BOSTON -- JetBlue for years built a reputation for customer service and accountability, but a 2007 ice storm that left flyers stranded and angry showed the company it needed a better way to measure customer feedback.
JetBlue was one of several companies to present at the Text Analytics Summit being held here this week, detailing how the technology has made a difference in its CRM initiatives.
While the market for text analytics software is still emerging and many companies are still focused on getting a handle on their structured data, a handful of businesses are simultaneously examining their unstructured data. Framingham, Mass.-based IDC predicts that the market for enterprise search technology, which includes text analytics software and applications, will reach $3 billion by 2012, up from a current estimate of $2.25 billion.
Two years ago, JetBlue found it had a wealth of unstructured information that needed analysis. The company relies heavily on net promoter scores -- the likelihood of a customer to recommend a service or product to others. As a result, JetBlue gathers customer feedback in a number of ways. It sends out surveys to 35 customers on each of its flight (taking special care to ensure that no one receives a survey more than once every 90 days). It consults with VisionCritical, its select customer panel, and with JetViews, its online customer community. All of that is paired with text analytics software from Palo Alto, Calif.-based Attensity Corp.
It was the Valentine's Day ice storm of 2007 that brought home the power of text analytics. The storm left JetBlue customers stranded on the runway for hours, leading to nationwide coverage in the press, a sudden backlash and the company's creating a customers' bill of rights.
"This really opened our eyes to the need for text analysis," said Bryan Jeppsen, senior analyst, customer feedback with the Forest Hills, N.J.-based airline. "The day after that storm, we had 15,000 emails. We're staffed to read 400."
With executives pushing for customer feedback and not enough staff to sort through the flood of email, text analytics became a natural experiment, Jeppsen said. JetBlue ran a test implementation and got a report in three days.
"It was eye-opening to see there's technology that can read and categorize huge amounts of data that fast that you could use," Jeppsen said.
With the ice storm, JetBlue knew why there was a sudden drop in its net promoter scores. That's not typically the case. A number of factors can influence that score, ranging from competing fares to on-time percentage.
"We learned that taking text analytics and structured data together is the most powerful way to [measure customer feedback]," Jeppsen said.
For example, JetBlue knows where people who took surveys flew to and from and knows which cities JetBlue members live in and how often they fly. Managers at each of the airports it serves receive their own weekly net promoter score.
Text analytics also allows JetBlue to better focus its efforts. While the company's overall net promoter score is generally high, analysis of customers with the higher scores found that there were still more negative comments than positive.
"They want us to do well, only they're asking us to do something," Jeppsen said. "We found they are the most important people to pay attention to."
Customers are always going to be frustrated by delays and cancellations when flying.
"Some will say there is nothing we can do about that, but that's not true," Jeppsen said. "During that ice storm, there were a lot of compliments to our crew."
With that feedback, JetBlue is changing the way it deals with customers faced with delays and cancellations.
Similarly, JetBlue can run reports on which planes have the most complaints about the entertainment systems and can alert ground crews. The company also runs reports on which airports get the most complaints about rude staff (JFK in New York is first, followed closely by Boston's Logan).
Jeppsen conceded, however, that determining ROI has been a challenge. Much of the savings from the program comes from efficiencies, and Jeppsen was reluctant to bring that to the finance department, lest they then ask him how many people he could get rid of.
"We finally just appealed to the fact that we are going to read each of our customers' feedback, and to our executives that was enough," he said.
JetBlue plans to extend its analysis to social networks. The company has established a Twitter account and is working with Attensity to figure out how best to use that data, Jeppsen said. JetBlue also monitors airline-based online communities.