Improving CX with text analytics and emotion

Research indicates that behaviours of service employees are critical to customer evaluation of the service quality as well as customer satisfaction.

The levels of satisfaction mostly reflect relationship quality and customer attitudes like loyalty, customer advocacy or switching over to another provider.

These findings also highlight the importance for organisations to understand the emotional responses of their customers to improve service design and/or employee training for better outcomes.

With the industry moving towards shorter questionnaires to improve response rate and engagement, it is vital to leverage open ended feedback to uncover deeper customer insights using text analytics and emotions.

Historically within our text analytics offering we have used sentiment as a proxy for detecting emotions to identify satisfaction/dissatisfaction based on customer comments. We have now applied academic research and decades of text analytics expertise to build an emotions model that captures nine key core set of emotions - Anger, Hate, Displeasure, Anxiety, Trust, Love, Surprise, Happiness and Excitement.

Most often when customers leave feedback it may be difficult to verbalise their emotions clearly. However, our emotions model is simple enough to contextualize variations of several words, sentiments and phrases that customers may use to express how they feel into these nine-core set of emotions.

It is also important to note that unlike quantitative scores that are precise, measuring emotions is not very easy.  If something can’t be measured well it cannot be managed. MaritzCX have extensive experience measuring emotions and studying the impact on customer experience across a variety of industries.

As an example, to study the impact of emotions we analysed the data gathered via a customer satisfaction study by linking emotions, text analytics with customer satisfaction scores. The results highlight that positive emotions are linked to higher NPS scores whilst negative emotions elicit lower recommend scores.

We have further combined the emotions with behavioural science and operational data to effectively integrate emotional measurement into our client’s CX efforts.

By grouping core set of emotions based on their intensity and customer hierarchy of needs we are able to identify clusters of customers with similar behavioral attitudes. These clusters help identify brand advocates, loyalists as well as those that don’t engage with the brand or at a risk of churn. These insights help organisations tailor their efforts to reward loyalty or deploy churn reduction strategies that contribute to long term value and profitability. Select examples from our recent work that we have delivered to the clients in emotions and text analytics:

• Trigger recover efforts during service disruption at airports.
• Rewarding dealers for positive behaviors and sharing best practices across network to improve dealer performance.
• Employee coaching to improve service standards at contact centers.
• Identifying at risk customers who potentially might deflect to competition and putting strategies in place to minimise churn.