Analyze Customer Calls with Natural Language Processing

by   |   September 12, 2014 2:24 pm   |   0 Comments

What is the value of an opinion?

For the effort marketers make analyzing customer sentiment, a Tweet that perfectly expresses the public’s take on a product is worth hours of patient hunting. A recent study revealed that nearly 90 percent of marketers review customer comments manually.

Despite such hands-on tendencies among marketing pros, sentiment analysis platforms are rising in prominence, as businesses hope to learn what people like and hate about the things they buy. Each quarter, new CRM solutions and applications that aggregate social media posts arrive on the global software market.

But outside of a brick-and-mortar setting, no business communication unit beats with the emotion of human dialogue quite like a call center. Seeing customer calls as a source of cogent, subtle, and sometimes no-so-subtle sentiment, a Virginia firm is marketing a technology that transcribes customer calls and analyzes them for opinion—as emotional, frustrated, and inflected with feeling as the opinion might be.

Related Stories

Natural-Language Processing Evolves into Vital Technology.
Read the story »

Siri Points the Way to Analytics-Driven Enterprise Assistants.
Read the story »

Take a Data-Centric Approach to Managing Your Unstructured Data.
Read the story »

Developing a Legal Risk Model for Big Volumes of Unstructured Data.
Read the story »

Clarabridge Speech uses Natural Language Processing (NLP) to transcribe and automatically analyze customer call recordings, including service calls, phone-based market research, or after-call surveys for a 360-degree view of the customer. The product is a result of Clarabridge’s partnership with speech-recognition technology provider Voci.

The product adds advanced speech analytics capabilities to Clarabridge’s customer analytics portfolio. Businesses can harness customer feedback from several sources, including multiple survey types, contact center agent notes, social media, chat, voice, and email. By listening to every channel, Clarabridge aims to deliver nuanced customer feedback with depth and accuracy of meaning.

As explained by Sid Banerjee, Clarabridge CEO, most solutions connect to data from a mass of sources: market research systems, CRM and customer interaction technology, and social media platforms. Clarabridge technologies apply classification, categorization, sentiment scoring, and analytics to answer the evergreen questions of consumer relations. “What type of policies drive loyalty?” said Banerjee. “What functions are complicated? What services may be driving away customers, or are non-compliant, or are not competitive?”

The solution is live with a major telecommunications company and a large retail banking firm, Banerjee said.

“The feedback we were getting from call centers were live chats, agent notes from the call—but what we didn’t have was the actual conversation. When you don’t have that, you don’t get the full voice.

“There’s a narrative arc to that conversation,” Banerjee continued. “If the customer is using emotional trigger words, if they are frustrated, you can pick that up in the transcript. You get more color with that.”

Yelp, Twitter, and the manifold social media platforms have taken their place in the marketer’s portfolio of customer insight tools next to internal feedback mechanisms. Together, these mouthpieces raise a clamor of sentiment, from giddy to hostile, on the efficacy of modern products and services.

Without a technology platform applying consistent measures to make sense of data from many different sources, argues Banerjee, a lot of transactional platforms fail to reflect the shades of human opinion.

The use of such feedback is dictated by scoring mechanics, which can be inconsistent between systems that consume structured or unstructured data. Banerjee explains: “If I say something like, ‘I went to the Starbucks, and my coffee was great, but the barista was rude,’ the majority of vendors will score that interaction as neutral”—a balance between one positive and one negative response. Yet the example includes two separate observations, each equally valuable to a business seeking information about the quality of its staff and products.

Along with highlighting staff or training issues, product problems, or confusing policies, the system delivers data to the appropriate people.

“Feedback exists everywhere,” said Banerjee. “If I’m in marketing, and anyone gives me feedback, no matter the source, it’s bubbled up into the right reports and sent to the right people.”

Joshua Whitney Allen has been writing for fifteen years. He has contributed articles on technology, human rights, politics, environmental affairs, and society to several publications throughout the United States.

Subscribe to Data Informed
for the latest information and news on big data and analytics for the enterprise.

Tags: , , , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>