Formula for customer happiness
Somehow knowing whether a customer is happy with a service provider’s services, and why is that, is extremely important. Regardless of the industry, customer experience rank at the top of reasons customers choose a service provider as well as at the top of areas for service providers to invest in.
However, having data and understanding of customer experience and happiness is not that straightforward: Customers might not want to communicate straightforward about their experience, collecting that feedback might be a challenge, experience might be impacted by factors external to service / product itself.
AI/machine learning technologies are useful in reducing wide range of data points into more simple inferences, e.g. how’s the customer experience, is the customer happy. Such inferences are at best predictions, but the better those predictions are, the more useful they are. We have defined a formula for customer experience index (CEI) in telecommunications. In telecommunications there’s lot of technical data available and practically everyone is a customer of a telecommunications service provider’s, so the answer to the question if the customers are happy with their experience and why, has very impactful applications resulting from it. Ideally, we would like to prescribe and automate predictive improvements of customer experience based on our CEI formula.
Our current CEI formula (depicted below), works like a three, collecting quality data from multiple branches of services i.e. data, voice, customer care, billing etc. New branches, e.g. digital channels experience, Social Media signals of even weather can be added into the formula. Branches are summarized together based on weights applied for each branch and the result is normalized at each level between 0 and 100. This results into top level CEI between 0-100, the higher the number is, the better is the customer’s overall experience supposed to be. In order for the CEI reflect customer experience and happiness, the weights can be tuned up either by experience or based on data, specifically by a machine learning process. If there’s a CEI formula that predicts customer happiness well, it can be very powerful as it contains the information how to impact it.
We would like to perfect the CEI formula to reflect customer happiness as precisely as possible. We got the telco services quality part covered, but would like to understand especially how the external factors impact customer experience. In this challenge the task is to study, by defining and using machine learning applications, how the external factors (digital channels experience, signals in Social Media, weather, events…) impact customer experience and how to collect and model the data automatically so that the experience modelling exercise can be automated. Also, insights how those factors could be addressed when dealing with customer and their impactfulness to overall happiness are of interest.