How does a business find out what customers think of it? Normally it would run a market research exercise with a set of multiple-choice questions which enable people to rate the company. This produces data that is great for structured data analysis, using traditional business intelligence (BI) tools, to gain an insight into customers' views.
However, more often than not, the most interesting customer feedback appears in free text form, which is difficult to analyse using traditional BI.
Using natural language processing sentiment analysis allows businesses to look at the information customers type into text fields, or the remarks they make on social media sites. For instance, an article on sister site SearchDataManagement.co.uk describes how hotel chain Marriot was able to identify which high-profile bloggers regularly posted positively about its brand.
"We have a voice as a company, but we also found that others were talking about Marriott," said Mike Keppler, senior vice-president at Marriott International.
Clarabridge opens UK office
Clarabridge is a US provider of sentiment and text analytics software for customer experience management which is expanding into Europe with the opening of an office in the UK. Customers include Gap, Best Buy, Walmart and Amex. "We have international customers looking to use software outside the US," said Sid Banerjee, CEO at Clarabridge.
He says the privately-owned company experienced high growth. Previously, IT investment in analytics software focused on transactional systems such as customer relationship management (CRM), and were used to track structured information. "We are now seeing investment in tracking unstructured information, such as the feedback from surveys and social media sites which contain 99% unstructured data."
According to Banerjee, this unstructured data contains very rich information about what customers need or like. Such information can be critical for customer retention, especially if the business is looking to cut back products and services.
Case study: United Airlines
United Airlines has been using the software for five years. Matt Hadfield, senior manager of customer metric at United Airlines, says the company began using Clarabridge when it started running online surveys.
He says that as it received more feedback in freeform text boxes, the information became more difficult to analyse: "As we grew our survey volume we could no longer manually analyse the text fields."
One could argue that a well-structured survey should provide enough insight from the multiple-choice structured question fields. But in many cases the freeform text fields provide the reason for the structured responses. "It gives us a second level of detail and answers the question 'Why?'."
By analysing the unstructured data using Clarabridge, United Airlines started putting a strong emphasis on customer experience management. The survey results helps United Airlines drive business decisions and product development.
"It allows us to communicate information to our stakeholders in a meaningful way," Hadfield says.
This is particularly relevant as United Airlines and Continental Airlines are undergoing a merger. "Products are being kept at the new company, based on the customer comments. If we do not include these products, we would definitely lose customers," he says.
Harness the potential of social networking
Analyst Gartner urges CIOs to begin exploiting the potential of social networking technology this year, or risk losing the initiative to other parts of the business.
While hugely popular, social media marketing is difficult to measure. Sentiment analysis may be the way forward for IT to gain a foothold in social medial by providing tools to analyse what customers really think about what the company is doing.
Clarabridge version 4.5 - Tower
Clarabridge has introduced version 4.5, the "Tower" release of its Enterprise and Professional editions. The release includes embedded connections to several social media sources and French and Portuguese natural language programming (NLP).
The Tower release includes an API, which provides real-time access to multiple sections of text (verbatims), and new security and administration capabilities. In addition, it offers a number of new reports and dashboards.