Why our methods of measuring conversations and emotions are fundamentally flawed
Why do we get customers to quantify their emotions mathematically instead of counting the feelings and priorities?
How many things in your life are a perfect 10? You probably have to think hard about that. We have very few moments that are perfect sadly.
How many things in your life are OK, a little, a lot, very, completely, or more than? You probably were able to think about that question more quickly and with less effort.
The problem with emotion measurement
Clients and practitioners talk a lot about the role of emotions in customer experience.
I think the general position of the customer experience industry is we think that emotions are important to the bottom line, but we really can’t understand them properly.
This position has fuelled a general approach over the last 30 years creating a world of bad processes and worse customer surveys. It has led us to a world of Net Promoter Score (NPS), customer satisfaction (CSAT), customer effort score (CES), etc.... which is to say it's led us up the garden path to a wilderness – far away from the real world.
It’s wrong. It’s really completely totally absolutely and factually wrong. It is scientifically wrong, and businesses and government organisations have wasted billions on not understanding customers. They have been led there by the big six consultancies and the “Leaders of Customer Experience” by which I mean those who shout the loudest in the marketplace.
I want to explain how you fix your customer experience best practice and improve your focus on two things.
1. Start listening to conversations rather than just sending out surveys.
Conversations are designed by the human race to be understood. We created conversation as part of our drive to become human. Did you know that conversations are made up of simple patterns? That’s how we all understand each other. Simple patterns designed to be understood. There is a science called conversation analysis that can extract massive amounts of insight from conversation and yet less than 1% of organisations train or focus on the subject.
Cheaper
- Conversation is the richest, most available, source of insight you have, and it is designed to be understood.
- AI and experts existing data helping you make small changes to your conversations that improve outcomes.
More accurate
- Observe what actually happens in real situations.
- Conversation is designed to be understood.
- Calibrate with actual outcomes.
- People talk about people process technology trust culture problems - in fact everything you need to understand in a set of patterns.
Practical
- Identify words and phrases that can be improved and those that seem to work.
- Understanding what people actually say is more practical.
- Easy to test those small improvements.
- Zero effort experience for customers.
- Provide actual VoC (Voice of the Customer).
- CHANGE OUTCOMES.
Surveys on the other hand are:
- Costly complex systems to get short answers and VoC.
- Data skewed by collection methodology.
- Less bandwidth for less insight.
- Survey fatigue.
- High effort extra experience for customer.
- Witness statements after the event based on memory which are less reliable then observation of the experience.
- CANNOT CHANGE OUTCOME.
These are facts not opinions so the question is - having read this will you start think about conversation analysis?
2. Measuring emotions – the practical approach.
The situation with emotions is even worse in terms of business culture. The advice and thinking on the subject are unusable.
I and my crew of experts have spent the last three years building an approach based on real science and not service provider spin and marketing.
Emotions are understood through conversations. Today you understand emotions as do your friends family customers and colleagues. Gen XYZ Baby Boomers, etc. understand them too. Look around you the next time you are in a crowd and you will see who is angry or happy, and who is in between.
We understand emotions in terms of feelings and priorities. We may not always be able to express the exact words we need sometimes but in one-way (like conversations) emotions are simple patterns that we all understand but our organisations do not. They are active or passive on a scale.
But that’s not completely accurate. Organisations who focus on people like medical, police and social services think about emotions in a simple way because they need to understand elements of emotions.
So why does business get it so completely wrong?
The answer is maths. You do the wrong kind of maths at the wrong stage so you can’t get correct answers. You are trying to get customers to quantify their emotions mathematically instead of counting the feelings and priorities.
Emotions actually work like this – we are OK, a little, a lot, very, completely, or more than. This is what it looks like.
This is how we talk about things and how we think about things. If you do the maths to convert what people say it is massively more accurate.
Working this way, we end up with multiple feelings and active priorities. For example:
- “I like working with James because he is a laugh, but I wouldn’t take him home to meet my family he’s just too ……”.
- “I think their service is bad, but I can’t be bothered to switch”
- “I love the brand but it’s too expensive”
- “This is the third time I have phoned about this….”
- “I want to close my account”
You can understand all of these emotions. You simply don’t have the right fields in your existing systems, and you are not using what is already there. You are not counting what people say and then start producing reports.
Or to put it bluntly you are looking for the wrong information in the wrong place at the wrong time.
So, what can you do that is practical?
The best thing I can suggest is to follow these two links. Our website has information we publish freely and there are no sign ups, logins or email captures. There you will find a starting point to thinking in a different way and all the info that you need to start taking steps towards using real science to understand emotions and conversation.
- Conversations - https://www.cxfo.org/conversation-traits
- Emotions - https://www.cxfo.org/e-score-thinking
They will help you start to think more simply about these issues. Would you be willing to see if there are better ways of doing things?
Related content
As a seasoned professional with nearly 40 years of experience in the contact centre industry, I have been a driving force in shaping customer and employee experience for some of the world's largest organizations. From the earliest days of cloud-based contact centres to the latest innovations in conversation and emotion auditing metrics and...
Replies (1)
Please login or register to join the discussion.
I concur on a few things. However, the measurement of emotion is fraught with elephant traps. I speak as the designer of one of the first commercial emotional measurement systems, so I know a fraud when I see one.
Firstly:
1. Emotions are downstream. We appraise we 'feel', emotions do not drive behaviour alone, emotions do not exist in an isolated part of the brain. Without recourse to understanding the prevailing event, any measurement is a waste of time
2. Emotions are felt by the individual. To apply an NLP algorithm is not 100% wrong, but is a muted effect and misses this meaningfulness.
3. Emotions in the commercial world are usually a lot simpler (see Marsha Richens and my own extensive analysis). Indeed, from a review of several thousand datapoints, we came to the conclusion that in fact measurements were hiding a super factor of satisfaction/ dissatisfaction (why don't we measure dissatisfaction directly, I digress). So no joy there then.
4. Emotions are fleeting. Frequently a measure conflates emotional response with attitudinal measures on a scale. Simple satisfaction is better (ie thats the same as NPS)
5. Emotions reflect dispositions. Most of CX data is dispositional not root-cause. More on this another time. But reflect on the impllications of Cynefin.
6. I agree on opening up conversation, I agree on the value of narrative, I agree on big data proxies (loss aversion only).
7. We do not need emotion all the time. This is very important. If emotions are for learning then once learnt we do not need to feel so much.
8. Emotion and cognition come together around the relationship based event. So start with the event and use emotions in narrative as flags.
Unfortunately emotions are being used to sell vendor ware under many false promises. In addition, some certifiable and self -proclaimed 'gurus' in the market are spouting heaps of steaming BS when they really should be working down the tattoo parlour.
Happy to share any further information on my views. Emotion based academic references are: Lazarus, Baumeister, Oatley, Marsha Richens, Dave Snowden, Olaf Hermans, Feldman Barratt. A number of these I have personally interviewed and trialed there techniques.
Final point: close the loop with your agile teams who build stuff and create new events that are meaningful. Be less concerned about buying technology systems and putting in cost to your contact centre. IMHO.
Thanks (0)