Using predictive analytics in the hospitality industry is in the early stages, but is primed to soon take off. Airlines, hotels, restaurants and casinos are beginning to realize that they can improve their margins while taking customer satisfaction into the stratosphere.
The key is in in pairing real-time, location-based condition monitoring with predictive analytics models. If you can gauge where and when an issue may arise, such as a cancelled flight or a hotel room no-show, you can deal with the situation to make customers feel like kings and capitalize on it.
Take flight cancellations, for example. If an airline used predictive analytics, as part of a predictive maintenance program, to determine when a certain plane might need maintenance, it might prevent a flight cancellation and subsequent passenger issues. Because airlines have a policy of overbooking flights in order to fill the seats of no-shows, any flight cancellations can seriously disrupt travel.
I was recently on a 5:30am flight from Denver to Chicago which was already overbooked, and the issue was exacerbated even more when the 8:00 am flight was cancelled. The airline offered a US$700 travel voucher for anyone who would take a different flight (I was tempted, but the next flight with seats available was at 6:00pm!) The result was disgruntled passengers whose flight had been cancelled (and they could not get another morning flight because all of the flights were overbooked to begin with), plus an expensive scrambling exercise for the airline.
The kicker here was that the 8:00am flight had been cancelled the day before. Let’s say the airline had been using predictive analytics for maintenance; it would have been able to predict with a great deal of certainty that the plane scheduled for the 8:00am flight would need maintenance days or weeks before then. It could then have pulled it off the rotation for maintenance during downtime, or replaced it with another plane.
There are other ways an airline could use predictive analytics, such as accurately estimating how many no-shows there will be for a given flight. Monitoring for conditions such as icy roads, traffic accidents, or how much traffic there is heading for the airport would underlie predictive models to tell them, with a great deal of certainty, that X% of passengers will not show. Then the airline could proactively offer empty seats at a discount, via text alerts, to loyalty customers. This could be a new, more reliable form of “standby” that pleases your frequent flyers and limits overbooking.
Hotels would benefit in a similar way. Using predictive analytics, again coupled with condition monitoring, a hotel could more accurately estimate how many empty rooms it will have by a certain time each day. Weather, traffic, big events in another town, etc could signal slack business or no-shows. Then, it could reach out to loyalty customers with a discount before dumping the rooms onto an aggregator site and possibly selling them at a loss.
With predictive analytics, the hospitality industry can change the way it deals with negative events; either by preventing them or by having a greater amount of time to prepare a response to an issue which enables the handling of exceptions in such a way that customers feel they have been treated well. Win:win.
Sean Riley is a director of strategic business solutions for Software AG and supports the supply chain practice for the Americas by working with the company’s largest retail and manufacturing customers. His focus areas include value discovery and enablement, process improvement, financial and economic modeling, and collaboration enablement.