The future is bright for the Forecaster Cup

Nobody can see the future, but University of Delaware's Alfred Lerner College of Business and Economics students sometimes give it their best shot.

The Lerner College doesn't offer a course in fortune telling. Instead, the prediction in question is part of a revenue management class run by Zvi Schwartz, professor of hospitality business management. This year, he started a new global competition called the Forecaster Cup, which welcomed teams of hospitality students from around the world to match wits using hotel data to predict occupancy in pursuit of a $1,000 prize.

UD made itself proud in the inaugural contests in April and May. One UD team finished third among more than 50 teams from 11 universities.

A team from Vienna, Austria, placed first among competitors from far-flung locations like Vietnam, Spain, Malaysia and Australia. In overall university rankings, UD finished second.

Over seven weeks, the teams analyzed information from STR, a large travel research and data company that tracks numbers and trends, and co-sponsored the competition. The goal was to make the best predictions on how many hotel rooms would be sold in Nashville, Tennessee, in the week ahead. Schwartz picked it because of its importance as a travel and event destination, which can lead to fluctuations in demand.

Figuring out how many travelers are coming to town might not seem like very glamorous work, but millions of dollars can be at stake.

Businesses use forecasting to adjust their prices based on demand - a practice known as dynamic pricing. It's why your plane ticket can cost quite a bit more or less, depending on when you book it.

"Multiply it by thousands of seats or tens of thousands of hotel rooms sold every day," Schwartz said. "That's why it's so critical. It's the same for car rental companies, cruise lines and other sectors."

Numbers don't lie, but forecasting requires a knack for educated guesswork. As the contest's choice of Nashville illustrates, conventions, concerts and sporting events can all mess with travel plans. Extreme weather and political unrest, such as protests, can also upend trends.

"Reality is complicated and messy, and it refuses to behave the way we think it should," Schwartz said. In short, forecasting is hard, and it requires skills and training.

At the same time, the hospitality industry tends to draw those who like relating to others more than those with an analytical bent, Schwartz said. However, skills in economics, math, and statistics, coupled with knowledge of consumer behavior, are necessary for hospitality revenue management, and forecasting is critical.

Going into his senior year in hospitality business management, Matthew Merrill was one of the participants on the third-place UD team. Merrill plans to work as an assistant manager at DiFebo's Restaurants at the beaches this summer. He hasn't refined his career goals just yet, but although he enjoyed the competition, he doesn't see himself as the analytical type and doubts his future will lie in demand forecasting.

Still, "If I need to run a sale or something if I'm a restaurant manager, that [skillset] can definitely help, and I can understand the numbers better," he said.

"Without this kind of competition, they really don't have the real-life experience," said Steve Hood, senior vice president of research at STR, who works with educational institutions. "Doing this for seven weeks makes them understand the difficulty of it and begins to teach them life lessons."

Schwartz, who worked in the hotel industry before he entered academia, has been a leader among scholars studying revenue management for some time. He founded the RevME (Revenue Management Education) group, a global association of professors. This network and connections through STR helped him get the word out about this contest.

"These [the members of RevME] are the top revenue management professors in hotel schools around the world," Hood said, with more than 100 members.

"What has happened over time is the hotel schools, especially the higher level ones like the University of Delaware, have realized that revenue management education should be a priority," Hood said. Forecasting is a key part of that.

Techniques include using exponential smoothing and moving averages to make estimates and making subjective predictions about the impact of events.

Schwartz said that UD has a unique doctoral program in hospitality analytics, and the industry has noticed Lerner's focus on the field. For example, four master's graduates have ended up working in Hilton's revenue management clusters, an elite group of a few dozen among the world's most advanced in dealing with revenue management, Schwartz added. "The fact that four of our master's students were able to get this lucrative job is really telling a lot about where we were able to go with our level of education."

Schwartz has been running the forecasting competition in his class for some time now, but this year decided to open it up to outside participants.

His forecast on participation, though, was an underestimate. He got many more participants than expected, which was a pleasant surprise.

Schwartz's forecast for the Forecaster Cup for next year is more optimistic, as he expects around double the participants.

"I hear very good feedback from my colleagues that it went very well for them and their students," Schwartz said. "It increased student engagement and created some interesting discussions."