Uso de machine learning para entendimento das relações de consumo no modelo de negócio (Saas Enabled Marketplace) do OLIST
Data
2022-07-23
Autores
Leite, Gabriela Amaral de Alencar 
Orientadores
Pinochet, Luis Hernan Contreras
Tipo
Trabalho de conclusão de curso
Título da Revista
ISSN da Revista
Título de Volume
Resumo
Com o crescente acesso a internet pela população brasileira, houve um expressivo crescimento da área do comércio eletrônico, com a expectativa de alavancar as vendas de empreendedores, que são um importante pilar da economia do país. Porém, há ainda uma dificuldade de inserção ou de alcance para empreendedores de menor porte, no que a startup brasileira Olist assume o papel de intermediar o lojista e os grandes marketplaces aumentando o alcance e automatizando processos. As técnicas de Machine Learning têm um importante papel atual na interpretação de padrões em um grande volume de dados, assim como no auxílio a solução de problemas complexos, entretanto, ainda há uma lacuna na adoção dos métodos para a tomada de decisão nas empresas. Portanto, o presente estudo tem como objetivo identificar os fatores que impactam no score dado pelo consumidor após a realização da compra via Olist Store, em algum marketplace, por meio da análise de dados e aplicação do modelo de Machine Learning para prever a nota baseada em CSAT. Como resultado, verificou-se que a variável expected_diff, que representa a diferença de dias entre a data que estava prevista para a chegada do produto e a data que realmente chegou, é a de maior relevância para todas as categorias, sendo mais importante ainda em beleza/saúde e brinquedos. Por fim, concluiu-se que é de suma importância que empresas que desejam obter vantagem competitiva no ramo do comércio eletrônico invistam em uma boa logística. O estudo teve como principal limitação a base do estudo, pois a que estava disponível não era a mais atual, não sendo assim possível de retratar as conjunturas pós pandemia da COVID-19.
With the growing access to the internet by the Brazilian population, there was a significant growth in e-commerce, with the expectation of leveraging sales from entrepreneurs, who are an important pillar of the country's economy. However, there is still a difficulty of insertion or reach for smaller entrepreneurs, in which the Brazilian startup Olist assumes the role of intermediating the shopkeeper and the large marketplaces, increasing reach, and automating processes. Machine Learning techniques have an important role today in the interpretation of patterns in a large volume of data, as well as in helping to solve complex problems, however, there is still a gap in the adoption of methods for decision making in companies. Therefore, the present study aims to identify the factors that impact the score given by the consumer after making the purchase via the Olist Store, in some marketplace, through data analysis and application of the Machine Learning model to predict the score based on CSAT As a result, it was found that the expected_diff variable, which represents the difference in days between the expected date of arrival of the product and the date that it actually arrived, is the most relevant for all categories, being even more important in beauty/health and toys. Finally, it was concluded that it is of paramount importance that companies that want to gain a competitive advantage in the field of e-commerce invest in good logistics. The main limitation of the study was the base of the study, as the one available was not the most current, thus not being possible to portray the post-COVID-19 pandemic conjunctures.
With the growing access to the internet by the Brazilian population, there was a significant growth in e-commerce, with the expectation of leveraging sales from entrepreneurs, who are an important pillar of the country's economy. However, there is still a difficulty of insertion or reach for smaller entrepreneurs, in which the Brazilian startup Olist assumes the role of intermediating the shopkeeper and the large marketplaces, increasing reach, and automating processes. Machine Learning techniques have an important role today in the interpretation of patterns in a large volume of data, as well as in helping to solve complex problems, however, there is still a gap in the adoption of methods for decision making in companies. Therefore, the present study aims to identify the factors that impact the score given by the consumer after making the purchase via the Olist Store, in some marketplace, through data analysis and application of the Machine Learning model to predict the score based on CSAT As a result, it was found that the expected_diff variable, which represents the difference in days between the expected date of arrival of the product and the date that it actually arrived, is the most relevant for all categories, being even more important in beauty/health and toys. Finally, it was concluded that it is of paramount importance that companies that want to gain a competitive advantage in the field of e-commerce invest in good logistics. The main limitation of the study was the base of the study, as the one available was not the most current, thus not being possible to portray the post-COVID-19 pandemic conjunctures.
Descrição
Citação
LEITE, Gabriela Amaral de Alencar. Uso de machine learning para entendimento das relações de consumo no modelo de negócio (Saas Enabled Marketplace) do OLIST. 2022. Trabalho de Conclusão de Curso (Bacharelado em Administração) - Universidade Federal de São Paulo, Escola Paulista de Política, Economia e Negócios, Osasco, 2022.