Business intelligence and analytics company SAS announced a few days back that it is acquiring revenue optimization company IDeaS. Both, SAS and IDeaS have major development centers in Pune. This article gives an overview of the software that IDeaS sells.
I always found airfares very confusing and frustrating (more so in the US than in India). A roundtrip sometimes costs less than a one-way ticket. Saturday night stayovers cost less. The same flight can cost $900 or $90 (and those two guys might end up sitting on adjacent seats). The reason we see such bizarre behavior is because of a fascinating field of economics called Revenue Optimization.
IDeaS software (which has a major development center in Pune) provides software and services that help companies (for example hotels) to determine what is the best price to charge customers so as to maximize revenue. The technology called pricing and revenue optimization – also called Revenue Management (RM) – focuses on how an organization should set and update their pricing and product availability across its various sales channels in order to maximize profitability.
First it is necessary to understand some basic economic concepts.
Market Segmentation
If you don’t really know what market segmentation is, then I would highly recommend that you read Joel Spolsky‘s article Camels and Rubber Duckies. It is a must read – especially for engineers who haven’t had the benefit of a commerce/economics education. (And also for commerce grads who did not pay attention in class.)
Here is the very basic idea of market segmentation
- Poor Programmer want to take a 6am Bombay-Delhi flight to attend a friend’s wedding. He is willing to pay up to Rs. 4000 for the flight. If the price is higher, he will try alternatives like a late-night flight, or going by train.
- On another occassion, the same Poor Programmer is being sent to Delhi for a conference by his company. In this case, he doesn’t care if the price is Rs. 8000, he will insist on going by the 6am flight.
If the airline company charges Rs. 8000 to all customers, a lot of its seats will go empty, and it is losing out on potential revenue. If it charges Rs. 4000 for all seats, then all seats will fill up quickly, but it is leaving money on the table since there were obviously some customers who were willing to pay much more.
The ideal situation is to charge each customer how much she is willing to pay, but that involves having a salesperson involved in every sale, which has its own share of problems. Better is to partition your customers into two or three segments and charge a different price for each.
Unfortunately, customers do not come with a label on their forehead indicating the maximum amount they are willing to pay. And, even the guy paying Rs. 8000 feels cheated if he finds out that someone else paid Rs. 4000 for the same thing. This is where the real creativity of the marketers comes in.
The person in the Rs. 4000 market segment (leisure travel) books well in advance and usually stays over the weekend. The person in the Rs. 8000 market segment (business travel) books just a few days before the flight, and wants to be back home to his family for the weekend. This is why the airlines have low prices if you book in advance, and why airlines (at least in the US) have lower prices in case of a weekend stayover.
This also keeps the rich customer from feeling cheated. “Why did I pay more for the same seat?” If you try saying “Because you are rich,” he is going to blow his top. But instead if you say, “Sir, that’s because this seat is not staying over the weekend,” the customer feels less cheated. Seriously. That’s how psychology works.
Exercise for the motivated reader – figure out how supermarket discount coupons work on the same principle.
Forecasting demand
This is the key strength of IDeaS revenue optimization software.
You need to guess how many customers you will get in each market segment and then allocate your reservations accordingly. Here is an excerpt from their excellent white-paper on Revenue Management:
The objective of revenue management is to allocate inventory among price levels/market segments to maximize total expected revenue or profits in the face of uncertain levels of demand for your service.
If we reserve a unit of capacity (an airline seat or a hotel room or 30 seconds of television advertising time) for the exclusive use of a potential customer who has a 70 percent probability of wanting it and is in a market segment with a price of $100 per unit, then the expected revenue for that unit is $70 ($100 x 70%). Faced with this situation 10 times, we would expect that 7 times the customer would appear and pay us $100 and 3 times he would fail to materialize and we would get nothing. We would collect a total of $700 for the 10 units of capacity or an average of $70 per unit.
Suppose another customer appeared and offered us $60 for the unit, in cash, on the spot. Should we accept his offer? No, because as long as we are able to keep a long-term perspective, we know that a 100 percent probability of getting $60 gives us expected revenue of only $60. Over 10 occurrences we would only get $600 following the “bird in the hand” strategy.
Now, what if instead the customer in front of us was offering $80 cash for the unit. Is this offer acceptable to us? Yes; because his expected revenue (100% x $80 = $80) is greater than that of the potential passenger “in the bush”. Over 10 occurrences, we would get $800 in this situation or $80 per unit.
If the person offers exactly $70 cash we would be indifferent about selling him the unit because the expected revenue from him is equal to that of the potential customer (100% x $70 = 70% x $100 = $70). The bottom line is that $70 is the lowest price that we should accept from a customer standing in front of us. If someone offers us more than $70, we sell, otherwise we do not. This is one of the key concepts of Revenue Management:
We should never sell a unit of capacity for less than we expect to receive for it from another customer, but if we can get more for it, the extra revenue goes right to the bottom line.
What would have happened in this case if we had incorrectly assumed that we “knew” certainty that the potential $100 customer would show up (after all, he usually does!). We would have turned away the guy who was willing to pay us $80 per unit and at the end of 10 occurrences, we would have $700 instead of $800.
Thus we can see that by either ignoring uncertainty and assuming that what usually happens will always happen, or by always taking “the bird in the hand” we are afraid to acknowledge and manage everyday risk and uncertainty as a normal part of doing business, we lose money.
The Expected Marginal Revenue
The previous section gave an idea of the basic principle to be used in revenue maximization. In practice, the probability associated with a particular market segment is not fixed, but varies with time and with the number of units available for sale.
One of the key principles of revenue management is that as the level of available capacity increases, the marginal expected revenue from each additional unit of capacity declines. If you offer only one unit of capacity for sale the probability of selling it is very high and it is very unlikely that you will have to offer a discount in order to sell it. Thus, the expected revenue estimate for that first unit will be quite high. However, with each additional unit of capacity that you offer for sale, the probability that it will be sold to a customer goes down a little (and the pressure to discount it goes up) until you reach the point where you are offering so much capacity that the probability of selling the last additional unit is close to zero, even if you practically give it away. At this point the expected revenue estimate for that seat is close to zero ($100 x 0% = $0). Economists call this phenomenon the Expected Marginal Revenue Curve, which looks something like this:
There is an EMR curve like this for each market segment. Note that it will also vary based on time of the year, day of the week (i.e. whether the flight is on a weekend or not), and a whole bunch of other parameters. By looking at historical data, and correlating it with all the interesting parameters, Revenue Management software can estimate the EMR curve for each of your market segments.
Now, for any given sale, first plot the EMR curves for the different market segments you have created, and find the point at which the rich guys’ curve crosses and goes under the poor guys’ curve. See the number of units (on the x-axis) for this crossover point and sell to the poor guy only if the number of units currently remaining is less than this.
Applications
Revenue Optimization of this type is applicable whenever you are in a business that has the following characteristics:
- Perishable inventory (seats become useless after the flight takes off)
- Relatively fixed capacity (can’t add hotel rooms to deal with extra weekend load)
- High fixed costs, low variable costs (you’ve got to pay the air-hostess whether the flight is full or empty)
- Advance reservations
- Time variable demand
- Appropriate cost and pricing structure
- Segmentable markets
Due to this, almost all major Hotel, Car Rental Agencies, Cruise Lines and Passenger Railroad firms have, or are developing, revenue management systems. Other industries that appear ripe for the application of revenue management concepts include Golf Courses, Freight Transportation, Health Care, Utilities, Television Broadcast, Spa Resorts, Advertising, Telecommunications, Ticketing, Restaurants and Web Conferencing.
Revenue Management software helps you with handling seasonal demand and peak/off-peak pricing, determining how much to overbook, what rates to charge in each market segment. It is also useful in evaluating corporate contracts and market promotions. And there are a whole bunch of other issues that make the field much more complicated, and much more interesting. So, if you found this article interesting, you must check out IDeaS white-paper on Revenue Management – it is very well written, and has many more fascinating insights into this field.