Coping with demand volatility in retail pharmacies with the aid of big data exploration

Abstract

Data management tools and analytics have provided managers with the opportunity to contemplate inventory performance as an ongoing activity by no longer examining only data agglomerated from ERP systems, but also, considering internet information derived from customers’ online buying behaviour. The realisation of this complex relationship has increased interest in business intelligence through data and text mining of structured, semi-structured and unstructured data, commonly referred to as “big data” to uncover underlying patterns which might explain customer behaviour and improve the response to demand volatility. This paper explores how sales structured data can be used in conjunction with non-structured customer data to improve inventory management either in terms of forecasting or treating some inventory as “top-selling” based on specific customer tendency to acquire more information through the internet. A medical condition is considered - namely pain - by examining 129 weeks of sales data regarding analgesics and information seeking data by customers through Google, online newspapers and YouTube. In order to facilitate our study we consider a VARX model with non-structured data as exogenous to obtain the best estimation and we perform tests against several univariate models in terms of best fit performance and forecasting.