Gaylord Hotels Taps Text Mining to Boost Guest Satisfaction
How can you quickly make sense of text-based comments on thousands of customer surveys? Text analytics helped this mega-hotelier uncover hidden obstacles to customer loyalty.
If the deployment at Gaylord Hotels is any indication, text analytics has emerged from the rarified domain of military intelligence and life sciences research and is gaining mainstream business use. With 1,400- and 1,500-room hotels in Orlando and Dallas and its well-known, 2,800-room mega hotel at the Grand Ole Opry in Nashville, Gaylord needed a better, faster way to "read" customer satisfaction surveys. Within six months (from pilot to production rollout), Gaylord replaced a slow, third-party service with an in-house analytics app that delivers overnight results through guest satisfaction dashboards and ad hoc reports.
Gaylord Opryland Hotel
Gaylord's pursuit of text mining began in late 2006 when executive Tony Bodoh investigated analytics as a way to scan thousand of RFPs and contracts to get a better sense of what meeting planners were looking for. When Bodoh was promoted to Manager of Operations Analysis in April 2007, he pushed for an even broader deployment.
"We were evaluating text analytics vendors when I took over guest satisfaction, and I quickly realized that the technology presented a big opportunity," he explains."By late June we signed a contract with Clarbridge to help us with a pilot project to examine two-and-a-half years' worth of guest satisfaction comments."
Some 90,000 hotel guests had completed the post-visit, Web-based poll since 2005, but Gaylord had only a primitive sense of the verbatim comments. The goal of the pilot project, which began in July and was completed by late September, was to use Clarabridge's content mining platform to automatically detect the sentiment of free-form comments and correlate them with structured responses on customer satisfaction.
"We previously had a third-party firm reading and categorizing comments about problems and keeping track with tick sheets, but we focused only on which categories received the most comments," Bodoh explains. "With text analytics, we were able to tie comments to structured one-to-five satisfaction rankings, and we quickly discovered that our biggest problems were not necessarily related to the most common compliants."
As an example, comments associated with "noise in the guest rooms" weren't the most common problem, "but we didn't realize that was so tied to 'wouldn't return or wouldn't recommend' from a loyalty perspective," says Bodoh.
We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.