• Home   /  
  • Archive by category "1"

Airline Revenue Management Case Study

0:16Skip to 0 minutes and 16 secondsHi Bill. Hi Christophe. From the start you've been involved in the evolution of revenue management. How has it all started? Well, when I was first in the industry, which was in the sixties, one of the major airplanes --a typical one-- was the 707. It was about a hundred and eighty seats, it went just under a thousand kilometers an hour, and it had a range of about five thousand kilometers. I brought some pictures for you. Here is the old airplane. And the reservation system in those days was not a computer, but business machines. There were people on telephones and maybe punch card machines. And before that, they had a big wooden drum with file cards in slots.

1:00Skip to 1 minute and 0 secondsEach flight had a file card and they would, when someone would call up, write his name on a card and put it back in the drum. So each day of the week or of the month was down in the drum and each flight was around the drum. And that was the reservation system, before revenue management came in. But, in those days, all the travelers were business travelers. It was quite high priced. Leisure travelers and poor people travelled in buses and trains. So it was all very fancy people and quite high priced in its time.

1:27Skip to 1 minute and 27 secondsNow today it turns out a typical airplane would be the 737-800 or the A321, which has about a hundred and eighty seats and goes just under a thousand kilometers an hour and has a range of about five thousand kilometers. So the airplane hasn't changed very much --in its performance at least-- but the whole reservation system and booking is a whole lot different, because it is all on computers and very controlled. The airplanes are no longer just full of business passengers. They're mostly --75-80%-- leisure travelers.

2:00Skip to 2 minutes and 0 secondsSo Bill, what is revenue management? Revenue management is really three different ideas.

2:08Skip to 2 minutes and 8 secondsI brought a slide for that as well. The first idea is to separate out the business travelers from the leisure travelers and charge them different fares. You have some way of separating them out. The second idea is to limit the number of tickets you sell in leisure class, so there is plenty of room for the high-paying business travelers. And the third idea is to make sure the discounts are deep enough that you fill a whole airplane with people, because that gives you the most revenue. So that's three ideas, right? That's three ideas and each one we came across by making a really stupid mistake. A mistake? Yes. Can you tell us what sort of mistake leads to these ideas?

2:47Skip to 2 minutes and 47 secondsWell the first mistake has to do with the 747, which was way too big when it first came out. It was twice the size of the 707, the normal airplane of its day.

3:00Skip to 3 minutes and 0 secondsThe reason they put a big airplane where a little airplane used to be is they had a demand model that said that an airline's market share was bigger if it had more seats in the market. Now it turns out that it was a stupid mistake. What really happens was if the airline had a lot bigger market share, it put more seats in the market to carry the passengers. The causality went the other way. So when they put the 747 in, it was empty. They had nobody on the seats. That was the first mistake.

3:28Skip to 3 minutes and 28 secondsHow is it related to revenue management? Well, they had to do something about it. So Boeing went around with this idea they called "surplus seat sales. " The idea was to find a bunch of customers that you weren't serving --that had nothing to do with the business travelers-- and offer them special discounts.

3:47Skip to 3 minutes and 47 secondsThen fill the empty seats with those discount customers. So what really happened is that they started offering low prices to students, to the military, to people in the ministry, to old folks; and to left-handed green-eyed dwarfs.

4:04Skip to 4 minutes and 4 secondsI brought a picture. Yes, I see the troll over there. So that sounds good? Well it was good but, it lead to the next big mistake. Which was what? The discount people bought their tickets early and they bought the whole airplane out. There wasn't any space left for the high-fare customers. So that was very quickly fixed. But it was fixed by a clever idea. Not by a bunch of fancy computer people. And what was this idea ?

4:34Skip to 4 minutes and 34 secondsWell in international flights they had three cabins: first class, business class, and what they call tourist class.

4:40Skip to 4 minutes and 40 secondsFor domestic flights they only had two cabins: a business- and a tourist-class cabin, what they call a full-fare cabin. But they had a computer ready to accept three cabins. So they imagined they would divide the seats that are called tourist-class seats in the back of the airplane into a high-fare group of seats and a discount group of seats, and only sell seats to the discounts from their. There is this little cabin, so they ended up with three

5:11Skip to 5 minutes and 11 secondscabins again: a first class, a full-fare, and a discount cabin. And the big idea was instead of having them fixed --as twelve seats in first class and a hundred and sixty, in discount or in coach-- they changed the size of the full-fare cabin depending on the demand every day for every flight. That was the beginning of the revenue management idea. By 1975 that idea was working.

5:37Skip to 5 minutes and 37 secondsWe had the seats divided in three cabins: first class (what looks like Business class)…… full fare, and then what we call K class as a discount class. I know that one. You're in it all the time, we call it the back of the bus. So that sounds very good and then they manage the system and after that, everything went OK with this? At that point you are in a position that resembles modern revenue management. Because they put it on the computer originally, while making estimates by hand of how much to protect for full-fare. When they put it on a computer, the math geeks got hold of it.

6:19Skip to 6 minutes and 19 secondsThey started optimizing the right size of the cabin against the uncertainty of the forecast for the full-fare demand.

6:28Skip to 6 minutes and 28 secondsThere were lots of studies about that and a lot of discussions of what the real problem was. They made a few little technical errors on the way. There were lots of conferences to try straighten those out. What was one little detail ? The silliest of them was, it turns out if you happen to sell all the full fare seats it is a separate cabin, it is not allowed to reach into the discount seats and use any available up empty seats for full fare. So they changed the cabins --instead of being separate-- to nesting for the high-fare cabin, inside the low-fare cabin. So they protected seats for high-fare but high-fare could reach down into the low-fare seats if it needed to.

7:08Skip to 7 minutes and 8 secondsThat's the kind of thing they came across. By then we're pretty much at what you might think of as revenue management. But there was still one big mistake left to make. Which is the third mistake, right ? That is the third mistake, and I happened to have been right in the middle of this one, so I can tell you the inside story that I haven't found any place else. Because you were there ! I was in American Airlines at the time. And in the 1980's there was a start-up airline in New York City called People Express.

7:34Skip to 7 minutes and 34 secondsPeople Express was a rough airline with crowded seats, and you had to carry a baggage on, and the fare is a very very low. They were connecting a lot of passengers across New York City. Those connecting passengers, American finally figured out --we figured out-- they were coming off of our flights that were connecting over our hubs. We did not want People Express to take our customers. So we decided to teach them a lesson and play mean, and we learned a lot from the lesson. This all happened because American was both one of the hardest playing airlines in competition and one of the most disciplined airlines in reviewing how it was doing. So what happened?

8:13Skip to 8 minutes and 13 secondsWhat happened was American matched the very low discounts of People Express. Expecting to lose money at those high discounted low fares, but hoping to push People Express out of business. Because American was the preferred airline, what happened was everybody who wanted a cheap flight went on American whenever they could get the seats. And because of revenue management they were still protecting the high-fare space, but they sold out all of the empty space every day of the week. On Wednesday they were full, on Sunday they were full. People Express all of a sudden only got the leftovers that did not make it onto American Airlines.

8:55Skip to 8 minutes and 55 secondsSo they had a terrible peeking problem, People Express, because they were full all weekend when everybody wanted to travel, but empty in the middle of the week when everybody was on American Airlines. And they went out of business. Which was the goal. Well it wasn't happy answer for American because Continental took over the People Express operation in Newark and they were much better competitor. So in the end they didn't get rewarded. But! They did look into how they did financially --how much money they lost trying to do it. That was how they discovered the third mistake.

9:27Skip to 9 minutes and 27 secondsWhen they went to study how much money they had lost trying to put People Express out of business, it turned out they made money by playing these dirty tricks. So from then on, the new paradigm for revenue management was to drop the discount fares enough --just enough-- to fill all the empty seats. So instead of having an annual load factor of 65%, it was 85% --now the standard-- and they filled all the seats almost all the time. Which is why the airplanes are always full whenever you get on. I am very happy you're doing this. It is a good thing you are doing. I am happy to be part of it. Thank you. Thank you very much, Bill. Polite applause.


I applied online. The process took 4 weeks. I interviewed at American Airlines (Dallas, TX) in November 2017.


1nd Round was a phone interview by an analyst in the team about work experience and general analytical questions like - You sell 3 glasses of Lemon juice on a Monday. What would you do differently the next day and couple of other questions like the google interview questions for which there are no wrong answers.

2rd Round - On Site - Hiring Manager and a Team of 3 Analysts. Here you need know in detail what revenue management is about . Why is it important for Airlines, RM Models, Airline Types, Some Statistical Knowledge about how to interpret trends and key RM Models and the usual Managerial type questions like why American, why Airline industry, and questions from your resume.

Last Round - Technical Assessment - Was given a bunch of tables and asked to write SQL queries in 15 minutes using pen and paper. Time was very less and queries were a little complex, ranging from using inner queries, joins, various aggregations.

Face to face interview was very comfortable. Except for the technical assessment part, everything else was stress free and very relaxed. Interview experience in my opinion was good.

Interview Questions

  • RM,Why is it important for Airlines, RM Models, Airline Types, Some Statistical Knowledge about how to interpret trends and key RM Models and the usual Managerial type questions like why American, why Airline industry, and questions from your resume.   Answer Question
American Airlines 2017-12-11 12:51 PST

One thought on “Airline Revenue Management Case Study

Leave a comment

L'indirizzo email non verrà pubblicato. I campi obbligatori sono contrassegnati *