Data science and marketing in e-commerce amid COVID-19 pandemic

Abstract

PURPOSE: The objective of this study involves the determination of data-driven solutions needed to increase the usability of e-commerce systems and its profitability.

DESIGN/METHODOLOGY/APPROACH: In the research implementation process, logic generalization and induction to identify and analyze the most beneficial data science tools in e-commerce. deign of the study is to generalize existing approaches of data science usage in e-commerce, to develop practical recommendations to ensure the competitive advantages of e-commerce market participants and to estimate the cost of technical tools needed to launch the data science project in e-commerce.

FINDINGS: The results clearly demonstrate that in 2020 businesses that have e-commerce system were financially successful and in next 3 years online sales will increase rapidly. The simple analytics will not cover the demand of online business and it is needed to implement advanced data-driven decisions now.

PRACTICAL IMPLICATIONS: The present research provides generalized knowledge on how to launch a data science project in e-commerce and how to choose the best programming and visualization app to ensure the profitability of a project. The scientific paper gives an instruction on the marketing contribution analysis, which is the tool of key importance for online marketplaces.

ORIGINALITY/VALUE: The main research value drawn from the study is to launch the datadriven models in e-commerce company it is needed to observe the real business need and available data, find the best programming and visualization tools. It was defined that the most beneficial data science solutions are demand forecasting, estimation of the marketing contribution, customers clustering, recommendation system and customers’ attitude analysis. The main business need for each e-commerce company is to estimate the contribution of all marketing channels and advertisement formats separately. This issue may be easily handled with a regression modelling, which helps to understand a set of factors influencing sales.