Impact of Emotion and Sentiment Dynamics on Sales Outcomes and Customer Satisfaction
Emotion drives human connection. We buy things we desire, not the things we need. We don’t understand our gut feeling, but we know that it is a good idea to follow it. Naturally, emotion also influences our communication. How often have you been frustrated with lengthy calls to customer service? What kind of skills the call centre agent need to have in order to calm you down and earn your loyalty back? Is it at all possible to do it through upgrading agent’s communication skills, or is there another, better way? What is the role of emotion with regard to the quality of communication and outcomes?
In recent CX Emotion conference we explored these questions, and the impact of AI to encourage and deliver a positive and emotion-centric approach to customer interactions.
As human beings, we take many things for granted. For example, an ability to recognise someone else’s emotion, or intent. In reality, many people struggle recognising emotions and responding with the right tone of voice. We think that we are all very good at showing empathy, but the data shows that we aren’t. This is where I believe Artificial Intelligence and Natural Language Processing have a big role to play, to help humans become better communicators, identifying flaws in our communication and pushing our abilities to the next, previously unachievable level.
We are entering a new era of empathic and ethical human machine collaboration. Ethical is very important in the context of AI, and we have already made progress on gaining more clarity on what ethical means here, and emphasising the importance of privacy. This era, where our emotions and privacy are not neglected, was going to come anyway, and with pandemic, many things, including this, have been accelerated.
“All hands on deck” approach in the Contact Centre Struggle
Here is an email I received during the lockdown, from one of my favourite companies when it comes to customer service:
“However, we are still struggling to get anywhere close to our normal level of service. Therefore, we ask that unless you need to tell us about something urgent, you try to avoid contacting us. If you really need to speak to us, please expect that you may wait close to 30 minutes. And, it is taking much longer than usual for us to reply to emails.”
Prioritisation based on customer judgement: While it’s nice that they are giving me a heads up, the challenge here is that they are asking me, their customer, to make a judgement on whether my enquiry is urgent or not. Everyone’s problems are their biggest problems, so the outcome of this email might actually be counter-productive. Urgency is relative.
Contact Centre Epiphany
What if, instead, they said something like this:
“Please get in touch with us and we will try to help even though we might not be able to personalise our interactions, and we also might delay our response to you if we decide that other customers are in the more vulnerable position than yours”
Prioritisation based on customer feelings + contact reason: This would mean that they are taking control on making the decision based on hearing their customers first and understanding why they are getting in touch. In reality, to handle potentially large requests of customer enquiries in this way, and redirecting ‘urgent’ queries to the relevant staff, it would require putting AI and automation at the core of the contact centre set up, which isn’t something you can do overnight, in the middle of pandemic.
Which is why many companies are now pushed to look into at least automated if not fully AI-driven solutions.
Pandemic is an Extreme Situation Though
Once Covid-19 crisis is over, do we need to care about customer emotions and AI? Absolutely, and even more so. We already know that Emotion Drives Outcomes. Your customers will remember how they felt in a particular situation especially if it was emotionally charged, and your response to them will largely influence your Net Promoter Scores (NPS), Customer Effort Scores (CES), Customer Satisfaction (CSAT), or any other measure you are collecting to learn about customer experience from their feedback.
Figure 1: Emotion Drives Outcomes
As for AI, at Sentient Machines, we’ve developed Sentiment Index — one number that reflects quality of a conversation and includes many elements related to the quality of customer interaction. Among other things it considers:
customer sentiment,
agent’s response to that sentiment,
agent’s empathy,
agent listening skills ,
conversation dynamics (e.g. the impact of agent’s responses to the customer’s feelings and reaction throughout the call),
differentiation of a call to the others,
positive smart language for agents.
What is Positive Smart Language?
Our definition of positive smart language is derived from millions of customer interactions and is defined as the language that is encouraged to be used by agents, to stimulate increased customer happiness while focusing on quick resolution.
Positive Smart Language is one of many factors that are included in our Sentiment Index.
What is the Relationship between Sentiment Index and Sales Outcomes?
We’ve analysed thousands of interactions for which we already knew if they resulted in a successful sale or not. In Figure 2, we show a comparison of the Sentiment Index derived from successful vs. unsuccessful sales calls. As we can see, we observed almost 27% difference between the two as shown in Figure 2.
Figure 2: Higher Sentiment Index Drives More Sales
What is the Relationship between Sentiment Index and Customer Satisfaction?
In Figure 3, we show how our Sentiment Index impacts Customer Satisfaction (CSAT) Scores derived from inbound Customer Support calls. Here we have:
CSAT Agent 1 — the lowest score for customers rating agent’s performance, and
CSAT Agent 5 — the highest score for customers rating agent’s performance, and also
CSAT Resolution 1 — the lowest score when customers are rating whether their issue was resolved or not, irrespective of the agent’s performance.
Figure 3: Higher Sentiment Index Drives Agent CSAT Scores, but not CSAT Resolution
We can see that higher Sentiment Index drives CSAT Agent Scores, but not CSAT Resolution. Even though the Sentiment Index for CSAT Resolution 1 is almost 5% higher than for CSAT Agent 1, customers are staying with the lowest CSAT Resolution score of 1, demonstrating that effective communication in contact centres is not only about training your agents, but also about making sure your products and services are up to the required standard, and that quick resolution of customer’s problems is as important as a well trained agent on the support line — if they are to aim for happy customers and high ratings.
What is the Relationship between Sentiment Dynamics within a call and Sales Outcomes?
We further look into the impact of emotional dynamics. At Sentient Machines, we have an algorithm that monitors the change of customer’s emotion and the impact the agent’s responses make. In Figure 4, we can see that ‘Positive to excitement’ dynamics change has the biggest impact in sales conversations. In other words, agents who are more successful in sales demonstrate the ability to turn a positive conversation into excitement 5.4 times more than those in unsuccessful sales.
Figure 4: Emotional Dynamics & Upward Turn in Sentiment Drives More Sales
What is the Relationship between Sentiment Dynamics within a call and Customer Satisfaction?
Looking into the Emotion Dynamics and CSAT in inbound calls, perhaps it is not surprising to see that there are no low CSAT scores for both agent and resolution, if there is a shift from positive to excitement (see Figure 5) . Whereas, such a dynamics exists in the calls rated with CSAT Agent 5.
Figure 5: Agent ability to turn positive sentiment to excitement drives high CSAT
On the other hand, if there is a negative emotion at the beginning of the call, the situation looks very different (see Figure 6). Now, despite the agent’s ability to turn that negative emotion into positive towards the end of the call, some customers are still rating calls with the lowest CSAT Resolution and CSAT Agent scores of 1. This confirms our previous hypothesis that no matter how amazing your agents are in their communication skills, customers want quick resolution.
Stats in Figure 6 also suggest that elimination of negative emotion could be the best approach in order to avoid low customer satisfaction.
Figure 6: Agent ability to turn negative to positive sentiment doesn’t eliminate low CSAT
Which leads us to the question of — how do we eliminate negative emotion?
A simple solution is to focus on building an emotion-centric contact centre.
How Do You Build an Emotion-Centric Contact Centre?
To build an emotion-centric contact centre, what you do is as important as how you communicate about it. More concretely, the quality of your product and service is as important as the speed at which you react to your customer feedback, as well as how you communicate to your customers about it.
In the next two blogs, we will explore in more details how to use Actions and Communication in order to build an emotion-centric customer experience.