ORMS Today
June 2000

www.informs.org/Executives -- Corporate OR The Sabre Story

The making of OR magic at AMR

By Peter Horner

Whenever operations researchers gather around the optimization model and start discussing the greatest OR group since the advent of the desktop computer, the conversation inevitably routes them through the Dallas-Fort Worth Airport. That's because DFW is the hub of American Airlines and home to Sabre, unquestionably the biggest and arguably the best and most influential OR group in the world.

The relationship between American, Sabre and operations research is as robust and complex as the models and problems they tackle every day, but most of all, it's a richly rewarding relationship for all concerned. AMR, the parent of American Airlines and stepparent of Sabre, provided a nourishing home in which OR soared to unprecedented heights in terms of impact — on American, on the entire airline industry and on a good chunk of the rest of the corporate world.

The making of OR magic at AMR is a legendary tale in the operations research community, one that is constantly pointed to as a shining example of just how much competitive advantage can be gained when good OR hooks up with supportive management over some juicy OR problems. We'll pick up the story in 1982, when erstwhile college professor Thomas M. Cook joined American Airlines. At the time, the airline's OR group consisted of a dozen talented individuals toiling away on potentially high-impact projects (like rudimentary yield management; see accompanying article) while buried in the IT Department, far removed from the consciousness of upper management.

"They didn't have a lot of corporate sponsorship," Cook recalls. "They weren't working on as many important problems as they should have, and there was almost no senior management visibility."

Once elevated to group director, Cook set out to raise the team's profile within the corporation. But first things first.

"Working on important problems was the obvious way to gain senior management's attention," Cook says. "But how do you get those problems? The first thing we did was transition the group from a very research/academic organization into one that looked more like a consulting firm."

Under Cook's guidance, the OR group created formal relationships with the internal client, American Airlines, and drew up formal proposals. Soon the group began to hit, in Cook's words, "a lot of singles before we went for the home run.

"We made sure that we wrote clear proposals so that the client knew what we were going to do, and then we made sure we delivered on our promise," Cook adds. "We did a better job of packaging our deliverables so that they looked professional. We had a lot of small wins along the way. We'd do a nice project, and people would hear about it. During the first year I was there, though, I don't think Bob Crandall even knew my name."

Crandall, the legendary boss of AMR, was widely considered the toughest, smartest CEO in the airline industry. Crandall never built a mathematical model, but he understood what optimization could do to improve his company's competitive position. He wasn't afraid to take calculated risks. As soon as he discovered the power of OR to quantify risks, he became the profession's biggest supporter, or, as MIT professor Arnold Barnett once described him, "The Patron Saint of Operations Research."

"Bob was tough, smart and intellectually curious," Cook says. "He always wanted us to explain how these models worked and why he should have confidence in the results. He was always probing. You could show Bob a thousand numbers, but if one of them was wrong, he would point it out in a nanosecond. You had to be on your toes because he had a reputation for taking no prisoners.

"If you didn't have the answer, you had to tell him. But when you knew you were right, you had to stand up for what you believed in. If you did those two things, you'd get along fine. Bob appreciated what operations research brought to the table."

Within five years of Cook's arrival, the OR department had grown to include 75 professionals who were, in Cook's words, "building models for important decision-support systems and performing strategic studies to support senior management decision-making at American Airlines." The work ranged from improving crew scheduling to analyzing a multi-billion dollar expansion at DFW. "Those were big projects, Bob Crandall-type projects," Cook says. "That meant we got a lot of exposure to Bob and senior management. After that, things really took off."

Cook recalls his first one-on-one meeting with Crandall in the mid-1980s. Impressed by what he saw coming out of the OR department, the CEO wanted Cook to draw up a business plan to expand the group. Cook came back with a proposal to triple the number of analysts, a plan that Crandall quickly approved. Shortly thereafter, Cook began attending all of Crandall's staff meetings and many of the Planning Committee meetings involving senior management. "I was there to recognize opportunities where operations research could make a difference," Cook says. "No matter what we were talking about, if I thought we could add value, I'd speak up. Over time, we got a lot of assignments."

In 1987, Cook explored the possibility of marketing the OR group's rapidly expanding expertise and portfolio of software solutions outside of American Airlines. He drew up another business plan that was approved by Crandall that created a new entity entitled American Airlines Decision Technologies, or AADT. Its charter was to continue to be the operations research arm of American Airlines, as well as create stand-alone external business as a wholly owned subsidiary of AMR.

AADT quickly became a profit center for AMR. Amtrak was the first external client, followed by Qantas Airlines. Business snowballed, and Cook's biggest headache was keeping up with growth, not creating demand. From 1988 to 1994, AADT grew from 75 people to more than 600, making it the largest OR group in the world. By 1994, AADT was doing about $70 million a year in business, 80 percent of it outside of American.

Rising costs and fuel prices combined with increased competition sent the airline industry into a tailspin in the early 1990s. AMR management responded by growing the airline where it was profitable, shrinking non-profitable areas (i.e., closing certain hubs), and expanding the non-airline businesses (primarily Sabre Travel Information Network, a $1 billion travel distribution business). AADT fell into the latter category.

A series of mergers and realignments followed. Since 1993, the OR contingent has operated under the banner of AADT, Sabre Decision Technologies, Sabre Technology Solutions, The Sabre Group and now simply Sabre. At each step along the way, the combined group grew — from 600 in 1994 to 10,000 professionals worldwide today. Crandall retired in 1998. Cook departed a year later, but not before building a hugely successful business with operations research at its core. Although just a fraction of the 10,000 current employees of Sabre are considered operations researchers, their influence and impact greatly exceeds their number. In the minds of many, the OR group founded at American Airlines sets the standard by which all others are measured.

Mother, Father of Invention Produce Golden Child: Revenue Management
If necessity is the mother of invention, then deregulation is the father, and revenue management (also known as yield management) is the couple's golden child — at least as far as operations research is concerned.

The wave of deregulation now altering the landscape of the communications, gas and electricity markets crashed down on the airline industry back in the late 1970s. Deregulation spawned a host of low-cost airlines such as People's Express and Southwest that could make money while selling seats at a fraction of the price charged by established carriers like Pan Am and American Airlines— a huge competitive advantage.

The big carriers, locked into labor contracts, had no clear way of reducing costs. So the question all major carriers faced was, How can we compete? If we lower fares, it doesn't matter if we have a 100-percent load factor — we'll still lose money. And if we don't match the low-cost airlines' fares, we'll lose market share, and eventually we'll be put out of business.

American Airlines solved the dilemma by offering different fare classes in each market. American matched the low-cost airline's prices with a portion of its seat inventory, and then saved another portion for late-arriving, high-yielding demand (typically, the business traveler). Thus was born the concept of yield management (now widely referred to as revenue management, since it doesn't manage yield as much as it maximizes revenue).

It took a group of operations researchers at American Airlines, however, to turn the concept into capital. After all, several major questions remained to be answered: How many seats do you sell at a discount? How many do you hold in reserve? When do you change the price of each "bucket" of inventory and by how much?

The answers were found in the OR tool bag: forecasting and optimization. Operations researchers used the former to forecast demand by inventory bucket, and then used the latter in allocating the discounted seats in an optimal fashion.

The initial implementation was very crude by today's standards, but it quickly earned its keep. Tom Cook, who arrived at American in 1982 and presided over the airline's OR group for most of the past two decades as it mushroomed, merged and morphed into what is now known as Sabre, counts at least four basic generations of yield management during his tenure. Fueled by constantly improving forecasting and optimization models fed by more and better data, each new generation produced in excess of $100 million in incremental profitability over its predecessor. By 1998, Cook estimates the revenue management system at American Airlines was generating nearly $1 billion in annual incremental revenue. To put that figure into perspective, consider that the airline's total operating profit didn't approach $1 billion until 1997.

Today, virtually every airline in the world employs some sort of revenue management system. The concept is also hard at work throughout the cruise, hotel and car rental industries, and wherever else you find perishable, limited capacity.
— Peter Horner

Peter Horner is the editor of OR/MS Today.

  • Table of Contents

  • OR/MS Today Home Page

    OR/MS Today copyright 2000 by the Institute for Operations Research and the Management Sciences. All rights reserved.

    Lionheart Publishing, Inc.
    506 Roswell Street, Suite 220, Marietta, GA 30060, USA
    Phone: 770-431-0867 | Fax: 770-432-6969
    E-mail: lpi@lionhrtpub.com
    URL: http://www.lionhrtpub.com

    Web Site Copyright 1999, 2000 by Lionheart Publishing, Inc. All rights reserved.