Transcom helped a big European telco increase debt collection rates with machine learning.

The background.

We’ve been partnering with a big European telco company for years, delivering both inbound and outbound debt collection services. 

In conversation with the client, we realized that the new digitalization processes in the telco industry required a new debt collection management approach. This is when we performed a detailed analysis of the processes and systems to find the best solution for this operation.

The end goal was to increase the debt collection rate and reduce the operating cost with the implementation of different self-payment solutions. We found that a machine-learning solution would benefit the client the most and deliver exceptional results. The solution was designed and implemented to reach the new service challenges.

The solution.

The machine learning solution was designed and implemented to reach the new service challenges and help the client reduce costs and increase debt collection rates. The project consisted of:

• Historical Data Analysis: more than 6 machine learning models were selected using +1M of historical leads.
• Data analysis & ML algorithm selection: a prior statistical analysis was conducted to understand the relationship among variables. A range of machine learning algorithms were tested, and finally, the prediction model was selected.
• New process definition: every day, an automated process characterizes each customer with a predictive debt collection classification. Based on the debt recognition and the ML prediction, the customer is pointed to the self-payment process or handled by an agent in inbound calls. The outbound campaign is configured based on machine learning predictions to maximize customer contacts.
• Monitoring & models retraining: all data is monitored daily, and machine learning models are retrained every month.
• Increasing business: the success of this machine learning initiative has implied new business opportunities to extend the service to the next phases of debt (not early debt).

The results.

Combining the power of analytics and operational expertise, Transcom was able to generate customer insights that have helped the company to continue delivering successful operations, adding value to day-to-day interactions and supporting any strategic decision based on the generated insights.

This implementation has had a positive impact across different areas of the business. Client satisfaction increased to 100%, while NPS increased from 17 to 33, comparing the results obtained in the previous 2 years. 

The client is more than happy with the achievement of initial objectives since the debt collection rate increased by 6.8%. The self-service debt collection now makes 17% of all debt collected, saving 860 operating hours for the client.

Thanks to this successful implementation of machine learning which improved the operation significantly, Transcom was awarded by Customer Relationship Excellence Awards for Best Debt Collection service in 2022.