“According to sample online social data, what are some ways that the COVID-19 crisis may affect a particular brand?”
Especially in view of the last two questions, as of March 2020, the keywords ‘panic buying’ and ‘panic shopping’ and their derivatives become significant search terms and hashtags on Twitter (Chen, 2020). After analysing social listening data for a case study, aspects such as the possible causes, reactions, advantages, and drawbacks can be discussed to better understand the business context. This research also proposes strategies for different business scenarios over time to take advantage of the social listening applications and data for companies.
  1. Literature Review and Theoretical Perspectives: Social Listening Concepts
A definition of social listening is given in Anderson et al (2017), whereby it is described as a process to identify and analyse big picture information about a company, product, brand, or individual from online social media. By contrast, social monitoring is an endeavour that focuses on the daily interaction and the details (Cuttica, 2016). Crawford (2009) gave a broader description by categorising social listening to three types, which are background listening, reciprocal listening, and delegated listening. These can involve different listeners such as individuals, politicians, and corporations. Background listening is a method that generally scans the social media, where content continuously flows in the background, without an intention to interfere (Crawford, 2009). Reciprocal listening is a method that uses a broadcast approach to respond to the comments and direct messages in social media, which is commonly used in politics. The delegated listening conducted by corporations, enables others listeners to take part and observes the participants and data in a controllable manner. Likewise, Paine (2011) emphasises that companies should focus on their content and the publicly available data about the customers rather than exhaust their own resources trying to monitor everything.

Social Listening Factors and Benefits

Social listening can bring multiple advantages to companies. According to Crawford, companies can have three benefits from social listening: hearing opinions, leveraging customer support, and increasing the brand’s influence (2009). According to Patrick (2017), Clutch surveyed 300 social listening software users who worked at companies with more than 100 employees. This survey indicated that improving products and services, attracting new customers, and improving customer service are considered the most beneficial aspects to companies (respectively 25%, 24%, and 21% of responses). Also, the survey indicated other benefits of social listening, such as monitoring content performance, recruiting and hiring new employees, and learning about the competition. In practice, 42% of the participating companies used these tools to improve the relationship with their customers by adopting various social listening strategies. The data from Clutch’s 2017 social listening survey also showed that 86% of the participants monitor customer requests, questions, and concerns, while 77% monitor their competition.
In the contexts of the company brand, a company’s interests are likely to be greatly impacted by any snowball effects that may begin with a single post. Identifying and coping with a crisis is vital for companies (Paine, 2011); Magdalena Urbaniak, the brand manager of “Brand24” company, said“With a social media monitoring tool, we are able to respond first and nip a potential social media crisis in the bud before it can escalate.”
In regards to social listening for market competition, brands and products need to be positioned flexibly and be able to respond and thrive in the market (Paine, 2011). There are so-called hot buttons or factors and different dimensions such as price, value, delivery, and variability. The positioning is to find what strategic approach a company might take, e.g. being a leader, follower, or laggard. For example, a company can listen to the comment of their rivals’ customers for what drives their consumption decisions and then make their own marketing strategies (Paine, 2011).

Social Listening Measurement

Measurement is a critical part of social listening. From the perspective of business functions, social listening can be applied to public relations, marketing, and advertising fields (Paine, 2011). The data for measurement can be based on public data, such as the number of mentions and followers, or a group of intelligence systems (Kietzmann et al., 2011). Optional metrics for measuring social listening mentions are strength, sentiment, passion, and reach (Kietzmann et al., 2011). The strength metric is the number of times a company or product are mentioned in the selected channel. The sentiment metric is the proportion of mentions i.e. positive versus negative. The passion metric is the frequency in which certain users mention a company or a product. The reach metric is the distinct users in mentions divided by the strength metrics (Kietzmann et al., 2011).
As a particular metric, sentiment measurement is a media-content-analysis-based method to understand how the image of a company projects in the outside environment (Paine, 2011). Through media content analysis, companies can try to measure the positivity of their image in the media and better understand their audience. Paine (2011) points out that there are several key aspects of sentiment measurement. One is the main subject for the sentiment. The second is the type of the measured item, such as opinion, feature story, Q&A, and customer feedback. The third is the visibility of companies in the measured item, which means whether the images of companies can be identified within public content. The fourth is the people who make the posts. The fifth is the tonality, which means whether a reader seems to have a conflict with the company. The sixth is the type of media in which the item appeared. The seventh is the key message in the item. The eighth is how customer satisfaction and responsiveness to customers appear in the content.
Social media software and data are quite accessible online to users around the world; therefore, technology transfer and exchange of new ideas can take place (Erturk, 2009). For social media sentiment listening, manual analysis may be impractical due to the enormous amount of data (Schweidel & Moe, 2014). By contrast, software-automated sentiment analysis is a viable way, both faster and easier (Paine, 2011). Also, Hopkins and King (2010) suggest a hybrid sentiment analysis by combing manual analysis and software. For example, in many cases, a human can detect the sentiments toward the price of a product which automated analysis cannot detect. Furthermore, most systems cannot identify the irony and sarcasm in a common conversation, which causes the vast majority of automatic sentiment analysis systems to only detect correctly at a rate of 50% or less (Paine, 2011).

Methodology

This research adopts a case study methodology. Since this research focuses on the strategies of social listening and monitoring practices, looking at multiple snapshots in time has its advantage of conducting in-depth research by seeing variations under different scenarios.
In terms of the data sources, this research aims to find data from Facebook, Twitter, and YouTube, which are mainstream social media. However, for those users who comment a company or product using multiple accounts across different channels, the results of calculating mentions may seem duplicated, i.e. they may overstate certain outcomes such as consumer sentiments. However, this research advocates a multi-channel approach review the influences of brands and trends. For 2020, during the period influenced by COVID, the paper uses data and keywords mainly from Google Trends as well as Wordtracker.com.
As to the social listening tool, this paper adopts SentiOne as the social listening tool. SentiOne is a professional social listening tool which can be used for monitoring and analysing various social media in 26 languages and can detect trends (Kucharska, Brunetti, Confente, & Mladenović, 2018). In this research, SentiOne is leveraged to perform the data gathering and present results in a dashboard. In addition, this research has used other tools such as Microsoft Power BI to visualise some of the findings.
Pak’nSave is a well-known discount food warehouse chain in New Zealand. This paper studies social listening or monitoring, including brand image reviewing and crisis detecting. Pak’nSave is a good example of a company influenced by social media trends.
The first step of conducting a social listening is keyword selection. For a company or a product, there may be probably multiple names or spellings. For example, the company registered name is the Pak’nSave, while on the Internet, users are likely to type it in other ways, as a result of typos, or easier ways of typing, or naturally-pronunciations, such as “Pak n save”,” pack and save”, or “Pak and save”. To find appropriate search keywords, this paper leverages Google AdWords, which provides keyword suggestions based on the target market. Table 1 is the search criteria of the subjects. The table lists major related names for the brands or products in this research. The AdWords tools calculated 12 months’ search results and provided the times of search for relevant keywords which are organised and listed for a “keyword ideas”. This research manually filtered out irrelevant keywords and focused on the most relevant brands and product synonyms and selected the most-searched keywords in the keyword ideas list.
Table 1: Keyword searches for three scenarios