The Evolution of Yield Management in the Airline Industry: Origins to the Last Frontier

By Ben Vinod

This book chronicles airline revenue management from its early origins to the last frontier. Since its inception revenue management has now become an integral part of the airline business process for competitive advantage. The field has progressed from inventory control of the base fare, to managing bundles of base fare and air ancillaries, to the precise inventory control at the individual seat level. The author provides an end-to-end view of pricing and revenue management in the airline industry covering airline pricing, advances in revenue management, availability, and air shopping, offer management and product distribution, agency revenue management, impact of revenue management across airline planning and operations, and emerging technologies is travel. The target audience of this book is practitioners who want to understand the basics and have an end-to-end view of revenue management.

• Language: English
• Publisher: Springer
• Release date: May 28, 2021
• ISBN: 9783030704247

Book preview

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021

B. VinodThe Evolution of Yield Management in the Airline IndustryManagement for Professionalshttps://doi.org/10.1007/978-3-030-70424-7_1

1. Origins

Ben Vinod¹  

(1)

Charter and Go, Grapevine, TX, USA

1.1 Introduction

In the early days of commercial aviation, private contractors had to bid for the routes to carry mail after the passage of the Contract Air Mail Act on February 2, 1925, more commonly known as the Kelly Act after Representative Clyde Kelly of Pennsylvania. For his work to revolutionize the delivery of mail service in the United States, he earned the title father of the air mail. The Kelly Act was an important piece of legislation to free the airmail from total control by the U.S. Post Office. The Kelly Act allowed the Postmaster General to negotiate independent contracts with private companies to carry mail. The airmail planes had one seat available for sale to a commercial passenger. Passengers would contact the airmail company in the departure city and make a reservation if a seat were available on the requested departure date.

When airlines started carrying more than a single passenger on a flight, automation was required to determine availability, accept the reservation, and update the inventory counts. During the 1930s most airlines operated a request and reply system (Copeland, 1995). Seat inventory was still controlled at the flight’s departure city, which resulted in a roundabout process of the airline agent communicating to the agent in charge of inventory from a departure city for changes to the reservation on behalf of the customer. After the sale was confirmed, the agent would record the details from the passenger on a PNR card before it was transmitted to inventory by teletype or phone.

In 1939, the request and reply system was replaced by a sell and report system at American Airlines. Agents in the Boston office realized that agents could sell seats freely until the flight was almost sold out, thereby reducing the number of phone calls, and speeding up the process of processing customer requests, boosting agent productivity. Once the bookings on the flight had reached a threshold, a ‘stop sale’ message was transmitted to the agents, and the process reverted to the request and reply environment. Although an improvement, this still suffered from inefficiencies and a lack of automation.

The Civil Aeronautics Authority was formed by the United States government in 1938, which was renamed as the Civil Aeronautics Board (CAB) in 1940. CAB was given regulatory power over civil aviation including airline tariffs, airmail rates, mergers, and competition. CAB virtually controlled the performance of the U.S. airlines until the late 1970s by regulating prices for each route and governing which carrier could operate on which route.

1.2 Origins of the Airline Reservations System

The inventory control domain of the airline reservations system, also known as the host computer reservations system (host CRS) serves as the execution component of a revenue management system to accept and reject bookings. It is the place that inventory controls recommended by revenue management are executed to accept and reject booking requests to maximize an airline’s network revenues. The core components of a host CRS include flight schedules, inventory, passenger name record (PNR), ticketing, and departure control. 

Even with the growth of commercial aviation after World War II, the reservation process was antiquated. The sell and report system operated by airlines was inefficient. To avoid oversales, airlines had to maintain seats in reserve so that the last few seats could be processed by the slower request and reply system. The age of interactive real time computing technology had not yet arrived.

In the 1940s, the President of American Airlines, C.R. Smith, was looking for a breakthrough in reservations processing with the aid of electromechanical devices. However, at that time no manufacturer of business equipment could fulfill American’s requirements for reservations processing: find flight availability, record the booking or cancellation, retain processed transactions on the agent’s device until it was manually cleared and automatically notify other stations on the status of the flight. Besides, the data processing had to be economical over the current manual environment. The first electromagnetic system designed to determine seat availability was the Boston Reservisor in February 1946 to replace the card files. Because of the time it took to reconcile the passenger’s name record against seat inventory, it required a buffer of a few seats to avoid denied boardings. In 1952, the Magnetronic Reservisor was installed at LaGuardia airport in New York which could handle 1000 flights for 10 days. With this system, many operators could look up information simultaneously and notify ticket agents over the telephone whether a seat was available. It was still considered inefficient because an operator and agent had to communicate at each end of the phone line. In addition, the issue of reconciling the passenger name record with seat inventory after each booking continued to persist.

The chance encounter between American Airlines CEO C.R. Smith and a young IBM International Business Machines Corporation (IBM) salesman R. Blair Smith on a Los Angeles to New York flight in 1953 revolutionized airline reservations processing (Copeland, 1995). Blair Smith was on his way to an IBM training session in New York.

Blair Smith¹ later recalled the conversation with C.R. Smith: I told [C.R. Smith] I was going back to study a computer that had the possibility of doing more than just keeping availability. It could even keep a record of the passenger’s name, the passenger’s itinerary, and, if you like, his phone number. Mr. C.R. Smith was intrigued by this. He took out a card and wrote a special phone number on the back. He said, ‘Now, Blair, … when you get through with your school, our reservation center is at LaGuardia Airport. You go out there and look it over. Then you write me a letter and tell me what we ought to do.’

At the training session Blair Smith informed IBM’s CEO Thomas J. Watson Jr. about his conversation with C.R. Smith. Watson urged Smith to do what C.R. Smith had requested: tour the reservation center, write a letter with his recommendation, and send a copy to Watson. Blair Smith recommended a joint project between IBM and American to create a computerized reservations system. The research project was called SABER—Semi Automatic Business Environment Research, to study the technical feasibility for automating the reservations process by integrating a passenger’s name to a seat reservation. When this project was completed in 1958, American signed a new contract with IBM to create the functional specifications for the world’s first passenger name record (PNR) system to match passengers to seats.

In 1951, IBM on behalf of the United States Air Force set up project SAGE (Semi-Automatic Ground Environment) to develop a real time computer system for air traffic control. In 1959, the federal government declassified SAGE. The air defense system, partially developed by IBM, had the promise of interactive real time computing. SAGE was a major invention and made several technical advances (Astrahan & Jacobs, 1983; Copeland & McKenney, 1988)—magnetic core memory, active standby dual processors, modems for digital communications over voice-band channels, time sharing the central processor, input-output control with memory cycle stealing and branch and index instructions. The IBM SAGE computer, the AN/FSQ-7 also had a communication front-end that received data in real time from tracking devices over communications lines. The knowledge gained from SAGE was invaluable to develop the world’s first computerized airline reservations system.

IBM viewed the Sabre partnership as a high-risk undertaking and hired the services of the consulting firm Arthur D. Little to evaluate the feasibility of the endeavor. At that time IBM only sold or leased hardware and relied on customers to develop the software. To avoid failure, IBM decided to make an investment to support program development. From American, Roger Burkhardt and Wilfred (Fred) Plugge were the driving force for this initiative. To augment the programming staff, American hired mathematician Mal Perry who played a central role in the development of the reservations system. Phased deployment was planned to begin in 1961. The joint IBM/American team moved into the American Airlines headquarters at 99 Park Avenue in Manhattan, NY. Bill Elmore programmed the first PNR demonstration on an IBM 650, which at 25,000 instructions was an early indication of how powerful the Sabre system could be. There was a software bug in the program that caused the 650’s drum memory to zero itself out unexpectedly, leading to more skepticism (Head, 2002). Losing sight that application programming in assembly language was the primary challenge, IBM made mistakes by proposing incompatible hardware in their proposals to Pan Am, American and Delta. American’s system used the binary IBM 7090 computers. Delta’s DELTAMATIC used the IBM 7070 and Pan Am’s PANAMAC used the IBM 7080 computers, which were decimal machines (Copeland & McKenney, 1988). These computers had different throughput capabilities and processors (Siwiec, 1977). In retrospect, to lower development costs by spreading the cost across the three airlines, they should have standardized on the more expensive 7090 computers. By 1960 Pan American and Delta had signed contracts with IBM and the umbrella project for all three carriers was called SABER. This prompted American to seek a new moniker for its system, and it was called SABRE (Semi-Automated Business Research Environment).

In 1962, the Sabre central site for the reservations system was in Braircliff Manor. Robert V. Head, who joined the Sabre team in October 1959, had an interesting story to tell. During a visit to Bill Elmore’s office, which overlooked the exercise yard of Sing Sing Prison, Elmore explained to Head the difference between the prisoners and Sabre programmers—the prisoners knew when they were going to get out! After his tenure on the Sabre project, Robert V. Head went on to become a prolific writer and contributed articles to banking, automation, and computer journals. He also wrote several books on information systems and management.

In 1964, after several cost overruns and delays, SABRE was launched. It was the first fully operational computerized reservations system. DELTAMATIC and PANAMAC became operational in 1965 followed by Eastern’s System One. Eastern’s System One was an important milestone since it was based on IBM’s Programmed Airline Reservations System (PARS), which became the new baseline for reservations processing in the early 1970s. TWA and United were allied with hardware vendors Burroughs and Univac, respectively. Neither Burroughs nor Univac had any experience with teleprocessing systems (unlike IBM’s work on SAGE) and these two initiatives ran into a range of technical problems. Forced to seek an alternate solution, both carriers decided independently to purchase Eastern’s software with installation support from IBM (Copeland & McKenney, 1988).

The SABRE system had two binary IBM 7090 mainframe computers in a new data center located in Briarcliff Manor, New York. With the new computing environment, American could economically check availability, update inventory, and record a passenger name record (PNR) automatically. American’s reservation error rate dropped to under 1%, and a reservation could be processed in seconds. The new system could process 7500 reservations per hour. The average time to process a reservation in the manual card system was 90 min, which was now done in a few seconds.

It was the world’s first commercial online transaction processing (OLTP) system. This fulfilled American’s requirement to economically check availability, update inventory, and record a PNR automatically. When the first fully operational computerized reservations system was launched in 1964, Sabre was the largest private real time commercial data processing system in the world, second only to the U.S. government. Agents accessed the mainframe reservations system with Raytheon cathode ray tube terminals (CRT) that were completely dependent on the host. The industry called these dumb terminals because they had extremely limited capabilities on their own.

A few years later, IBM released a generalized version of its reservations system called PARS for Eastern Airlines on the powerful System/360 computers. By 1971, PARS became the industry standard with the application software for a passenger name reservation and seat inventory. United introduced the Apollo system in 1971 based on PARS. The operating system for the reservations systems was called the Airline Control Program (ACP). ACP was a low-level programming language in Assembly and code was written in 128-byte, 381-byte, 1 K or 4 K blocks. If the code could not be contained in a block, the blocks were chained for continuity. Remarkably, conceived in the 1960s, ACP has survived conceptually unchanged into the twenty-first century. ACP eventually became Transaction Processing Facility (TPF).

Sabre went through a series of upgrades to the PARS-based system from 1971 to 1973 which culminated with the replacement of the IBM Selectric terminals with cathode ray tubes in 1973. After Sabre upgraded to PARS based system in 1971, nine out of the top ten U.S. major airlines were on PARS (IPARS is the international version). These transaction-oriented systems were developed in TPF, a low-level assembly language that could process a high volume of messages per second. They supported all reservations related functions—schedules, inventory, passenger name record, shopping, itinerary pricing, ticketing, and departure control. ACP and later TPF high performance operating systems revolutionized transaction processing on an unprecedented scale, improving efficiency of airline passenger operations and ultimately airline profitability. The generic name for IBM-based reservations systems is Passenger Service System (PSS).

Today, Amadeus’ Altea, Sabre’s SabreSonic Customer Sales and Service and Navitaire New Skies reservations systems host many airlines. Navitaire, a wholly owned subsidiary of Accenture, was acquired by Amadeus in 2015. Other vendors that provide hosting services for reservations are Shares, SITA Horizon Customer Sales and Service, Hitit Computer Services and Radixx International. Radixx was acquired by Sabre in 2019. UniSys Aircore® in a partnership with TravelSky provides hosted reservations services for carriers based in China. The Civil Aviation Administration of China (CAAC) mandates that all Chinese carriers should be hosted for reservations in China. While the industry has been consolidating and using the services of CRS vendors that provide hosted services, a third of the world’s airlines continue to use proprietary systems for reservations processing.

1.3 Airline Deregulation

The Airline Deregulation Act that was signed into law in 1978, despite the unanimous opposition from airline executives, under the administration of President Jimmy Carter. Deregulation resulted in increased service, a wider choice of air services to choose from and increased competition. Airline deregulation dismantled a comprehensive system of government controls (Kahn, 1988a, 1988b). Prior to deregulation, airlines operated in a tightly regulated environment; regulated by governments and self-regulated through organizations such as the Civil Aeronautics Board (CAB) and the International Air Transport Association (IATA). A powerful voice in favor of deregulation of the airline industry, economist Alfred E. Kahn, was the chairman of the CAB. Kahn was the chief architect and promoter of deregulating the airlines in the United States despite opposition from labor unions and industry executives. He is acknowledged worldwide as the Father of Airline Deregulation.

The act called for the complete deregulation of all fares by January 1, 1983, a year after the CAB’s route authority was to be terminated. The deregulation of the airline industry led to the birth of pricing and revenue management as it is called today. For the first time in commercial aviation history, airlines had the liberty of changing their routes and fare structure in response to consumer demand and competitive pressures. With airlines around the world going through various phases of deregulation today, the luxury of investing in new technology and solutions in the pursuit of incremental revenues has become a necessity to survive in a fiercely competitive environment. Every one of the major airlines has been practicing some form of pricing and yield management since deregulation, although with varying degrees of sophistication. The pricing and yield management process is the most important determinant of airline profitability. The practice is mature on several fronts such as business process adoption of best practices, decision support capabilities for sophisticated forecasting and inventory control recommendations and cost-effective product distribution.

Deregulation saw a major transformation of a regulated industry (Crandall, 1995; Dempsey & Gesell, 1997) overnight. During the regulated era, competition was artificially limited, and profitability was all but guaranteed. The CAB approved routes and set fares that guaranteed airlines a 12% return on flights that were 55% full. Deregulation changed the dynamics of airline profitability, enhanced competition and productivity and ushered an era of survival of the fittest. The U.S. Domestic Airline Deregulation Act mandated the end of route restrictions by December 31, 1981 and the end of rate regulation by January 1, 1983. The CAB accelerated this effort and ended route regulation in 1979 and rate regulation in 1980. With deregulation, CAB was sunset on January 1, 1985.

Airline deregulation also had its negative impacts on the population at large. Several smaller cities lost commercial service which in turn resulted in the movement of people toward the larger cities at the expense of the smaller cities. At that time this had an adverse impact on the general health of the U.S. economy.

With deregulation, the government mandated fare levels through the CAB’s Standard Industry Fare Level (SIFL) came to an end (Kretsch, 1995) for the domestic U.S. market. This led to an explosion of fares offered by competing airlines from the traditional full fare to discounted and deeply discounted fares that brought air travel within reach of a whole segment of the population that had never flown before. Between 1976 and 1990 average yields per passenger mile declined 30% after adjusting for inflation (Kahn, 1988a). Besides changing the domestic U.S. market, deregulation also influenced the airline industry worldwide.

The dismantling of government control with deregulation of the U.S. airline industry in 1978 resulted in an exponential increase in fare filings and new markets that were served by airlines. Faced with intense competition after deregulation, several airlines like Braniff Airlines (1982), PEOPLExpress (1986), Eastern Airlines (1991) and Midway Airlines (1991) ceased to operate. There was also a wave of airline consolidations after deregulation.

1.4 Yield Management: The Early Period

Yield management is discussed in the context of scheduled air carriers that operate under a Federal Aviation Regulations (FAR) 121 certificate granted by the Federal Aviation Administration (FAA). Charter operations (FAR 135 certificate and variations) is not addressed in this book.

Prior to deregulation, airlines operated in a tightly regulated environment; regulated by governments and self-regulated through organizations such as the Civil Aeronautics Board (CAB) and the International Air Transport Association (IATA).

An overview of key economic, regulatory and market factors that led to the discipline of yield management is discussed in Cross (1998). In the 1940s CAB had regulatory powers over tariffs, air mail rates and competition. Following United’s lead in 1940, airlines introduced a second class of service (coach class) at a lower fare in the 1940s. Charter passenger airlines (airlines that offered non-scheduled passenger services), offered lower fares than scheduled airlines in 1948. Faced with this competition, schedules airlines like Capital Airlines sought permission from CAB to offer discounted fares. The era of discount travel had begun and led to a range of restricted fares in coach class. Common restrictions were off-peak travel, no stopovers, and no refunds. Discounts expanded into the 1960s with qualified fares such as youth fares, family fares with discounts for children, clergy fares (required an identification card) and Discover America excursion fares. Discount fares continued to flourish in the 1970s with a wider range of discounted fares. The Peanut Fares were introduced by Donald Burr when he was President of Texas International Airlines in a few markets with low load factors with CAB approval. A month later, with CAB approval, American introduced the SuperSAAver ™ fares valid only on roundtrips with a 45% discount over coach fares, a 30-day advance purchase restriction and limited between 7-days to 45-days for the return. The justification provided by American to CAB was the competitive threat from transcontinental charter flights from New York to Los Angeles and New York to San Francisco. The CAB introduced a free zone in 1978 in which fares that ranged between 70% discount over coach fare formula and a 10% increase over the coach fare formula were deemed lawful.

Southern Airways conduced the first planned overbooking/controlled oversales experiment in 1949 (Cross, 1998). In the mid to late 1950s and 1960s, airlines practiced overbooking, the process of accepting more bookings than available capacity to compensate for the effects of cancellations and no-shows. It was reported by Marvin Rothstein (1985), Director of O.R. at American Airlines, that prior to 1961, U.S. airlines did not publicly acknowledge overbooking, even though they practiced it discreetly. In 1961, the CAB acknowledged that one in every 10 passengers on the 12 leading airlines at that time no showed for flights which caused economic distress to the airlines. Even though the CAB did not officially approve the practice of overbooking, it mandated a no-show penalty of 50% of the value of the ticket to passengers and also put in place a reciprocal policy wherein airlines paid a penalty equal to 50% of the value of the ticket for passengers who were left at the gates. These penalties were abandoned in 1963 and a subsequent study was conducted by the CAB in 1965–1966. Their study determined that there were 7.69 denied boardings for every 10,000 passengers boarded (CAB economic Regulation Docket 16563, 1967). The CAB concluded a year later in 1967 that airlines could accommodate more passengers by practicing controlled overbooking. Hence, overbooking became an accepted practice, but the CAB also raised the penalty to 100% of the value of the ticket. The involuntary denied boardings reported as the number of denied boardings for every 10,000 passengers boarded was monitored by the CAB and today is still reported by the U.S. Department of Transportation (DOT).

Prior to airline deregulation, the first discount allocation model was proposed by Kenneth Littlewood (1972) from the British Overseas Airways Corporation (BOAC). Littlewood proposed that airlines maximize revenues instead of passenger occupancy on a flight for the perishable seat inventory. Known as Littlewood’s rule for two booking classes, this can be extended to multiple classes. In 1980, William Swan, a member of the O.R. department at American, made extensions to Littlewood’s model, with the logit approximation to the normal distribution. This was first deployed in 1982 (Smith, Leimkuhler, & Darrow, 1992).

Yield management came into existence on a significant scale a few years after the deregulation of the U.S. airline industry.

The term yield management is attributed to Robert L. Crandall when he was Senior Vice President at American Airlines (Cross, 1995). When Robert Crandall was named President of American Airlines in 1980, airline deregulation was already underway. American, under Crandall, was concerned about relentless price cutting by competitors in an unregulated market that would wipe out profits. When this came to pass, innovation began in the airline industry. Carriers operating in a deregulated environment quickly realized the importance of passenger mix to improve yields and the importance of increasing the seats occupied by revenue passengers because the cost of carrying a marginal passenger on a flight was negligible, and it was almost pure profit.

Crandall succeeded Albert Casey as American’s Chairman, President and CEO in 1985 (Serling, 1985). The first yield management system was created in 1985–1986 by American Airlines under the leadership of Robert Crandall to counter the competitive threat from PEOPLExpress (Crandall, 1995; Cummings, 2007). For American, yield management served as a strategic and tactical weapon to counter the competitive threat. An Operations Research team under Tom Cook, Director of the Operations Research group (and later President of American Airlines Decision Technologies) developed the system which was deployed in 1986 (Cook, 1999; Smith et al., 1992). This system, called DINAMO (Dynamic INventory Allocation Maintenance Optimizer), with overbooking and multi-class discount allocation controls was launched in 1985 (Cummings, 2007). Work on yield management at American started in 1979 and went through five iterations (Smith, 2007); starting with a rules-based system to the launch of DINAMO in 1986. The five iterations were the Multiclass Optimization Modeling System (MOMS) in 1979, Discount Allocation Decision System (DADS) in 1980, City Allocation Reporting System (CARS) in 1981, the Super City Analysis and Reporting System (SCARS) in 1982 leading to DINAMO in 1986.

These first-generation applications were deployed on IBM mainframe computers with the MVS operating system. The SCARS reports were produced from a MARK IV program and given to the analysts for review. By the time DINAMO arrived, the yield management analysts accessed data and reports with an IBM mainframe 3270 terminal emulator.

As President of Texas International Airlines under Chairman Frank Lorenzo, Donald Burr revitalized the airline industry with the introduction of the Peanut Fares in 1977. This was before airline deregulation, and Texas International Airlines sought permission from the CAB to offer these fares at half the price on selected low-density routes, making it the fastest growing airline in the country, earning $8.2 million on revenues of $145 million (Easterbrook, 1987). Burr left Texas International Airlines in January 1980 (Gordon, 1989) to start his own low-cost airline, PEOPLExpress which was launched in April 1981. The airline grew at a rapid pace with its primary hub in Newark, New Jersey. By 1985, Donald Burr’s PEOPLExpress Airlines was listed by Fortune as one of the fastest growing companies. By undercutting fares offered by competing airlines, PEOPLExpress market share increased rapidly to become the fifth largest airline in the U.S. Left with no choice but to match the low fares, Crandall relied on American’s new yield management system, DINAMO, to capacity control the availability of these deeply discounted fares. Changes were also made in the reservations system, Sabre PSS, to accept these new discounted inventory controls from DINAMO, and seat availability was based on the net nesting method (Vinod, 2006) with segment limits. On January 17, 1985 American introduced the nonrefundable Ultimate Super Saver fares which were priced lower than the PEOPLExpress fares. For the very first-time yield management analysts were able to explicitly control the availability of the deeply discounted fares. In the absence of yield management controls, PEOPLExpress sold all seats at the deeply discounted prices, which were not sustainable over the long run to support the airline’s cost structure. The implementation of yield management by American Airlines had a significant impact on PEOPLExpress. By September 1986, PEOPLExpress was in deep financial trouble and eventually ceased operations on February 1, 1987 when they merged with Continental Airlines. PEOPLExpress failed for two reasons: the absence of yield management controls and the rapid uncontrolled expansion of the fleet with the acquisitions of Denver-based Frontier Airlines, Britt Airways and Provincetown-Boston Airlines which placed an enormous debt burden on the airline. American Airlines used their new strategic weapon to control seat inventory and maximize network revenues. American matched PEOPLExpress fares but offered only a few seats at these deeply discounted prices. Donald Burr, Chairman and CEO of PEOPLExpress Airlines summed it up in 1986 (Cross, 1997).

We were a vibrant, profitable company from 1981 to 1985, and then we tipped right over into losing $50 million a month. We were still the same company. Still at Newark. There were no changes in this company. What changed was American’s ability to do widespread Yield Management in every one of our markets. We had been profitable from the day we started until American came at us with Ultimate Super Savers. That was the end of our run because they were able to underprice us at will and surreptitiously. There was nothing left to defend us.

All we had left was our cost structure, which at the time was a billion dollars a year less than American. You figure that at a billion dollars cheaper, you ought to be safe. We kept naively hoping that our billion- dollar cushion would give us enough room even if they underpriced us here and there. But all they needed to take away from us was that marginal traffic above breakeven. You don’t have to take away half the guy’s market. All you have to do is take away a few seats on every flight and the guy’s dead.

Donald Burr

Chairman and CEO

PEOPLExpress Airlines

Burr realized the significance of yield management late in the game and was quoted as saying (Bryan, 1989) "What you don’t know about Revenue Management could kill you!"

There is another aspect of what American did. When American Airlines introduced the deeply discounted Ultimate SuperSAAver™ fare in 1985, prices were discounted 70% and ranged between $39 and $129 but were capacity controlled, bringing air travel within reach of an entirely new population that had never flown before.

Yield management maximizes revenue from perishable inventory based on the fundamental premise that all customers are not created equal. Seat inventory is perishable, since once the flight departs, the unsold seat inventory is lost forever. Yield management is the process of selling the right seat to the right customer at the right price at the right time. Yield management is called revenue management today since it is revenue and not yield (revenue per revenue passenger mile) that is maximized. The transition happened officially in 1993 when the IATA Conference on Yield Management, which was organized in 1988, was rebranded as the IATA Conference on Revenue Management.

Development of the first yield management system started off as a rules-based system (Smith, 2007) before being enhanced to a decision support system based on marginal revenues. The economic overbooking model (Smith, 1982) in DINAMO evaluated tradeoffs between denied boarding costs and the cost of a spoiled seat. The discount allocation model was an expected marginal seat revenue model, an extension of Littlewood’s Rule (Littlewood, 1972) to multiple booking classes, which generated the joint protection levels for higher valued booking classes relative to the lower value booking class. This method was later called EMSRB (Belobaba, 1992) as a computationally viable heuristic to optimal booking limits (Belobaba, 1992; Curry, 1990) that requires the evaluation of multi-dimensional numerical quadrature (convolution integrals) and more compute power. Seat inventory controls for this first-generation yield management system were based on leg/segment controls—leg class nested inventory controls with segment close indicators and segment limit sales by booking class.

Beyond the decision support capabilities, the success of yield management at American was made possible with advances in inventory control on American’s Sabre Passenger Sales and Service (PSS) system (the host CRS) (Vinod, 1990) where American’s schedules and fares were stored.

American Airlines established its hub at Dallas/Fort Worth International Airport (DFW) on June 11, 1981. Central to the launch of Growth Plan I and Growth Plan II in the 1980s by American’s CEO Robert Crandall was the introduction of the McDonnell Douglas MD-80 (Super 80) aircraft into the fleet in 1983. The large number of new MD-80s entering service spurred the transition from a point-to-point network to the creation of a hub and spoke network at American’s two primary hubs, Dallas/Fort Worth and Chicago O’Hare to provide air services for a wide geographic area.

With the growth plan the fleet more than doubled in size between 1984 and 1990. Crandall also created the B-scale, a lower pay scale for new pilots. New hubs were added in Nashville, Tennessee, Raleigh-Durham, North Carolina, San Jose, California, Miami, Florida and San Juan, Puerto Rico. Expansion of the fleet and workforce led to lower operating costs, increased revenues, and profits. At one point, American had 363 MD-80s in its fleet. Crandall also found other creative ways to cut costs (Mayerowitz, 2011). In 1987 he famously removed one olive from each salad service on American’s flights. He reasoned that passengers would not notice, and the airline would save at least $40,000 per year. With cost in mind, he also preserved the airline’s distinctive look. The planes were polished but not painted. "No paint means less weight, Crandall once explained. The unpainted look, keeps the sun glinting off our ‘silver birds.’"

In a post deregulated world, a hub and spoke network provided service to more markets through the hubs and increased aircraft utilization. Hub economics were based on a mix of 30% local traffic and 70% connecting traffic. With the transition from a point-to-point network to a hub and spoke network, leg-based inventory controls generated by DINAMO had limitations and steps were undertaken to transition from a leg/segment yield management system that optimized controls by flight to an Origin and Destination (O&D) based system that maximized network revenues by capturing the value of the reservation request by O&D. The first O&D yield management system was deployed in 1987 with virtual nesting (Smith, 1986; Smith et al., 1992) controls, which won the prestigious Edelman award in 1992. With virtual nesting, O&D classes were mapped to virtual buckets to control inventory by O&D. Virtual nesting produced an additional 1%–2% in incremental revenues over leg-based controls. Inventory in the Sabre PSS was upgraded to support virtual nesting controls. Optimal mapping of itineraries to virtual buckets based on value produced an additional 0.4% in revenues (Vinod, 1989). American calculated that the systematic use of yield management enabled the company to generate $1.4 billion in incremental revenue between 1989 and 1991, while AMR’s (the parent company) profits were $892 million over the same period. Among the European carriers, SAS deployed O&D controls in 1993 (Petersen, 1996). Keeping abreast with the evolution of yield management, the real time transaction systems, the Global Distribution Systems (GDS) and airline central reservations systems (host CRS), brought to market some of the key enablers for O&D control such as seamless (interactive) availability and sell for true last seat availability by point of sale (POS), married segment control and journey control from the 1980s to early 1990s.

In the 1990s, yield management was renamed revenue management since revenue and not yield is maximized. Based on the premise that customers pay different fares for the same origin and destination, airline customers are conditioned to paying different amounts for the same product depending on when they book and the details of the itinerary. Every time a person boards a flight, it is highly likely that the person in the next seat paid a different price.

When Tom Cook was asked to compare his management style with that of Bob Crandall after they had both departed from American Airlines, he remarked (Horner, 2002): Bob is a very unique and incredibly talented chief executive. I admire him tremendously. As for my own management style, you would have to ask the people who work with and for me. I consider myself a pretty much hands-on guy and I share that with Bob. I look for the right people and give them the authority to get things done. I’m results-oriented.

1.5 Origins of the Frequent Flyer Programs

Texas International Airlines created the first mileage-based frequent flyer program in 1979, but it was dismantled within a year and later merged with Continental Airlines in 1982. A year later, in 1980, Western Airlines created the Travel Bank program. This program introduced by Western Airlines (‘the only way to fly’), which awarded a US$50 travel certificate for every five trips, was not mileage based. American Airlines launched the AAdvantage frequent flyer program on May 1, 1981. Robert Crandall is credited for launching the frequent flyer program as an instrument to promote the brand to repeat customers, track frequent flyer consumption patterns and create targeted offers for the most loyal customers. The idea evolved from William Bernbach (Levenson, 1987), the legendary founder, and CEO of Doyle Dane Bernbach, American’s advertising agency, based on his observation of commercial banks who were wooing their best customers with free products such as toasters and electric blankets. He advocated a special loyalty fare for frequent flyers to American Airlines. Though the loyalty fare was never rolled out, it eventually led to the mileage based AAdvantage program.

The first wave of mileage based frequent flyer programs, all launched in 1981 after American’s AAdvantage program, is summarized in Table 1.1.

Table 1.1

Frequent Flyer Programs in 1981

It is a well-accepted fact that the loyalty programs are a resounding success for the airlines. These loyalty programs create new demand, build brand equity although the end product is more or less a commodity, build a lifetime relationship with loyal customers who are married to the frequent flyer program and above all serve as a secondary source of revenue through joint marketing programs with credit card companies, telephone companies, and numerous promotions with packaged goods companies.

At present the largest airlines in the world boast memberships of more than 50 million and partnerships with hundreds of companies. Since inception, frequent flyer programs have been embraced by airlines worldwide in a highly competitive marketplace to lock-in customers with frequent flyer miles and add-on incentives. Loyalty programs have expanded into hotels, rental car, casinos, financial services, grocery chains, apparel retailers and restaurants. Loyalty programs offered by travel suppliers such as airlines and hotels have become so pervasive that it is difficult to imagine a world without them. Over the years, they have enabled travel suppliers to extend their reach by selling miles and points to credit card companies and retailers who in turn use it as an incentive to sign up new customers, stimulate demand and gain repeat sales. Although loyalty programs should be leveraged for revenue growth, they are frequently wrongly perceived as a necessary evil, the cost for doing business.

A challenge faced by a loyalty program is to estimate the share of wallet of a customer. A customer may be loyal based on the amount of business transacted. However, the customer may spend twice as much on a competitor. Unfortunately, this data are not accessible. A combination of surveys and predictive analytics based on a loyal customer’s purchase behavior patterns can be used to make inferences on this important statistic. Large brick-and-mortar retailers like Walmart and Target face a somewhat similar problem. To gain insight into the share of wallet of a customer, credit card companies monetize their data by selling aggregate spend data by postal code back to the retailers. This in turn influences marketing spend by retailers on free standing inserts (FSI), print advertising in local newspapers and circulars. Harrah’s uses predictive analytics to determine share of wallet. A lot has been said on Harrah’s successful total rewards program (Loveman, 2003). The loyalty card tracks a customer’s playing time, money won and lost, and credits accumulated to build history of each gambler’s behavior at Harrah’s properties. Central to the success of this program are customer segmentation and the use of advanced predictive analytic tools to determine which customers should be targeted with offers and incentives to return and play. Although data are readily available on past customer visits from historical data, predictive models reveal which customers are likely visiting other casinos in the market and be used to determine which customers should be targeted with offers and incentives to return and play.

Loyalty programs also have a significant impact on revenue management. Every frequent flyer redemption displaces revenue passengers, which represents a cost to the airline. Frequent flyer redemption policy is influenced by its impact on revenue management. Frequent flyer redemption bookings are capacity controlled.

1.6 Origins of the GDS

From the mid-1960s to the mid-1970s, both airlines and travel agencies promoted the concept of a neutral industry-wide reservations system to standardize reservations workflows and enhance agent productivity (Vinod, 2009). The Reuben H. Donnelly Corp. (publisher of the Official Airline Guide) initiated development of the Donnelly Official Airline Reservations System (DOARS) on a Univac platform in 1967. It subsequently failed because of lack of financing from the 21 airline participants.

The Automated Travel Agency Reservations System (ATARS), a system that was exclusively for airlines and travel agents based on PARS, sought approval from the Civil Aeronautics Board (CAB). The U.S. Justice Department interpreted the exclusivity feature as a per se violation of anti-trust laws, which in turn triggered a CAB investigation. The ATARS agreement was modified to meet the requirements but was abandoned before CAB could reach a verdict.

The contract to develop a common integrated travel agency system between the American Society of Travel Agents (ASTA) and Control Data Corporation (CDC) in 1973 also failed when the airline constituents could not accept a computer vendor controlling access to travel agents.

To gain control of the distribution channel, American proposed the creation of a joint task force consisting of airlines, ASTA and hardware vendors to develop the Joint Industry Computerized Reservations System (JICRS) in 1974. American’s Robert Crandall and Max Hopper played key roles in defining the functional scope, benefits, and costs of JICRS. In July 1975, the technical evaluation team concluded that the development of JICRS was technologically feasible and would produce significant cost savings to the airline participants. However, United did not find the terms of the financing of JICRS acceptable. It was tied to passenger volumes and United, by virtue of being the largest domestic U.S. carrier, would become the largest investor. JICRS failed when United announced that it would actively promote its PARS-based Apollo system to travel agents within 9 months.

Carriers such as American and TWA were forced to adopt similar plans and the race was on to develop and actively promote these systems to the travel agency community. ASTA made a final attempt to create a joint system, known as the Multi-Access Agent Reservations System (MAARS). It failed when CAB refused antitrust immunity over concerns with exposing the other airline CRSs being promoted to the travel agencies.

When attempts to create a joint system for travel agents failed in the 1970s, airlines started promoting their reservations systems to travel agents to sell their products. Agency features were added to the reservations system and the host CRSs with the enhanced capabilities used by travel agents identified during the JICRS study came to be known as the Global Distribution Systems (GDS). These systems developed agency point-of-sale support worldwide, whereas the airline system where seat inventory was stored, is known today as host CRSs or simply CRSs. Hosting multiple airlines in a single system followed quickly.

The display of flights on the Sabre system did not follow the OAG (Official Airline Guide) format. The OAG format displayed nonstops followed by direct flights, online connections and interline connections. The Sabre display was much more sophisticated and based on total travel time (also called elapsed time), displacement from requested time and carrier preference. Differences in the display order between OAG and Sabre led to the perception that the display was biased toward American. Interestingly each of these three schedule parameters feature very prominently in consumer choice model calibrations today to understand consumer behavior and how they select an itinerary from a choice set.

In 1976 there were 130 travel agencies that subscribed to Sabre. Max Hopper, the chief visionary of the Sabre system owned by American Airlines was the first to market with a product for travel agents. American aggressively marketed the reservations system as an extension of the airline sales force, to sign up new subscribers. It was perceived by competitors that the product display of flight schedule and availability favored American over other carriers.

United’s Apollo was introduced to the travel agency community a few months later. Apollo did not include the travel agency features, identified in the JICRS study, which Sabre supported and United quickly realized that travel agents preferred the Sabre system. Trans World Airlines (TWA) introduced their system (TWA PARS) shortly after Apollo. They also introduced PARS II with limited features, which found a niche with small travel agencies., followed by the Trans World Airlines (TWA) PARS system. The common perception was that the airline owned proprietary reservations systems gave these airlines an unfair competitive edge over airlines that were not promoting a host CRS of their own.

MAARS Plus, a system not owned by airlines or travel agencies was launched in 1977 by ITT. They made the lofty claim that it was the only unbiased system, which was not true. This system had unique features; it offered direct connections to airline reservations systems of all participating carriers. MAARS Plus also stored the reservations on the airline reservations systems. This system failed for two reasons. First, the absence of a common language made it difficult for travel agents to understand the various codes from the individual airline systems and, second, the revenue model was flawed because investors could not recover subscription fees from the travel agents.

From a product development and capability perspective, Sabre was a year ahead of the other CRSs. Most of the CRSs had similar marketing strategies—in return for the hardware and training that was provided at no cost, the agencies agreed to pay a subscriber fee based on the volume of bookings made by the agency. Agencies offset these expenses with the standard commissions paid by the airlines, about 10%. In addition, they also received override commissions to encourage bookings on specific airlines, based on thresholds achieved for key performance indicators (KPIs). The volume of business at travel agencies expanded rapidly with the agency channel providing most of the bookings for the airlines.

After airline deregulation, facing stiff competition, airlines with vested interests in agency reservations system started investing in back office systems. American bought Agency Data Systems (ADS), after United had acquired a license to use ADS. All systems today have back office systems which are required for agents to view agent sales reports (ASR) and commissions.

When smaller airlines and travel agencies complained, CAB identified four anti-competitive behaviors that were prevalent in 1983. The Code of Federal Regulations (CFR) was introduced in November 1984 which prohibited anti-competitive behavior from airlines that owned the dominant CRSs. They were:

1.

Carrier preferencing in displays was eliminated and all displays were required to display results which best met the query parameters submitted by a travel agent. Hence, on United’s Apollo, for a specific market, an American flight may be ranked first and appear on the first line of the display.

2.

Capabilities available in the CRS for a specific airline should be universally available to all participating airlines.

3.

Discrimination between CRS providers was also banned which ensured that all airlines that owned and marketed a CRS to travel agencies participated in all the CRSs. Hence an airline that participated in one CRS had to participate in all of them. It was also mandated that travel agency booking data collected by a CRS were required to be made available to the competing CRSs for a fee. This is now referred to as MIDT (Marketing Information Data Tapes).

4.

Booking fees charged by the CRSs to participating carriers were required to be non-discriminatory. Before this ruling, price discrimination prevailed, and some carriers paid no segment booking fees whereas others did. Booking fee billing data previously provided on microfiche or paper is provided as BIDT (Billing Information Data Tapes) by the CRS vendors.

The CRS rules went into effect on November 11, 1984, though it did not dispel the debate over anti-competitive behavior.

During this period, a notable feature launched by Sabre was Sabre Traveler Automation Records (STARS) that eliminated the manual customer contact list maintained by travel agents. The CRSs came to be known as the GDSs as the systems developed agency point-of-sale support worldwide, whereas the airline system where seat inventory was stored is known today as host CRSs, or simply CRSs. Hosting multiple airlines in a single system followed quickly.

When airlines realized the value of GDS ownership, several new GDSs appeared in the market in the 1980s. Eastern Airlines launched System One Direct Access (SODA) in 1981. It became operational in 1982 and was based on PARS. Delta Air Lines launched DATAS II in 1982 based on PARS technology, terminating the joint marketing agreement with United’s Apollo system. Abacus was founded in 1987 by Cathay Pacific Airways, Singapore Airlines and Thai Airways International PLC. Thai Airways later dropped out of the partnership and other airlines joined the partnership. PARS and Abacus signed an agreement that created the foundation for Worldspan.

Amadeus Global Travel Distribution was created in 1987 as a joint venture between Air France, Lufthansa, Iberia, and Scandinavian Airlines System (SAS). System One, developed by Eastern Airlines was the baseline for the Amadeus reservations system (passenger name record) code running on TPF and the pricing engine was from Air France, running on Unisys. The system became operational in 1991 by integrating four national reservations systems, Esterel in France, Savia in Spain, Smart in Sweden and START in Germany (Kärcher, 1996). These