22/05/2017 | Written by: Mando Rotman
Categorized: Analytics | Generic
Share this post:
“A dissatisfied client is a loyal client”.
Even the most accepted theories of CRM might be not applicable to your organization´s situation and Data Scientists can discover such fascinating insights from your data.
The massive volumes of data represent opportunity to acquire information and make an organization much more intelligent and adaptable. It enables us to challenge assumptions and beliefs such as: “a dissatisfied client is a client at risk”.
IBM has worked with a major European bank to better assess and anticipate the risk of seeing individual customers move to the competition.
To establish a preliminary position, we pursued our “best practice process“ to first understand the business situation and to clarify the definitions involved. We then explored the available data and worked with business subject matter experts to come up with ideas and hypotheses. We generated features from the available data based on these hypotheses, based on our data exploration insights and based on our experience in similar situations. These features were fed to different machine learning algorithms in iterative cycles.
An expected result was that one of the algorithms picked up the feature “relative number of complaints” as an explanatory variable for the risk of customer churn. However, a closer look showed that it actually had a negative correlation to churn risk! This was indicating that customers with relatively more complaints were actually less likely to churn. That was counter-intuitive at first and therefore a clear example of one of the added values of applying analytics to solving business challenges!
“Customers with relatively more complaints turned out to be…
less likely to churn.”
We now hypothesise that these may be the customers that still have an underlying positive motivation to complain, they are trying to improve something to benefit from this in the future. Customers who do not complain may well be customers who have already ‘given up’ on the relationship with their bank.
To retain the informed consumers of today, banks need to put the client at the heart of their activities. This requires deeper understanding of clients’ needs, their preferences and their behaviour. Now that businesses can analyse so much available data through advanced algorithms they can get really serious about this.
IBM offers a suite of solutions for the banking sector (and the rest of industry) designed to create a business focused on the client. Put your intuition to the test with IBM Data Science Experience.