Sentiments and emotions are the two sides of the same coin. These are those cognitive characteristics of a person that govern his decision making process, and the concept of sentimental analysis is designed to harness the same.

Sentiment analysis or opinion mining, a branch of speech analytics, is a way of processing, interpreting and analyzing the indented sentiments behind the responses of users about an organization, business or institution on social platforms within the internet space. This is done through assigning of a metric to a piece of text which provides a detailed blueprint of how positive, negative or neutral a given text or response is.

Today, 92% of marketing professionals think that social media has deep impact on their business.

As of June of 2017, Facebook achieved a monumental growth with its two billion users, while Instagram has reached 700 million users and Twitter about 317 million.

Former U.S President Barack Obama used this technique to gauge public opinion for policy announcements and campaign messages ahead of his 2012 presidential election.

There is a common proverb in business parlance that: Customer is the King. Satisfaction of customers is of utmost importance for the survival of any business organization.

Sentiment analysis can be beneficial for a business in broadly two ways:

• It can help a business to keep a track of what people think of its products and services
• A business can also know about the opinion of people regarding its competitors through sentimental analysis

Customer sentiment analysis is an algorithm based interpretation of complex factors which includes measuring of some of the prominent indicators that helps to determine a person’s stress level such as:

• The amount of frustration in a customer’s voice
• Rate of speech
• Close examination of the intonation of a person’s speech caused due to tension or any other stress element

Sentiment analysis- A boon for customer care services

The deadly combination of sentiment analysis and machine learning can do wonders. The use of technologies like NLP (Natural Language Processing) that helps the machine in understanding the text and identifying specific keywords has become a common phenomenon in sentiment analysis these days.

Sentimental analysis along with artificial intelligence can:

• Help your business track the audience’s response in real time
• Helps your business to identify with latest and current trends, and
• Segregate the results.

Machine learning has a very pivotal role to play in the process of sentiment analysis. It is here to improve and automate the low-level text analytics functions that sentiment analysis depends upon, including Part of Speech tagging.

A Customer’s sentiment isn’t often sufficient on its own for acquiring the required feedback because it does not describe why that customer feels about a particular product or service, in a way he feels. So how can a business find this out?

Business should focus on two main factors:

• The aspects that the customer is commenting on. For example: In a restaurant, a customer may like the food or the live music.
• Specific recurrent themes. For example: ‘Food service is snail paced’ or ‘home delivery of food is not on time’ etc.

One of the key roles of machine learning in case of customer care is to help the data aggregators to identify double meaning words, phrases or sentences. For example: Suppose a customer wrote: ‘Not Bad…’ about a home appliance product on an online shopping website.

Now what does that mean? Is the product just good or does it needs to be better? This requires a lot of brainstorming on the part of data aggregators, and machine learning assist in the same. Language is evolving, and so is technology. In order to keep with the pace on latest techniques like LSTMs (long short-term memory) which are units of a recurrent neural network (RNN) is used, where a whole sentence is made compact into a single vector (a list of numbers) retaining the actual meaning taking the word count into consideration while deciphering it.

Sentiment analysis tools that are automated aid business analysts in providing useful and detailed information on how the customers feel about a particular product or a brand. This is done through analysation of tweets, online reviews and news articles etc. Thus, such automation facilitates the media managers of a company to convey the issue and complaints, if any, of the customers to the top most management, enabling them to make informed decisions.

Online British supermarket Ocado became overloaded with large amount of customer support tickets. After which the company decided to use sentiment analysis for organizing the tickets and respond to them more quickly. They used a dataset built from thousands of customer support tickets and by using a trained machine learning algorithm to tag tickets as positive, negative and neutral to help the customer support team decide how to prioritize them.**

At the 2017 Google Cloud Next conference in San Francisco, as a part of a case study Dan Nelson, the head of data for Ocado explained how his company has implemented machine learning. He said that Ocado gets roughly 2,000 emails into its contact centers a day, ranging from refund requests, to general feedback, to website trouble, and more.***

Talking about the Natural Language API, Nelson said they were able to label message with tags such as “Feedback” or “Positive.” However, they then built a custom solution with Cloud Machine Learning and Tensor Flow to get more detailed filtering. It sits outside of their storage layer and allows the company to more effectively triages their customer service requests, Nelson said.

As a result, Ocado was able to respond to urgent emails four times faster, and saved money on headcount in the contact center. He recommended that companies investigating machine learning would define their success criteria early and be careful to set control groups and perform a lot of testing.

Nick Martin, coordinator of social engagement at HootSuite said: “Sentiment analysis is an effective tactic for blogger outreach, customer service, and support representatives as they can react efficiently and monitor dissatisfaction before negative sentiments from consumers spread”.

These pointers can prove really valuable to a business in devising a better course of action to meet the expectation of the customers effectively.

References:
* Understanding sentiment analysis
** Sentiment analysis saves time
*** Machine learning to improve customer service