How Sentiment Analysis Can Improve the Customer Experience

Risk Management, Sentiment Analysis, Social Media Analysis

Sentiment Analysis for Customer Experience Management

Customer Sentiment Goes into a Downward Spiral

It’s an all too common customer experience. Bob is not happy at all. He called his device carrier’s customer service line and got disconnected on the first couple of tries. When he finally reached someone, the call-center person tells him it’s not his job and says to send a complaint via email. The call-center person says Bob will hear back within 24 hours.

After four days and not hearing, Bob is majorly unhappy. He’s seriously considering switching carriers. What’s worse for the carrier is that he let all his friends know about how bad his experience had been and he posted a detailed account on the carrier’s Facebook page.

How did things get so bad? And what can the carrier do about it?

Sentiment Analysis Is Critical for Customer Care

The ways in which customers interact with companies have become extremely complex in the last couple of decades. Customers can post opinions and recount their experiences in the many channels that the Internet offers, in particular on social media. Companies have many more ways to ascertain the customer’s experience than were available in the days when their main tools were surveys and focus groups. They are able to develop far more accurate and detailed pictures of their customers. The downside is that one unhappy review of any aspect at all has the potential to go viral and hurt a company’s reputation substantially in no time.

This is where a new technology, Sentiment Analysis, comes into play. It uses the technologies of Text Analytics and Natural Language Processing to analyze a person’s attitude to something and to assign “positive” and “negative” values to it. It also assesses how strongly the customer expresses the sentiment. Sentiment Analysis does this with high accuracy and, just as important, it can handle Big Data-sized quantities of information.

Sentiment Analysis Enables Customer Experience Management

Sentiment Analysis also enormously improves Customer Experience Management (CEM). The latter focuses in detail on the customer’s experience of their interaction with a company.  It aims to understand all aspects of the customer’s journey through the various stages in the sales and support process. CEM aims to answer questions like the following: Did the customer have a pleasant experience interacting with a company, e.g., in placing an order through its web site? CEM wants a better view of each individual customer – what apps they use to communicate with the company, what opinions they have about the company’s web site, what their opinion is regarding the products’ strong points and drawbacks, and anything else important to the customer.  The ultimate goal of CEM is to reduce the number of Bobs having unhappy experiences and to maximize good experiences.

How Sentiment Analysis Works

As part of a company’s CEM process, a company could use a Sentiment Analysis product to analyze all incoming emails or social media posts and determine the degree of happiness or unhappiness contained in them.  All communications could be ranked by their level of anger to prioritize those who require a quick human response to contain any potential damage.

More advanced Sentiment Analysis can even zero in on what precisely the customer is unhappy about. Sentiment Analysis doesn’t just identify customers’ attitudes as a whole to a brand, it can focus in on what their opinion is on specific features of a brand’s product, e.g.,

“I like my new water heater, but the installation team wasn’t very professional. They were late and took way too long to install.”

Here Sentiment Analysis can identify the overall positive attitude to the product, but also capture the more negative sentiments expressed about the installation. This information, when aggregated over a very large number of communications, offers a company a gold mine of information of how they and their products are doing and points to directions on what to improve and how to improve it.