AWS for Industries

T-Mobile US, Inc. Modernizes Sales Commissions with AWS

T-Mobile US, Inc. One of the largest US telecom carriers is rapidly increasing their subscriber base. They had concerns about meeting daily sales level agreement expectations for frontline employees and business stakeholders. They were using a legacy Software as a Service (SaaS) application for compensations calculation and reporting. The application lacked agility, scalability, and did not provide the value proposition that T-Mobile was expecting. T-Mobile US, Inc. teamed up with Amazon Web Services (AWS) to develop a fully agile, auto-scaling, and cost-effective solution for sales performance management. This solution drives sales and revenue growth for T-Mobile by effectively managing the sales performance. The solution exceeded the performance goals set during initial scoping even though the sales traffic volume doubled over the last two years.

The solution

Here are the key design principles the AWS and T-Mobile team adhered to when outlining the framework for this new application: 1) Embrace a microservices architecture. 2) Enable parallel processing and synchronous task handling. 3) Allow customizable flow execution. 4) Operate without the constraints of licensing fees. 5) Maintain a stateless design with minimal infrastructure overhead and infrastructure as a service (IaaS) on demand. 6) Incorporate auto-scaling capabilities to accommodate unpredictable processing demands.

In our pursuit of improved scalability and reliability, we strategically chose to decouple the application capabilities from business rules implementation, utilizing SQL for submission and processing within T-Mobile’s custom Spark application. This strategic maneuver not only enhanced the platform’s capabilities but also accelerated time-to-market, enabling swift development of new business rules necessary. Furthermore, it relieved concerns about application-level details by abstracting away complexities, allowing the application to handle computational tasks seamlessly in the background.

The following is a diagram of the solution:

T-Mobile’s custom Spark application - figure 1

We decoupled the solutions into two applications to improve the agility and eliminate business dependencies, the applications are:

1. The Sales Compensation Data application is a core application that handles all upstream data necessary for compensation calculation. This application acts as a landing zone for data pushed by the upstream systems (billing, customer relationship management (CRM), payments, and others). It uses Amazon EventBridge for event-driven actions such as batch data transfer, AWS Lambda functions to trigger the data migration via AWS Database Migration Service (AWS DMS) with the AWS Schema Conversion Tool (AWS SCT) to convert database schemas and Extract Transform Load (ETL) jobs. The application uses Amazon Simple Notification Service (Amazon SNS) to push a notification to the Sales Calculation and Reporting application.

2. The Sales Calculation and Reporting application, which calculates, adjusts transactions, and reports on the overall process state. AWS Step Functions fully orchestrates the execution of synchronous jobs. Jobs are running in Amazon EMR using T-Mobile’s custom code as Spark applications. To process batch traffic efficiently and cost-effectively, we used Amazon Elastic Cloud Compute EC2 spot instances with auto-termination enabled for these scalable jobs. This application uses Amazon Simple Queue Service (Amazon SQS) to queue up the messages coming from the Sales Compensation Data application. It leverages Amazon RDS MySQL to store the metadata, and Amazon DynamoDB as a key-value pair store for sales agents’ information. While Amazon Simple Storage Service (Amazon S3) is the central object storage for both raw and processed data, Amazon Athena is being used to query data that resides in Amazon S3, and AWS Glue is being used for Spark jobs and integrate with external systems such as SQL databases on-premises.

The data flow

Data ingestion from upstream involves gathering data from over fifty various systems such as business inputs or manual overrides. This data may encompass transactions, quotas, human resources (HR) and CRM information, metrics, or configuration data. Upon ingestion, rigorous validation is performed to ensure both accuracy and completeness. Any missing or inaccurate data is promptly flagged for thorough investigation and correction.

Following validation, eligible transactions are identified based on predetermined criteria’s, including payee eligibility and quota/target earnings calculations. These identified transactions are then allocated to suitable payees, using specified business logic, to support compensation constructs like Sales Performance Incentive Funds (SPIFs), individual performance, or store metrics. Subsequently, once transactions are appropriately assigned, payouts are processed accordingly, integrating seamlessly with the downstream systems such payroll applications, Customer Relationship Management (CRM), and HR systems. This complex integration is done via AWS Lambda and Amazon EventBridge.

The elasticity in batch operations

The auto-scaling of AWS services such as Amazon EMR and AWS Step Functions is a critical feature for handling batch operations. Elasticity in Amazon EMR enables clusters to automatically scale up or down in response to workload fluctuations, optimizing resource usage and cost efficiency. It dynamically adjusts cluster instance size based on metrics (such as CPU and memory utilization) ensuring sufficient capacity for efficient batch processing under peak loads. AWS Step Functions, a serverless orchestration service, facilitates the coordination and automation of workflows, including Amazon EMR steps. It offers flexible resource allocation and supports parallel processing, enhancing the efficiency of batch operations. To save on costs, T-Mobile leveraged an ephemeral design which involves releasing the assets when not in use.

The results

T-Mobile is seeing significant improvements across the sales compensation process:

  • Improving the system reliability with a decrease of 80% in application related issues.
  • Reducing the platform cost by 40%.
  • Reducing the processing time by 50%.
  • Eliminating the latency between sales reporting and compensation statements.

Conclusion

T-Mobile US, Inc. built a highly scalable, reliable, and cost-effective solution by leveraging AWS services to digitize their sales performance management. The new solution saved them time, money, and improved the application reliability, resiliency, and operational excellence. To learn more about how telecommunications companies are leveraging AWS Services, visit Telecom on AWS or contact an AWS Representative today.

Further Reading

 

TAGS: , , , , ,
Mounir Chennana

Mounir Chennana

Mounir Chennana is a Lead Solutions Architect within the AWS Telecom business unit. Mounir has 20-years’ experience in the telco industry. He’s passionate about telecom transformation, OSS/BSS modernization, and leveraging the emerging technologies such as big data, AI, generative AI, and advanced data analytics to build “Telcos of the future”.

Paul Heng

Paul Heng

Paul Heng is Software Development Manager at T-Mobile within the frontline experience organization supporting sales compensation for over 40K employees. Paul has been with the company almost 20-years—starting off his career in frontline as a retail sales representative. His experience there keeps him grounded to ensure that T-Mobile’s sales employees come first.

Robin Sayed

Robin Sayed

Robin Sayed is a seasoned Solutions Architect Manager within the AWS Telecom business units, boasting nearly two decades of extensive experience in the telecommunications industry. With a robust background spanning Radio Access, Network, and Packet Core domains, 4G/5G, Voice Core, and WIFI calling solutions tailored to meet the needs of governments, enterprise businesses, and consumers alike.

Sanjay Mahobia

Sanjay Mahobia

Sanjay Mahobia is Sr Software Engineer at T-Mobile in sales compensation who led the compensation transformation effort within the organization. With a keen eye for detail and eagerness to learn new technology stack, Sanjay enjoys diving into the depths of data-driven solutions, leveraging the power of modern architecture to drive impactful outcomes that directly benefit the business.