approval for the Eastern purchase, Lorenzo turned his attention to People Express.