According to researches, it has been found that it is cost effective to retain your old customers rather than acquiring the new ones. On average, retaining existing customers is five times cheaper than the cost of recruiting new ones. By such project, one gets an estimate of which customers are likely to leave and one can put in efforts to retain them.
It is when a customer is likely to stop doing business or leave you. Also known as Customer Attrition.
There are different types of Churn : a) Contractual churn - when a customer cancels under contract. e.g Cable TV b) Voluntary churn - when customer leaves on their own e.g. streaming subscriptions c) Non-contractual churn - customer leaves under non-contractual terms . e.g. online browsing d) Involuntary churn - expiration of service or utilities by the provider e.g credit card expiration
We go through multiple churn datasets to get a better understanding how different customers behave in a different manner.
For Telco dataset , we first do exploratory data analysis. And further build up a model on the data.
Another is the Bank customer dataset. Through this project, we observe the multidimesnional behavior of the customer and try to figure if a pattern is being reflected. This can aslo contribute in identifying trends and improvising in business.
The Multidimesnional Analysis of Customers gives pretty good insights. It has well classified data. We go through multiple steps such as Exploratory Data Analysis , Cluster Analysis and finally Principal Component Analysis to thoroughly understand the data in Python.
Once we have understood the Customers' behavior patterns , we further move to predict the customer attrition. The data is properly processed through Problem Definition, Solution Discovery , Best fit Algorithm Search and Finally deploying the same.