Flight Cost Management for a Cruise Line
Overview
When vacationers book cruises, they often purchase a combined air and cruise package. Cruise lines typically set prices for the package based on market conditions. If the actual airfare paid to transport the customer is less than expected, the cruise line earns a greater margin. If airfare is more than expected, the margin decreases. Even when a cruise line charges separately for the cruise and for airfare, the “air fare supplement” is typically only an approximation of the actual airfare incurred to transport the passenger.
For a large cruise line, airfare expenses can exceed 20% of revenue. The cruise line typically negotiates airfares on a city-pair basis (e.g., Dallas – Miami, Grand Rapids – Miami) with each airline. Fares for the same city-pair can vary widely between airlines. For a city-pair, cruise lines may or may not have “guaranteed access” to its negotiated fare for a certain number of seats. When additional seats are needed, the cruise line may pay higher airfares.
Reducing airfare costs can have great leverage, as each additional dollar paid on airfare often represents one dollar less profit. Consequently, cruise lines have a strong interest in minimizing the total cost of routing all its passengers across its flight network, subject to meeting certain customer service standards.
A major cruise line was having difficulty controlling its airfare expenses. We were asked to review the models and procedures they used for assigning cruise passengers to flights, and if possible, design decision support models and procedures that would help them reduce their airfare costs.
Approach
Our review of the cruise line’s efforts included direct observation of how cruise staff assigned passengers to flights, interviews with cruise staff, and an extensive review of the cruise line’s information technology architecture and system capabilities. We also downloaded historical reservation and flight and rate data so we could benchmark the costs of the cruise line’s flight assignment decisions with the costs that would have been incurred using the network optimization model we designed.
We designed both tactical and strategic decision support tools to help the cruise line reduce its airfare expenses. For strategic use, we designed two decision support tools intended to give the cruise line more information and leverage when it negotiated rate and space needs with the airlines. In addition, one of the models was also designed to help identify opportunities for the cruise line to use charter flights to reduce its airfare expenses.
We designed a tactical route optimization tool to assist the operational efforts of cruise staff assign the least cost air routings for passengers, subject to meeting specified customer service levels.
The following example illustrates the flight assignment problem faced by the cruise line. Suppose 20 passengers need to be routed from Buffalo to Miami and 10 need to be routed from Philadelphia to Miami. Further, suppose there are two routes and fares available from Buffalo and two from Philadelphia. Table 1 defines the routes and fares. The leg or segment flight numbers are given, as are the leg capacities. The capacity number given is the number of seats that the cruise line can obtain from the carrier.
| Route | Routing | Leg 1 Flight Number | Leg 1 Capacity | Leg 2 Flight Number | Leg 2 Capacity | Fare ($) |
|---|---|---|---|---|---|---|
| 1 | Buffalo-Newark-Miami | 180 | 25 | 555 | 20 | $220 |
| 2 | Buffalo-Cleveland-Miami | 200 | 25 | 244 | 25 | $240 |
| 3 | Philadelphia-Newark-Miami | 311 | 20 | 555 | 20 | $200 |
| 4 | Philadelphia-Miami | 455 | 20 | N/A | N/A | $250 |
The lowest cost feasible routing of the 30 passengers is obtained by placing 10 Buffalo passengers on Route 1, 10 Buffalo passengers on Route 2, and all 10 Philadelphia passengers on Route 3. This results in an airfare cost of $6,600. Note that if you first route all the Buffalo passengers on the cheapest available routing (Route 1), then the Philadelphia passengers must be routed on Route 4 (because of the limited capacity on Flight 555 of 20 passengers). This alternative routing results in a total airfare cost of $6,900, an increase of $300 or 4.5 percent. While determining the least cost solution is relatively simple to calculate for just two gateway cities and one cruise port, optimizing over hundreds of gateway cities and multiple ports, and factoring in reservation cancellations and the potential need/opportunity to reassign passengers, requires a flexible network optimization methodology.
While the financial goal is to minimize the total cost of routing passengers, there are a variety of customer service standards that need to be considered during this process. For example, passengers may desire nonstop flights rather than connections, “high quality” carriers, jet service instead of propeller planes, etc. Additionally, certain passengers must be routed together. The cruise line also had “soft” preferences with respect to the service level provided and volume commitments to the airlines that needed to be met. As noted before, a limited number of seats are available on each routing, based on blocked seats held by the cruise line and their ability to obtain additional seats at various price points.
We designed a network minimum cost flow model to consider all the needed constraints. In our benchmark analysis, using a simplified prototype model, we found that our proposed approach would reduce the cruise line’s airfare costs by 3 percent to 16 percent per voyage.