It took Airbnb no more than five years to disrupt the hospitality industry. An online marketplace, it connects guests looking for short-term vacation rentals to homeowners wishing to rent their spaces. The company’s earnings come through 12 percent reservation fees to guests and 3 percent service fees to hosts.
By 2015, Airbnb had connected more than 20 million guests with more than 800,000 hosts worldwide, and its business model even appeared to have influenced that of a major competitor. How could it keep ahead of similar companies? It had found success helping hosts learn how to best showcase their homes; perhaps it could help guests better find what they were looking for.
One possibility would be to use text mining, a technique that could translate website comments into helpful data. In text mining, software uses algorithms to assign values to positive and negative words and patterns, thereby quantifying consumer preferences.
Researching public information on Airbnb, insights gained from text mining could result in changes similar to those Airbnb has implemented, such as promoting the best hosts, suggesting price points to them, and highlighting different features in different geographies — in one destination, price may matter more to renters, and in others, reviews. Sentiment analysis through data text mining can help to customize a user experience and take the guesswork out of a company’s marketing and strategy.
The preceding is adapted from Darden Professor Rajkumar Venkatesan’s article “By Scraping Data, Airbnb Could Scrap Guesswork,” which appeared in the 29 January 2017 issue of The Washington Post as part of the Darden School of Business/Washington Post “Case in Point” series.
The article is based on the case Have Text, Will Travel: Can Airbnb Use Review Text Data to Optimize Profits? (Darden Business Publishing), by Shea Gibbs and Rajkumar Venkatesan.