Segmentação comportamental de clientes usando agrupamento automático de dados
Arquivos
Data
2023-07-10
Tipo
Trabalho de conclusão de curso
Título da Revista
ISSN da Revista
Título de Volume
Resumo
Tendo em vista a crescente importância da análise de dados para a personalização de mar keting e a identificação de oportunidades de mercado, este trabalho investiga a aplicação
de técnicas de aprendizado de máquina para a segmentação de clientes, visando aprimorar
a eficácia das estratégias de marketing. A abordagem RFV (Recência, Frequência, Valor)
é utilizada, permitindo traçar o perfil do consumidor com base em seu comportamento de
compra. Os algoritmos K-means e DBSCAN são aplicados para o agrupamento dos cli entes. A eficácia desses dois algoritmos na tarefa de agrupamento é comparada em uma
pesquisa empírica, que analisa dados de milhares de clientes de uma empresa de varejo
online. Os resultados mostram que o algoritmo K-means superou o DBSCAN na tarefa de
agrupamento para este estudo, sendo capaz de distinguir efetivamente entre os grupos de
clientes, o que facilita a construção de estratégias personalizadas para cada agrupamento.
Por outro lado, o algoritmo DBSCAN agrupou uma grande proporção dos clientes em um
único cluster. Esses achados reforçam que a escolha do algoritmo apropriado é crucial para
a eficácia da segmentação de clientes e, consequentemente, das estratégias de marketing.
Além disso, as empresas podem utilizar essas descobertas para refinar suas estratégias de
marketing, garantindo que as ofertas sejam direcionadas de maneira mais eficaz para os
diferentes segmentos de clientes.
Given the growing importance of data analysis for the personalization of marketing and the identification of market opportunities, this work investigates the application of machine learning techniques for customer segmentation, aiming to improve the effectiveness of mar keting strategies. The RFV (Recency, Frequency, Value) approach is used, allowing to trace the consumer profile based on their purchasing behavior. The K-means and DBSCAN algo rithms are applied for customer clustering. The effectiveness of these two algorithms in the clustering task is compared in an empirical study, which analyzes data from thousands of customers of an online retail company. The results show that the K-means algorithm outper formed DBSCAN in the clustering task for this study, being able to effectively distinguish between customer groups, which facilitates the construction of personalized strategies for each cluster. On the other hand, the DBSCAN algorithm clusters a large proportion of customers into a single cluster. These findings reinforce that the choice of the appropri ate algorithm is crucial for the effectiveness of customer segmentation and, consequently, marketing strategies. In addition, companies can use these findings to refine their marketing strategies, ensuring that offers are more effectively targeted to different customer segments.
Given the growing importance of data analysis for the personalization of marketing and the identification of market opportunities, this work investigates the application of machine learning techniques for customer segmentation, aiming to improve the effectiveness of mar keting strategies. The RFV (Recency, Frequency, Value) approach is used, allowing to trace the consumer profile based on their purchasing behavior. The K-means and DBSCAN algo rithms are applied for customer clustering. The effectiveness of these two algorithms in the clustering task is compared in an empirical study, which analyzes data from thousands of customers of an online retail company. The results show that the K-means algorithm outper formed DBSCAN in the clustering task for this study, being able to effectively distinguish between customer groups, which facilitates the construction of personalized strategies for each cluster. On the other hand, the DBSCAN algorithm clusters a large proportion of customers into a single cluster. These findings reinforce that the choice of the appropri ate algorithm is crucial for the effectiveness of customer segmentation and, consequently, marketing strategies. In addition, companies can use these findings to refine their marketing strategies, ensuring that offers are more effectively targeted to different customer segments.