TV Recommendation System - Targeted offers for telecom users

A major telecom provider sought to improve the relevance and efficiency of its customer marketing campaigns. The goal was to move beyond static segmentation and deliver personalised offers that adapted dynamically to each customer’s profile and behaviour. This required consolidating fragmented datasets and creating a system capable of processing millions of customer profiles at scale.

I designed and implemented a cloud based recommendation system that integrated customer demographics, product usage patterns, viewing data, and contract details into a unified analytics layer. Machine learning models were developed to predict the most relevant offers for each individual, with the capability to evolve as customer behaviour changed.

The system was embedded directly into the campaign management workflow, allowing marketing teams to create and launch targeted campaigns in hours instead of days. Continuous A/B testing ensured measurable impact and informed iterative improvements to the recommendation logic. This closed feedback loop enabled the model to learn from real-world performance, improving accuracy and driving sustained gains in engagement over time.