Abstract
This article systematically applies machine learning and data envelopment analysis (DEA) to analyze Twitter messages, Twitter metrics, and organizational financial metrics to gain insights into impactful messaging typology on social media network. Automated machine learning is employed for the classification of tweets of select US Furniture Retail Stores while various DEA models are utilized to analyze multiple input metrics to obtain an efficiency ranking for the selected brands. Based on these analyses, the article discusses the implications of the findings for small and medium-sized enterprise marketing managers at the industry level. Recommendations for industry practice are also provided in addition to the directions regarding future research.
INTRODUCTION
Awell-designed social media strategy that creates and curates engaging content helps organizations maintain competitiveness in today’s dynamic business landscape [1]–[3]. Conventional marketing activities and strategies fall short in addressing specific requirements for efficient and effective social media marketing (SMM) operations because it seeks to sell through multiple means [4], in contrast to SMM which focuses on making individual connections and building long-lasting relationships. SMM customizes and personalizes marketing for a more meaningful and targeted marketing option [1].
CONCLUSION
While this article integrates various disciplines including data mining, social media marketing, and performance evaluation—there are limitations and opportunities for further research. This research was based on limited data from a single Social Media Network, Twitter. Twitter is a good case example for a proof of concept; but it provides only a certain type of information. Further research would benefit from including a larger dataset and other competing Social Media Networks.
Overall, given the pervasiveness and growth of SMM, more and better methodologies are needed for managerial and research investigations. This article adds to the discourse and helps build additional foundation for further investigation.
