Publication: DISEÑO E IMPLEMENTACIÓN PARA UN SISTEMA RECOMENDADOR DE MARKETING BANCARIO
Date
2019-07
Authors
FERNÁNDEZ SOTO, EDUARDO ARIEL
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Una de las problemáticas que se afrontan en el contexto del marketing bancario, es como ofrecer adecuadamente los productos que se poseen, es decir, la realización de una estrategia ad-hoc enfocada en los requerimientos de cada cliente, sin incurrir en el desagradable marketing viral. En este ámbito, se busca generar un modelo, que formule políticas que ayuden en la toma de decisiones, al momento de ofrecer los productos que posee la institución, de manera que esto se realice de forma efectiva, con el objetivo de maximizar los beneficios asociados a las políticas generadas por el método utilizado. En base a esto, en esta memoria se propone abordar el problema anterior, diseñando e implementando un sistema recomendador que permita ofrecer productos enfocados en cada cliente. Dada las características específicas del negocio, es necesario considerar métodos alternativos a los sistemas recomendadores tradicionales para diversificar las soluciones de forma tal que se aprovechen las ventajas estratégicas que cada método a utilizar posee al momento de generar políticas de recomendación. Una de las principales problemáticas a considerar es la naturaleza cambiante en las condiciones económicas de cada persona, y los cambios de comportamiento que esto genera, lo que provoca que tanto el consumo y preferencias por cada producto, vayan variando en el tiempo. Por otra parte, se debe considerar la cardinalidad de usuarios en contraposición a los ítems ofrecidos, debido a que se posee una mayor cantidad de clientes que productos a ofrecer. Finalmente se espera que el sistema ayude a generar un modelo capaz de crear un mapa de viajes de clientes. Así mismo, ayudaría a definir una política de marketing que potencie la venta cruzada de productos acorde a los cambios en las preferencias futuras que tendrá un determinado cliente.
One of the problems that is faced in the area of bank marketing is how to adequetly offer the available products to clients, that is, the development of a specific strategy focused on the needs of each client without using the unpleasant viral marketing. To address this problem, it is useful to generate a model that formulates policies that help the bank to make decisions to offer products effectively in order to optimize the benefits of such institution. In this thesis, the problem described above will be addressed by using a recommender system that will offer specific products to each client. Due to the nature of the problem, it is necessary to consider alternative methods to the traditional recommender systems in order to diversify the solutions and to exploit the strategic advantages that each one posseses when recommending policies are generated. On the one hand, one important issue to be considered is the changing nature in the economic conditions of each person and the corresponding changes that these trigger in their behaviour. On the other hand, there is a larger amount of clients than products to be offered which is not the case in traditional recommender systems. Finally, it is expected that the recommender system will help to generate a model able to create a journey map. Additionally, the system will help to define a marketing policy to strengthen the cross-selling of products in relationship to the variations in future preferences of a specific client.
One of the problems that is faced in the area of bank marketing is how to adequetly offer the available products to clients, that is, the development of a specific strategy focused on the needs of each client without using the unpleasant viral marketing. To address this problem, it is useful to generate a model that formulates policies that help the bank to make decisions to offer products effectively in order to optimize the benefits of such institution. In this thesis, the problem described above will be addressed by using a recommender system that will offer specific products to each client. Due to the nature of the problem, it is necessary to consider alternative methods to the traditional recommender systems in order to diversify the solutions and to exploit the strategic advantages that each one posseses when recommending policies are generated. On the one hand, one important issue to be considered is the changing nature in the economic conditions of each person and the corresponding changes that these trigger in their behaviour. On the other hand, there is a larger amount of clients than products to be offered which is not the case in traditional recommender systems. Finally, it is expected that the recommender system will help to generate a model able to create a journey map. Additionally, the system will help to define a marketing policy to strengthen the cross-selling of products in relationship to the variations in future preferences of a specific client.
Description
Keywords
MARKETING , PLANIFICACION ESTRATEGICA , BANCOS Y OPERACIONES BANCARIAS