The use of recommender and decision support systems for sales personalization in a mobile application

In the process of shopping, users are today faced with a large volume of information and a broad range of products and services that prevent them from being able to make rational decisions regarding the purchase of those products and services they actually require at a particular time and place and which meet their preferences, interests and needs. By defining and confirming this problem faced by users, we began with the analysis, design, development, testing and implementation of an information and recommendation system for the personalization of sales. This information system operates on the basis of a business model, where in exchange for providing important feedback, the user receives special offers or loyalty points. A lack of qualitative data about customers, their habits, future purchases and past experiences is one of the key factors in preventing companies from implementing effective personalization. Thus, even in real time, companies lack answers to important questions that concern marketing, sales and business operations. With the assistance of recommendation and decision making systems and by processing large amounts of smart data, we can offer the customer personalized products and services and thereby accelerate and increase sales volume while simultaneously improving the user and shopping experience. In the analysis and development of the information and recommendation system, we developed a hypothesis which proposed that with the use of qualitative data on user desires, needs, past experiences and future purchases, we could offer the user more personalized special offers. Personalization will also enable an increase of the CTR (Click to Rate) conversion between views of special offers and relevant responses, or rather, the execution of sales campaigns. On the basis of the developed and tested recommendation system, we conclude that the most appropriate solution for our purposes is the use of hybrid recommendation techniques which, depending on different types of situations, implement either the CF or CB method of filtering in combination with other decision rules and conditions