Carnival Improves Forecast Accuracy Using Optimiza
Carnival is “The World’s Most Popular Cruise Line”, with 24 ships operating 3 to 18-day voyages to the Bahamas, Caribbean, Mexican Riviera, Alaska, Hawaii, Canada, New England, Europe, Bermuda, South America, Panama Canal, and Tahiti and Fiji Islands.
The English division of the company, Carnival UK, has 1,200 employees and 10 ships under two brands, Cunard and P&O Cruises. Cunard has three ships with names of famous queens: Queen Mary 2, Queen Victoria and Queen Elizabeth. P&O Cruises has seven: Adonia, Arcadia, Aurora, Azura, Oceana, Oriana and Ventura. In 2015, the 11th ship, Britannia, will complement the fleet.
Organising cruises all around the world on luxury liners that can accommodate up to 3,500 passengers feels like a dream holiday. However, when you take into account the challenge of feeding, entertaining, and accommodating this number of people for several days and the logistical complexity induced by the diversity of ships, routes, times, profiles of passengers, products, as well as anticipating and planning the needs of these ships before each port of call can easily turn into a headache and one that requires accurate forecasting and contingency planning.
In essence, each ship is a floating town separated from its many suppliers for lengthy periods. Providing for its inhabitants during this time is nothing short of an epic undertaking. As patterns and behaviours must first be established before predictions can be made, no cruise line has ever been able to implement forecasting software to support preparations before embarkation as itineraries differ considerably in terms of destination, course, consumption, and time at sea. There are also seasonal fluctuations with availability of fresh produce and passenger numbers.
P&O Cruises and Cunard sought to find an accurate demand forecast solution that could handle this multifaceted model whilst remaining easy to understand, operate and maintain.
To help the planners of their ships, Carnival UK decided to optimise its complex forecast planning using the Optimiza demand and forecasting tool, with a centralised operation unit. Many intricate considerations had to be incorporated such as analysis of timelines, the usage patterns of consumables, restocking with local foodstuffs, purchases from bars, and demand for fresh linen and towels.
Existing processes used by the two companies were reviewed with a view of applying Optimiza’s Demand Planner module, an advanced demand planning software application designed to compile and maintain an accurate requirement forecast. The management tool provides planners with an early and accurate prediction of future sales and consumption that supports rapid and effective decision making. Typically, customers have reported 15-20 percent improvement in forecast accuracy.
Optimiza used multi-level forecasting techniques that intelligently select the right algorithms for each product level and collected results up or down the product, geographical, or channel hierarchy. Optimiza integrated seamlessly with the Carnival UK ship-based MXP software, which provides powerful tools to manage cruise ship operations.
In terms of results, Carnival UK firstly measured the accuracy of the software’s predictions, an initial target of 78 percent accuracy was set for Cunard’s Queen Victoria, which was achieved. Predictions corresponded to a minimum safety stock requirement for three days for cruises of 14 days on average.
At present, the forecasts include 4,000 references by ship which is 95 percent of the volume and value. The forecast quality is good and additional work on capacity management and safety stocks will improve the results further.