Predicting Brand Personality in Social Media Networks with Machine Learning (Coursework Sample)

Topic Review
Student Name
Course Name
Professor's Name
Date of Submission
Article Identification and Citation
This paper is a literature review from the Journal of Interactive Marketing article on 'A Brand New Look at You: Predicting Brand Personality in Social Media Networks with Machine Learning.' The article was published on 8th July 2021. It was written by Utku Pamuksuz, Joseph T. Yun, and Ashlee Humphreys.
Citation
Pamuksuz, U., Yun, J. T., & Humphreys, A. (2021). A Brand-New Look at You: Predicting Brand Personality in Social Media Networks with Machine Learning. Journal of Interactive Marketing, 56, 55-69.Available at: file:///C:/Users/USER/Downloads/1-s2.0-S1094996821000311-main.pdf (Accessed: 8th July 2021).
Introduction
Brand personality is a subtle feature of brands with compatible characteristics apart from its practical benefits. Brand personality has affected various factors of consumer behavior, such as social interactions and individual preferences. However, there is little research on the relationship between brand personality and social media networks. This article focuses on advanced methods of deriving brand personality from the content posted on social media. The report uses data analytics methods that can be used to trace efforts to substitute brand personality over time, assess conformity in brand personality, and measure competitors' brand personality. This article research aims to improve theoretical comprehension of streamed and recognized brand personality as constituted in social media platforms and offer practitioners the capacity to promote branding strategies using resources from big data. The research establishes a novel hybrid machine learning algorithmic design (LDA2Vec), which enhances coding tasks, thus providing a scalable and adaptable tool used for various management studies. Practitioners and marketers use this machine to analyze how brand personalities are perceived and channeled through social media.
This paper is an article review that summarizes predicting brand personality in social media as outlined by Pamuksuz et al. (2021). The documents summarize the article's main points, highlight the positives and facts represented in the publication, and critique the article by identifying gaps, contradictions, unanswered questions, and disparities.
Summary of the Article
The brand is one of the most treasured assets in a firm that improves performance and builds a strong relationship with consumers. Recently firms are expanding their efforts on growing social media platforms as an effective strategy for marketing and brand-building. Therefore, branding in social media has become a crucial way of marketing communication to pass on fundamental brand personality. A firm should have the ability to review both the perceived and intended brand personality via interaction with consumers to comprehend their brand personality in digital media platforms.
The firm has to use advanced analytics methods with extensive content static and customization to understand social data. Traditional methods of examining brand personality like Likert Scale surveys cannot cope with high content creation speed on social media. Also, the conventional techniques are expensive, time-consuming outdated, and might report partisan issues. Therefore, Pamuksuz et al. (2021) advocate for an automated machine learning approach and provide future machine advances to examine brand-consumer relationships in social media platforms. They combine supervised machine learning with secure vocabulary-based methods and an unsupervised open vocabulary approach to developing one rectified framework. The designed model takes the open-ended text data from social media platforms and reinstate the five brand personality dimensions in real-time. The dimensions offer a novel method of evaluating the content, increasing the scale and scope of brand research.
Human personality detection is achieved through evaluating content-based documents like Facebook posts, likes, shares, and essays. The researchers combine these content-based texts with other factors like gender and age to assess human personality. On the other hand, standard personality questionnaires and self-reported surveys are used for hypothesis testing and data collection to detect brand personality. Marketers use online social media platforms to engage consumers hence creating user-generated content that can detect brand personality. Data analytic is used in open vocabulary approaches like LDA, LSA to understand the phenomena in marketing. The method in this article uses hybrid LDA to gain benefits of receptive vocabulary approaches and use a closed vocabulary to link topics within the data. The tool is comprehensive in analyzing data and connects with other theoretical constructs in marketing.
The research collects data from Twitter and Facebook, where they post several selected brands, retrieve the post, and use machine-learned implementation to analyze and interpret the data. The findings indicated that the most competent companies had one of the following traits, i.e., intelligent, responsible, hard-working, reliable, or leader. The results showed that some brands like Burberry and Louis Vuitton presented their brands as exciting. Some brands were ranked as sincere, sophisticated, or rugged based on the word used by consumers to describe the brand.
Critics of the Article
Pamuksuzet al at., (2021) has demonstrated well that the considerable amount available online can be measured and systematically analyzed to bring consumers' perception about the brand. Researchers can use this methodology to test future propositions since the approach used to detect brand personality is not limited to dimensions or keywords. According to Aaker (1997), brand personality is a set of various human traits accompanied by the brand. This research reflects both human personality and brand personality aspects within social media, allowing marketers to establish the relationship between consumers and brands. The model can help practitioners monitor and evaluate social media data over several years and explore the impacts of changes within a firm or brand personality over time. The approach also serves as a crucial metric to evaluate how their brand is conveyed through social media platforms.
However, this research has some gaps and limitations that need to be addressed. First, their automated method does not consider consumers and brands not active on social media platforms. The research is only limited to brands and consumers who maintain a presence on Facebook or Twitter. Therefore, such brands can apply this model in their marketing strategies. In addition, this approach only relies on the text data provided by the brand. According Dörner & Edelman (2015), to identify the actual context of the posts in social media, a fir