Multivariate Sales Forecast Model Towards Trend Shifting During COVID-19 Pandemic A Case Study in Global Beauty Industry

ABSTRACT

COVID-19 pandemic has changed the economic weather and business performance in multiple streams. The uncertainty condition caused by the pandemic needs to be carefully taken care by all companies and organizations due to rapid consumer trend shifting and volatile market condition. The sales and marketing strategy needs to be carefully taken during organizational decision-making process to avoid further loss. PT XYZ as one of the leading consumer goods in beauty industry experiences the same condition and challenge reflected by down-trend in the organization KPI. This research aims to introduce and provide predictive data analytics tools for enhancing sales forecast by comparing Random Forest and Neural Network as part of machine learning methods also Vector Autoregression (VAR) as conventional statistical forecasting methodology. As the result of this research, neural network returns better evaluation for skin care and Vector Autoregression for makeup category. Meanwhile data visualization is found necessary to provide additional factual information, includes the external factor, to support knowledge management for better rational decision-making process. © 2022 ACM.