Flight Centre taps more data in customer sentiment analysis

Flight Centre Travel Group is broadening the range of data sources that feed into a central voice of customer function, giving it more precise and well-rounded insights into customer sentiment and experience.

Speaking to the iTnews Podcast, global voice of the customer program specialist Ashley Taylor described his two-and-a-bit years at the company - setting up a dedicated voice of customer function and supporting technology.

The company had voice of customer capabilities in its operations prior to Taylor joining, but they had been set up independently of one another, recorded metrics in different ways, and were overseen by different functional areas.

“In some places it was run by our customer care functions, or it was coming from digital. In other countries, it was coming from marketing,” Taylor said.

“When you don't have a dedicated function like we do now, the focus isn't necessarily there 100 percent of the time.”

Emerging from the pandemic drove Flight Centre to globalise some of its functional capabilities, including voice of customer.

“We wanted to look after our customers,” Taylor said. “We wanted them to return to us based on having a good experience with us.”

Upon joining, Taylor worked to assess the varying capabilities in place across the group’s operations, and to set a new target endstate.

A request for proposals (RFP) process undertaken by the company led it to establish Qualtrics as the core technology underpinning voice of customer capabilities globally.

Qualtrics was already serving such a purpose in the group’s corporate travel businesses, courtesy of an earlier deployment.

More recently, Taylor has started using natural language processing functionality in the Qualtrics platform that came from the vendor’s acquisition of Clarabridge in 2021.

This is helping Flight Centre to “categorise and properly understand what customers are talking about” across a “very high volume of comments and feedback”.

While some of that feedback comes directly from Qualtrics-powered surveys that are put in front of customers, the group has this year expanded its use of the technology to analyse both “solicited and unsolicited feedback”, as well as structured and unstructured data.

“We are having our Qualtrics surveys flow into the natural language understanding,” Taylor said.

“We're also integrating Sprinklr, which captures all our social media commentary feedback on their review site comments and pulls that in as well and then applies things like sentiment and emotion to the language so that we can categorise and understand what all these customers are talking to us about.

“We've just implemented Genesys, so I'd like our telephony to be able to flow through into a centralised location, as well as operational data from Power BI so that we've got a much more rounded picture of what our customers are talking about and we’re able to really use the natural language understanding to categorise that with good precision and good recall and then drive inner and outer loop activities, and more strategic and process-related improvements.”

Taylor highlighted a recent sentiment analysis exercise that identified issues with some customers not receiving extra legroom seats they had paid for in their bookings.

“On some specific airlines, the seating wasn't being allocated properly,” Taylor said.

“If I've got tens of customers a month saying this, we know safely based on response rate and what we're not capturing this is probably affecting hundreds of customers a month … and thousands of customers every year.

“If we consider their lifetime value, this is a ‘millions of dollars over years’ problem, and it's a bad customer experience when they're buying something and not getting what they purchased.

“We were able to fix this problem, and we saw a 20 percent drop in these types of comments within the space of two weeks.

“We're continuing to work on our processes to improve this, and that's the benefit of having people in the business that are getting this information or are hungry to do something with it, and the information being useful and allowing them to do that.”

Taylor said that insights from the voice of customer program may also be used in the future to power ‘next best actions’ or ‘next best conversations’ to have with customers, although this would require some additional development work.