Sentiment Analysis in CX is calculated using IBM Watson Natural Language Processing in aggregation with NPS. We have designed a uselful yet easy way to provide NPS driven sentiment analysis on customer comments. This is how Sentiment Analysis widget in CX looks like:
To go into details, these are the 2 major components in sentiment analysis:
Keywords: Keywords are the relevant words identified from the customer's comment by Natural language processing. These keywords are highlighted with the comment in the Sentiment Analysis Widget.
Lets take an example: This is a comment from a customer who gave feedback for a restaurant - Loved the ambience of the restaurant. We would like to visit again.
The NPS is an indicator of the overall sentiment of a customer. Hence, the NPS given by a customer is displayed along with his comment in the widget, highlighted in a color representing the sentiment of the customer where Red is Negative, Yellow is Neutral and Green is positive
The bubbles in the widget represent the 10 most talked topic in all the comments. The color of the bubble represents the overall sentiment of that topic which is determined by the average NPS of all the customers who talked about the topic in their comment.
For more details, check out Sentiment and Root cause analysis