Nearly all of us have stayed in a hotel. And, we probably realize that at peak times of the year, the prices are higher. We also notice the location of a hotel might make the price higher than elsewhere. But, you might not realize that price forecasting can make or break the success of a hotel and it’s something academics have been studying for years.
Jean-Pierre van der Rest, professor of business administration at the Leiden Law School at Leiden University in the Netherlands, is an expert in strategic pricing and revenue management in the hospitality industry. Van der Rest is a visiting scholar to the Fritz Knoebel School of Hospitality Management at the Daniels College of Business and shared his latest research at a talk April 3. In his study, “Forecasting in Hotel Revenue Management: Do Judgmental Forecast Adjustments Improve Forecast Accuracy?” the scholar studies fully automated revenue forecasting systems.
“This is hot off the press,” he joked. Van der Rest is looking into forecasting prices based on behavioral tendencies, rather than on purely numerical figures. He explored several questions, including:
- Are group forecast adjustment more accurate than transient forecasts?
- Are adjustments made closer to the date of arrival more likely to improve forecasting accuracy?
- Are positive adjustments less accurate than negative adjustments?
The answer to these inquiries led to a potential answer to the biggest question of all: Can a computer forecast hotel pricing better than a human can? Two years of data collected by van der Rest and his mentee surprisingly proved that the answer to this query was no. But, as with most situations in business, it really depends. Van der Rest hypothesized that the computer may not have learned certain pricing patterns because humans were overriding the system so often. When looking at data from 1,516 hotels worldwide, 260,000 override decisions were made my managers.
While Van der Rest is an expert in revenue management, he acknowledges that computer forecasting is an emerging field. “I feel very insecure about this data because I can’t see it,” van der Rest said. In other cases, “I can look, I can feel…this is a whole different world.”
It sounds like he and other academics will have plenty to study in this growing field.