Estudo de métodos para avaliar as variáveis representativas nas publicações de uma empresa varejista

Abstract

With the event of the COVID-19 pandemic, the need to accelerate the digital transformation in retail companies was realized, especially the smaller ones, which ended up suffering more from the changes imposed during this period. An interesting way to increase the efficiency of social networks is through the analysis of the metrics of the posts made, hence the importance of developing a process that helps them in this regard. Therefore, this work proposed to study the different methods of design of experiments (DOE), in order to identify the most appropriate method to assess the variables that significantly influence the engagement of publications made through the Instagram® platform. It was noticed that the 2k-p fractional factorial is the most suitable method for this study, so it was the target of the work. The methodology used is based on the seven stages of planning experiments defined by (MONTGOMERY, 2017). A complete plan (including an Experimental Matrix) was obtained to conduct the experiment, respecting the limitations of the application in question and its particularities. Finally, it is concluded that the plan developed here extends to areas other than the one mentioned above, such as a business consulting profile or even digital influencers. Allowing and aiding the application of statistical methods and data analysis in decision making when managing a social network.