Text analytics software packages can be very beneficial in terms of helping companies identify, understand and
respond to customer needs, but they’re no substitute for verbal communication and good old-fashioned face time, according to one speaker at the Text Analytics Summit 2010 in Boston last week.
More on text analytics
See how NASA is using text analytics to improve aviation safety
Learn how companies are bringing meaning to satellite data
Find out more about the value of social media analytics
Chris Jones, manager of analytics at Intuit Inc. -- the Mountain View, Calif.-based creator of popular accounting applications like Quicken and TurboTax -- has worked extensively with text analytics tools for more than three years. Speaking to a roomful of Text Analytics Summit attendees, Jones said the experience has taught him some highly valuable, and occasionally bittersweet, lessons. Chief among them is the idea that text analytics software should serve only as a complement to frequent human interaction with customers.
It’s a message that was driven home for Intuit on April 15th -- the annual deadline for filing tax returns -- when Jones and a large number of his colleagues from various departments gathered to take customer service calls from TurboTax users who were scrambling to get their taxes submitted on time.
“When you’re actually on the phone talking to a customer with that April 15 deadline hitting and they’re trying to file their taxes to pay the rent, it’s very emotional and it’s very powerful,” Jones told conference attendees. “So, I caution that while [text analytics] tools are great and they allow us to tap into loads of data, don’t use them as an excuse not to reach out to your customers and talk to them, interview them and get to understand their problems.”
Text analytics -- also known as text mining or text data mining -- helps organizations identify and understand what people think about them, their products and their services. The software, which combines advanced search techniques with speech-pattern-recognition capabilities, allows users to quickly scour mountains of unstructured or free-form text in the hope of identifying trends in customer sentiment that will ideally lead to positive change. That unstructured text might reside within an organization’s email files, in online user communities, or on social networking sites like Facebook and Twitter.
Text analytics tools from Reston, Va.-based Clarabridge help Intuit hear the “voice of the customer,” Jones said, by allowing the accounting software giant to more quickly pore over and analyze call center data, comments in feedback forums, and the results of myriad online customer surveys. Other vendors vying for position in the growing market for text analytics software include Autonomy Corp., Endeca Technologies Inc., SAP BusinessObjects, SAS Institute Inc., and IBM, which gained text analytics tools primarily through its 2009 acquisition of predictive analytics pioneer SPSS.
The global market for text analytics software licenses, vendor-provided support, and professional services grew about 25% from $350 million in 2008 to $425 million in 2009, according to Seth Grimes, founder and analytics strategist with Alta Plana Corp., a Washington, D.C.-based business analytics research and consulting firm. Grimes, who delivered the opening comments at the Text Analytics Summit, estimates that in 2007, the global market for text analytics software and services checked in at about $250 million.
"Two factors have driven the steep global growth in text analytics uptake. First, organizations now understand that they need to go beyond databases and numbers to get at critical business information,” Grimes said in an email interview after the conference. “And secondly, the technology has matured, as have providers and solutions."
Text analytics and the gift of feedback
It takes more than fancy software to achieve the full benefits of text analytics. Companies interested in the technology first need to create a culture that both thrives and acts on customer feedback, Jones said, whether that criticism is positive, negative or somewhere in-between. Firms that can’t handle this essential first task will undoubtedly find text analytics to be a major waste of money, he added.
“If there is one thing that I want you to take away from this presentation, it’s that feedback is a gift,” Jones said. “If this is not something you can convince your organization to get behind, then I really don’t think you should be doing text analytics.”
In one instance, Jones told conference attendees, the gift of feedback and text analytics allowed Intuit to help customers help themselves.
After using text analytics software to make sense of loads of data garnered from online customer surveys and other feedback, Jones and his team uncovered a disturbing trend. They found that a significant number of customers were having trouble finding the Frequently Asked Questions (FAQs) page on one of Intuit’s websites – a section which Jones admits was “buried” about five levels deep. As a result of this problem, many of those users would end up dialing into Intuit’s call centers.
The discovery prompted Intuit to redesign its website and place a prominent link to the FAQs section right on its homepage. Jones said this simple move ultimately increased customer self-service by a whopping 40%.
“By listening to why customers couldn’t do something, we said, ‘You know what? We need to rethink [our approach],’” Jones explained. “Text analytics has really helped the company focus.”
In another example, Intuit’s commitment to acting on customer feedback resulted in a decision that was considerably harder for the company to stomach.
About two years ago, Intuit decided to change the pricing policy for its TurboTax Desktop product, Jones explained during another conference session. In previous years, customers could purchase the product once and use it to file up to five tax returns free of charge. But several factors led the company to end that policy and begin charging a fee for each additional tax return filed. Intuit has said that the incorporation of eFile technology into the base price of the product, as well as the company’s plan to give away some free copies of the software, meant that only a small percentage of customers would actually end up paying more. But the general public simply didn’t take kindly to the news.
“That minority of people who were going to be paying more for the product were extremely vocal on Amazon, and we went from having [a customer feedback rating of] four to four-and-a-half stars on Amazon to one [star],” Jones said. “Even the people that weren’t impacted jumped on and agreed.”
Despite the fact that Intuit felt the price increase was just, the company’s commitment to creating positive customer sentiment won out, and the software giant ultimately decided to undo the new pricing policy.
“Essentially, we gave away $150 million worth of the product to the point that we had to restate our earnings,” Jones said. “This is about doing the right thing, it’s about building loyalty and it’s about how we show up when things go wrong. We could have sat proudly and said that we still think the pricing is fair, but [instead] we took a big hit.”
With text analytics, a little beer goes a long way
When embarking on a text analytics initiative, it’s also important for IT pros to create strong partnerships with the business users in their company, Jones said. After all, he said, it is those business users -- people like product managers, engineers, marketing associates and high-level executives -- who will use the information gleaned from text analytics to generate revenue.
One way to enhance this partnership and strengthen the company’s commitment to customer feedback is to create a regular “voice of the customer day,” when business users and IT folks get together to discuss trends that are emerging within the customer community.
“IT people aren’t always scary. Take them out for a beer or a cocktail and you’ll get a lot more done that way,” Jones said. “[Text analytics] should definitely become a company-wide initiative.”