Amazon Sagemaker and Machine Learning Solutions

The Challenge and Why Amazon Web Services

The rate of customer churn is a critical metric for every business. Customer churn refers to when a customer ceases their relationship with a company. For example, this could be lack of renewing a service or agreement, closure of an account or cancellation of a subscription. The cost of customer churn includes both the lost revenue and the marketing cost in replacing those customers with new ones. Boostability’s challenge was to predict customer churn, identify customer behavior and shed light on the growth of their business.

Boostability recognizes the benefits they have received from leveraging Amazon Web Services (AWS) and continues to rely on the help and direction provided by 1Strategy, an AWS Premier Consulting Partner.

“AWS provides Boostability affordable and scalable services, which allow us to grow faster and more efficiently,” said Jared Turner, CTO of Boostability. “We continue to partner with 1Strategy due to their extensive AWS experience, as well as their strong capacity in applied machine learning and deployment in the AWS ecosystem.”

The Benefits

The customer churn rate measures the speed at which a business is losing their customers within a specific time period. 1Strategy developed a customer segmentation and churn prediction model to identify high-risk customers who may leave Boostability’s services in the following month, and they provided guidance and support with the following:

• Creation of an architectural plan to support customer segmentation and churn model in AWS
• Creation of customer segmentation and churn model infrastructure
• Extraction of source data from Amazon RDS and other datastores and the transformation of data to fit within the proposed architecture created in the above deliverable
• Creation of customer segmentation and churn model and evaluation of both for accuracy
• Establishment of location inside customer AWS account to store and secure customer data for use in the model
• Deployment of the project model to Amazon SageMaker to be accessed through application programming interface (API) endpoint for programmatic access

1Strategy designed and created a churn prediction model for Boostability that utilizes key AWS services, including Amazon SageMaker, Amazon S3, Amazon API and AWS Lambda (see image below). This allows Boostability to submit a customer profile dataset and receive a prediction ranking of the customer’s churn risk level. In addition, this churn model is based on Boostability’s data and predicts their churn rate more accurately and reliably while delivery forecasting insights quicker than their previous, manual process.

Amazon SageMaker is an end-to-end solution for machine learning in the AWS ecosystem and gives freedom for data scientists by providing the additional capabilities and benefits that companies like Boostability are able to utilize, including:

• Automatic scaling to meet performance needs
• Support of any model built into a Docker image

Customer churn impedes growth, but with Amazon SageMaker and the churn model developed by 1Strategy, Boostability is able to monitor their churn rate on a monthly basis, enabling them to identify high-risk customers before they churn. This also allows the company to take a proactive approach to offer retention incentives that prevent customer loss and gain feedback on their customers’ behavior.

“Leveraging Amazon SageMaker enables Boostability to determine our customer retention success rates and aides in identifying strategies for growth,” Turner said.

The work described in this engagement was originally completed by 1Strategy, a TEKsystems Global Services company acquired in 2019. As of June 2023, 1Strategy has fully integrated with TEKsystems Global Services to continue to deliver AWS expertise to customers. Learn more about our AWS solutions.