MULTICRITERIA METHODS ANALYSIS AND CONSUMER SATISFACTION SCORE: A CONSENSUS ANALYSIS ON BRAZILIAN E-COMMERCE
This paper aimed to analyze the odds ratio between the results obtained by the multicriteria methods (Adriana, DP2, and EDAS) and the consensus of the consumer review score in e-commerce. This study was conducted through archival research and is characterized as experimental. The data refers to the Brazilian E-Commerce Public Dataset available by Olist Store at the online community of data scientists and machine learners – Kaggle. The database contains records of the year 2016 to 2018 made at multiple marketplaces in Brazil. For data analysis, we used a machine learning technique and logistic regression models. Logistic regression makes it possible to analyze the odds ratio of the occurrence of an event about the consumer review. Both the multi-criteria methods (Adriana, DP2, and EDAS) and the consensus analysis of the responses can allow insights into the financial information of companies. Therefore, these results highlight the importance of analyzing the consensus of consumer reviews in addition to the managerial processes that can contribute even more to improving the processes involved. The results observe the relevance of not breaking consumer confidence regarding the time processes estimated, as this fact directly impacts the review score. The consensus analysis presented that, in addition to seeking a high average in consumer reviews, managers must observe the consensus on reviews, so that inconsistencies can be reviewed.
Keywords: Multicriteria methods; Consumer satisfaction; Consensus analyses, E-commerce.
