Understanding Customer Churn with Approachable AI

Jeff Kimmel
Oct 26, 2020 · 4 min read
Image for post
Image for post
Photo by Ibrahim Rifath on Unsplash

The elipsa Analytics Platform provides predictive data solutions driven by approachable AI. Our plug and play predictive tools enable data users to conduct complex data experiments with just a few clicks.

Data Drivers is a diagnostic analytics tool that allows users to quickly and easily identify and extract patterns within their data to go beyond the analysis of what happened and start to explain why it happened.

Elipsa’s intuitive platform allows data users to unlock the answers hidden in their data leading to accurate data-driven decisions.

Customer Churn Analysis

We will examine the use of Data Drivers to help explain the problem of Customer Churn.

Every business needs customers. These businesses seek to add new customers but they also work hard to retain existing ones. The loss of these existing customers is called churn. The more you understand about churn, the more you can prevent it, and in turn, increase revenue.

Through the elipsa platform, users can seamlessly build a no-code predictive model to monitor for potential churn (predicting what is likely to happen). Through Data Drivers, users can extract the key data sets that are driving actually driving these customer decisions (the why).

Computers are much faster and more accurate at connecting the dots amongst big data, but to understand the cause of that relationship, the why, you need to partner a computer’s speed with human intuition. Data Drivers seeks to do just that.

The Data Set

For this example, we are using the Telco Customer Churn dataset from Kaggle. The idea is to connect the dots, finding patterns that help to explain why a customer will churn by focusing on the customer attributes that have the strongest indication of predicting churn.

The data set includes customer account information, demographic data, and services that each customer signed up for. The goal is to automatically build a model using the elipsa analytics platform and to extract insights from that model indicating what attributes are driving customer behavior.

Understand the Why

The elipsa platform automates the analysis of connecting the dots and reports back which ones are most important for determining who will churn and who will not. This automated analysis uses machine learning to efficiently turn data into information. With Approachable AI (useability, explainability, and accessibility), this information is put directly in the hands of the user who understands it best. The result is actionable insights in seconds allowing the user to apply domain expertise and help reduce churn and increase revenue.

Results

By dropping the data file into the elipsa platform, we are able to select the Churn column as the target, and the remaining columns as potential drivers. Within seconds we are presented with insights helping to explain what is driving customer churn.

Image for post
Image for post
Customer Churn Drivers

Customer Churn Drivers

Based on the results, you can see that attributes such as customer tenure, their monthly fees, and whether they are on a month-to-month contract are the top attributes in terms of predicting customer churn. The numbers in the chart indicate the % increase in model performance from including the specified attribute. The color indicates whether higher values of that driver tend to result in churn (the blue items) or not (the orange items).

The data seems to show that this telco company has built up brand loyalty, or a high cost of switching, as tenure is the most predictive attribute, and as tenure increases the probability of churn decreases. From there we can see that customers on month-to-month contracts with high monthly fees are most likely to churn. The data gets interesting as you move further down and see patterns emerge that show things such as senior citizens being a high churn demographic, and customers that don’t purchase tech support or online security are less likely to stay.

Through approachable AI, we were able to quickly analyze a large number of records to extract critical insights. With insights in hand, this business can focus its efforts on better understanding the issues within the senior citizen market, make a push for long-term contracts, or focus sales efforts on what appears to be ancillary services such as tech support and online security. Effectively, the business user is given the tools to better understand what is causing churn so that he or she can stop it and increase revenue.

Conclusion

There are a lot of insights in your data, and with Approachable AI those insights are now able to be extracted directly by the domain experts across industries and organizations. The computer is much better than humans at sorting through large amounts of data finding relationships that indicate causation. However, those insights are still better utilized in the hands of the business user to enrich those insights with domain expertise. The elipsa platform allows data users to directly apply advanced analytics leading to faster and better insights and in turn better data-driven decisions.

Elipsa

Approachable AI for Business Users

Medium is an open platform where 170 million readers come to find insightful and dynamic thinking. Here, expert and undiscovered voices alike dive into the heart of any topic and bring new ideas to the surface. Learn more

Follow the writers, publications, and topics that matter to you, and you’ll see them on your homepage and in your inbox. Explore

If you have a story to tell, knowledge to share, or a perspective to offer — welcome home. It’s easy and free to post your thinking on any topic. Write on Medium

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store