Safe bets

: forecasting who’ll buy home alarms

Abstract

Telecommunications companies face constant pressure to adapt to technological advancements and updates. Simultaneously, these organizations explore new revenue streams to sustain growth and address competitive pressures. NOS expanded its operations into the home alarm market in 2022 in partnership with Securitas, a European leader in the security sector. This strategy aligns with the prospects of increased demand for smart devices in Portugal and the increased benefits of targeting adjacencies. Cross-selling to existing customers is one of the growth levers of this new service. NOS already has segmentation criteria for the next best offer system. However, for Alarms specifically, business rules are used to select the most likely-to-buy customers due to its recency. The aim of this thesis is to (1) collaborate with the marketing product team to identify key drivers influencing the purchase of alarm products and (2) develop a propensity model using features that best capture these drivers. Logistic Regression, Decision Tree, Random Forest, and XGBoost were tested. Emphasis was placed on geographical variables, as significant work had already been conducted to map alarm installations at specific coordinates across Portugal. The results of this study showed that financial variables play a crucial role in calculating the likelihood of a customer buying this service, together with the previous relationship with NOS. Geographical variables showed relevance for improving the performance of the models but with a low degree of significance. Finally, it was possible to develop models with higher performance compared with NOS current approach.

Date of Award	31 Jan 2025
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Nuno Filipe Loureiro Paiva (Supervisor)