Today, even a small reduction in subscriber churn can result in millions worth of benefit for service providers. Studies have shown, acquiring new customers’ can cost up to five times more than satisfying and retaining existing customers.
In the past analyzing reasons why a customer has decided to stop being a customer used to be a straightforward task because the data was limited, but in the era of Big Data, a host of data dimensions are given due consideration that can no longer ‘just be ignored’, making it a much more complicated task. With Big Data, we can now gain deeper and more relevant insight into customer behavior, enabling companies to providing better customer service and increasing revenue significantly.
Aware of the potentially huge financial impact of subscriber churn, and the advantage of operating in an environment with abundant customer data, most of the providers traditionally, have invested in human capital and technological infrastructure to enable the use of this data to understand customer churn & tried to predict the same using predictive statistical modelling. The process and methodologies adopted reveal a right mix of variables to predict churn.
How then does big data help us in variable selection and transformation? Let’s visualize the unstructured big data. Visualizing the data takes fraction of time it would traditionally take to identify not so obvious events & patterns. Natural Language processing & Sentiment analysis can reveal consumer’s emotional gradient as expressed in speech, social media, emails and other unstructured forms of data. Employing big data techniques coupled with sharp business acumen, reveals event and even sequences that lead to churn, that in turn can be transformed into variables that feed back into our traditional models.
Integrating new data about customers from emerging contact points gives much more meaningful insights into customer behaviour that can in turn be tested and closely acted upon. These inputs not only give us more in-depth understanding of customer behaviour but also help incrementally update the methodologies for churn modelling while increasing accuracy of existing models that have been traditionally adopted.
The service providers can accurately predict on basis of customer behaviour, likelihood of the customer to churn, if they know which customers are at high risk of churn and when they will churn, they are able to design customized customer communication and treatment programs in a timely efficient manner
No comments:
Post a Comment