Welcome to the InfoSci Platform
Reference Hub1
Applying Machine Learning to Study the Marketing Mix's Effectiveness in a Social Marketing Context: Fashion Brands' Twitter Activities in the Pandemic

Applying Machine Learning to Study the Marketing Mix's Effectiveness in a Social Marketing Context: Fashion Brands' Twitter Activities in the Pandemic

Sibei Xia, Chuanlan Liu
Copyright: © 2022 |Volume: 9 |Issue: 6 |Pages: 17
ISSN: 2334-4547|EISSN: 2334-4555|EISBN13: 9781668483732|DOI: 10.4018/IJBAN.313416
Cite Article Cite Article

MLA

Xia, Sibei, and Chuanlan Liu. "Applying Machine Learning to Study the Marketing Mix's Effectiveness in a Social Marketing Context: Fashion Brands' Twitter Activities in the Pandemic." IJBAN vol.9, no.6 2022: pp.1-17. http://doi.org/10.4018/IJBAN.313416

APA

Xia, S. & Liu, C. (2022). Applying Machine Learning to Study the Marketing Mix's Effectiveness in a Social Marketing Context: Fashion Brands' Twitter Activities in the Pandemic. International Journal of Business Analytics (IJBAN), 9(6), 1-17. http://doi.org/10.4018/IJBAN.313416

Chicago

Xia, Sibei, and Chuanlan Liu. "Applying Machine Learning to Study the Marketing Mix's Effectiveness in a Social Marketing Context: Fashion Brands' Twitter Activities in the Pandemic," International Journal of Business Analytics (IJBAN) 9, no.6: 1-17. http://doi.org/10.4018/IJBAN.313416

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

This study examines the effectiveness of the marketing mix practiced on Twitter across high-end and low-end fashion brands and explores whether any four Ps activities have changed across the different pandemic stages. A quantitative research method was designed to analyze text data scraped from identified fashion brands' Twitter accounts. A classification instrument was developed to group tweets based on the four Ps marketing mix. Then the developed instrument was applied to a small set of 145 tweets randomly sampled from the collected data. Logistic regression models were then trained using the sample set to predict four Ps activities on all the collected 144k tweets. The numbers of likes per tweet and frequencies of being retweeted per tweet were used to measure engagement effectiveness across brands.