Lisa Fairbanks
Share
  •  
  •  
  •  
  •  

Customer Experience (CX) has quickly become a strategic priority for organizations across industries. In fact, Gartner predicts that “50 percent of consumer product investments will be redirected to customer experience innovations by 2017.” But assessing the return on those redirected investments will require a shift in how organizations measure CX.

Take social media as an example, which has played an enormous role in shaping today’s CX landscape. One of the most commonly referenced metrics in social media monitoring is sentiment, which is used to determine the emotional state of the author. Unfortunately, when we look at sentiment thru the lens of CX, we find that it’s an old metric that no longer fits the new model.

As background, sentiment is often represented in three ways— negative, neutral or positive. In some cases, those dimensions are presented on a sliding scale, for example, very negative, somewhat negative, neutral, etc. – to determine the degree of the author’s opinion. To this end, we find sentiment analysis to be one of the core measurements for standard industry reports in all of the major social listening platforms, including Salesforce Social Studio, Sprinklr and Simplify360.

Sentiment versus Expectations: Improving your CX analysis

Social Studio Dashboard

Traditional sentiment analytics does offer value. You can use it to determine the public’s response to your brand. You can find out if they are talking about you, and if their comments are positive or negative and, if negative, how negative and why. But as we discussed in the article Fourth Listening Post in TCS Perspectives Volume 6: The Responsive Enterprise, traditional sentiment analysis should be extended beyond just positive, negative and neutral word tagging to better account for context.

Consider the following examples:

  • Customer tweet #1: Does @Tmobile have ANY idea how frustrating that my iPhone 6 doesn’t work? Ever? In their HQ city? #NoSignal.”
  • Customer tweet #2: “@ATT do you have a tower out? I’m not getting a signal?”  

The first customer tweet will be categorized as negative, since the word frustrating will trigger sentiment. With the absence of the word frustrating in customer tweet #2, there’s no term indicating the user’s tone. Following traditional sentiment analysis techniques, the absence of a term to indicate tone means that the tweet would be classified as neutral. However, when you consider the context of the sentence, we know that a missing signal is a poor experience for a wireless customer.

Customer Expectation Score

Creating a customer expectation score is one approach to overcoming the missing context issue found in sentiment analysis because it incorporates if the organization/product/service has met, missed or exceeded the customer’s expectations by assessing the context of an interaction or transaction. To execute such an approach the analyst creates several linguist libraries of terms which account for both the tone of the sentence and the expectation of the interaction.

Consider the following examples which illustrate customer expectation scores within the context of a specific CX interaction with a mobile app.

  • “My mobile app won’t open.” This will be categorized as a missed customer expectation score because the user expected the app to be operating properly.
  • “It only took five minutes to set up my account in the app.”  This comment is considered an exceeded customer expectation score because the customer was likely expecting the process to take more than five minutes.
  • “Why don’t you offer a mobile app for the BlackBerry?” This indicates a missed customer expectation score because the organization/product/service hasn’t developed an app for all devices used by customers.

The chart below shows an example of measuring customer expectations versus sentiment in the banking industry using TCS’ Customer Experience Insight tool.

Sentiment versus Expectations: Improving your CX analysis

In the journey to improved CX, organizations must gain a complete view of the existing experience to understand where to make improvements. Relying upon traditional analytical approaches, such as using sentiment analysis, may not provide the complete picture. Please share your experiences or questions with us here and learn more about measuring customer expectations by visiting TCS experts at the Salesforce World Tour London on May 21.


Share
  •  
  •  
  •  
  •