Question

Description:
Sprint Nextel has the highest rate of customer churn (the number of customers who discontinue a service) in the cell phone industry, amounting to 2.45 percent. Over the past two years, Sprint has lost 7 million subscribers. Management wants to know why so many customers are leaving Sprint and what can be done to woo them back.

Requirement:
Does Data Mining help the company in unfolding the reasons behind why their customers are leaving? Explain why or why not.

Based on the information provided in the the beginning, answer the question

Answer 1

There might be a lot of possible causes for Sprint’s customer churn, which are mentioned as follows.

• New entrants in the market providing similar values with less price
• Bugs (Network issues, downtimes) in product / service leading to customer
• Poor Customer support not addressing customers’ queries, complaints properly
• Unfulfilled Promises, leading to customer churn and negative word of mouth
• Wrong target Audience, or not identifying valuable customers
• Economy crisis, when people become extra price sensitive and survive with minimum requirements
• Lack of Competitive Advantage (pricing offers, promotions, discount plans, Referral incentives)
• Lack of customer engagement or communication
• Lack of long term Customer retention Plans
• Excessive advertisements to bother customers
• Lack of Brand value or offerings as per customer needs

Based on the data provided, Sprint Nextel might have a huge number of users or subscribers. Above mentioned factors are the most likely causes of its sudden customer churn. Now Data Mining can interpret large amounts of existing data from a cenralized data source or data warehouse and provide us with a lot of analysis, forecasts, insights and patterns. It's really helpful for a Telecom company like Sprint Nextel to monitor Customer's usage, understand their needs, behavioral trends since their purchase and analyze them for insights. Definitely Data mining will unfold the exact reasons behind this customer churn, no matter how big the data is. Sprint Nextel has all the customers and their usage data, who abandoned their service. Analysing seven million customers' data is not a hectic task here, if it's stored in a centralized database. Also these responses are highly reliable, as customers' behavior can be drawn from their usage, not from surveys. But identifying the required data to draw patterns and evaluate these factors or metrics is very important and might be difficult in order to find accurate reasons. Suppose in order to find Customer dissatisfaction, all the possible reasons such as Network issues, Downtime schedule, Technical glitches, Feedback from customer care and many more unknown reasons need to be segregated. Then these issues are tracked, with when the customer abandoned the service and we find relevance to the above issues. Next we can conclude whether it’s because of customer dissatisfaction or due to some other reasons. And the analysis goes on. Off-course with the ease of analytics tools, it has been made easier to carry out these processes. Involvement of ML (Machine Learning) and AI (Artificial Intelligence) programming into Data mining has made it hassle free to cleanse the data, find the bottlenecks and analyze them efficiently within very lesser time than expected.

If we think about other alternatives, when companies are relying on Surveys, market research, studies, FGDs (Focussed Group Discussions). It’s easier to find out the reasons for customers leaving the service. But can we trust their response, can we verify them and also people hardly respond to surveys. Another downside is that it can have very less number of sample space compared to Data Mining analysis. So for such a large datasets, Sprint Nextel must use and rely on Data mining to find the bottleneck and easily draw out the reasons of customer churn.