NORWALK, Conn., Sept. 18, 2013 – It’s one thing to detect a customer’s tweet and figure out (accurately) if the person is happy or frustrated - but then what do you do with that information?
While social media listening tools on the market today make it easier to monitor and analyze the millions of conversations happening across Facebook, Twitter, YouTube, LinkedIn, blogs and online communities - using the information to make sales or improve customer satisfaction remains a very hands-on, time consuming process.
To help, Xerox (NYSE: XRX) researchers are working on an automated data analytics platform that teaches computers to more accurately determine the sentiment of comments (pleased, angry, confused), and quickly route that information to the right person or team. The automation helps organizations respond faster to customer data, and in a much more relevant way.
“When humans have to step in and evaluate the context of a tweet or route a post, it slows things down and reduces the overall value of social media data,” said Tong Sun, who leads the data analytics laboratory at Xerox Research Center Webster in New York. “We’re piloting a platform that lets computers do the heavy lifting.”
The advances made in automation and accuracy could help a customer care agent address an issue before it becomes a crisis or allow a sales executive to make the most of a real-time event in the market. In a recent pilot, businesses were able to respond to comments in hours instead of days or weeks.
“Studies have shown responding quickly to an issue raised on social media channels leads to happier customers,” Tong said. “And we know that targeting the right customers and communities makes for a more profitable marketing campaign.”
How it works
Many of the social media monitoring tools on the market today use a simple keyword-centric approach for determining the sentiment of an online comment, making it difficult to detect things like sarcasm or abbreviated wording. To understand the context accurately, Xerox researchers with expertise in text mining, machine learning and predictive modeling created an analytical platform that quickly and accurately extracts the sentiment by reviewing it in context of ongoing conversations. The platform is then able to assess, prioritize and deliver the insights to the best contact (a customer sales agent or a sales or marketing team), so they can address issues or opportunities.
Already piloted with several customers, the technology is available now as part of Xerox Customer Care offerings and also will be offered to sales and marketing organizations across many industries including financial services, telecommunications and retail.