Crisis Communication on Social Media: What Types of COVID-19 Messages Get the Attention?
Abstract
This study assessed internet users’ attention to hospitality companies’ COVID-19 messages on social media. We used 657 Facebook and 754 Twitter messages initiated by eight of the world’s largest hotel chains between January and mid-June 2020 for the exploratory analysis. Under the situational crisis communication theory, the analysis reveals that hotels shared five types of COVID-19 (Prevention, Reminding, Ingratiation, Victimage, and Updates) versus Non-COVID-19 messages. Descriptive analysis and a series of t test, analysis of variance, and post hoc analyses reveal that hotels did not share any COVID-19 information until March 2020. Moreover, COVID-19 messages only accounted for about 20% of all messages, among which hotels shared Ingratiation and Updates messages most often. COVID-19 messages received more reactions, comments, and shares/retweets than Non-COVID-19 messages on both Facebook and Twitter, indicating the attention paid to the COVID-19 messages posted on a business’ social media page, which can help businesses spread the information in their networks. Specifically, Prevention, Reminding (although underused), Ingratiation, Updates, and messages with photos and videos received more attention. Such findings extend the crisis communication literature and help businesses develop effective communication strategies to engage their stakeholders on social media during the pandemic.
Proper crisis communication is critical to keep stakeholders informed, protected, and anticipated. Well-prepared and designed crisis communication strategies can promptly and effectively communicate with one another during a crisis that allows organizations to protect employees and customers and ensure business continuity.
Introduction
The World Health Organization (WHO) officially announced the widespread of COVID-19 a global pandemic on March 11, 2020. Several countries and areas worldwide became the epicenters of the pandemic and have since taken various measures, such as lockdowns, stay-at-home order, mobility and travel restrictions, and massive quarantine, to prevent the further spread of COVID-19 (Fernandes, 2020). People are urged to avoid having direct contact with others, and non-essential workers1 were asked to work from home. Because people are staying at home and most of the regular business activities are put into a pause mode, the global supply chains are disrupted, and the travel and tourism industry collapsed (van Barneveld et al., 2020). COVID-19 brought a decimated effect on the travel and hospitality industry and is, by far, a “once-in-a-generation” crisis that is expected to create long-term economic and societal consequences (Krishnamurthy, 2020).
At the organizational level, a crisis is referred to as “an event that is perceived by the managers and stakeholders to be highly salient, unexpected, and potentially disruptive” (Bundy et al., 2017, p. 1663). Although crises or disasters often bring unpredictable catastrophic changes to the tourism industry, effective management of crises can reduce the events’ negative impacts and shorten the recovery time (Mair et al., 2016; Ritchie & Jiang, 2019). Research that helps travel and hospitality companies identify the effective strategies to cope with the COVID-19 crisis is timely and can support a rebound of the industry. Organizational research has long been interested in crisis management. Through an in-depth conceptual literature review, Bundy et al. (2017) summarized the primary management research in a two-dimensional framework: internal versus external perspective and the three stages of the crisis, namely, pre-crisis prevention, crisis management, and post-crisis outcomes. Although a great deal of research concerning the external perspective of crisis management focuses on how stakeholders perceive and react to a crisis and how organizations influence those perceptions, Bundy et al. (2017) found no studies that have ever “systematically explored the real-time disclosure and information exchange that occurs between an organization and its stakeholders” (p. 1682). Social media have the hybrid elements wherein companies can control the content, timing, and frequency of the information being shared with consumers and, simultaneously, consumers can carry an open conversation about a brand or a product in a real-time setting (Kwok & Yu, 2013). These elements enable researchers to fill in such a gap through a systematic analysis of the real-time information exchange among an organization and its stakeholders as a crisis evolves. Within the hospitality and tourism industry, social media remain underused in crisis communication (Ritchie & Jiang, 2019). Research on hotels’ crisis communication strategies during the COVID-19 pandemic is timely and adds to the scattered, relevant research efforts in hospitality and tourism literature (e.g., Möller et al., 2018; Zhai et al., 2019).
Hence, this study examines how hospitality companies communicate with their stakeholders in the real-time social media setting during the COVID-19 crisis and how their stakeholders react to the companies’ messages. Different from the current literature (Möller et al., 2018; Zhai et al., 2019), this study’s investigation is framed under the situational crisis communication theory (SCCT) to analyze 657 Facebook posts and 754 Twitter messages that were initiated by eight hotel chains between January 1, 2020, and June 19, 2020. SCCT was introduced to help businesses select effective strategies in crisis communication (Roshan et al., 2016). SCCT suggests that organizations should choose the type of crisis communication strategy based on the level of responsibility that the public attributes to the organization (Coombs & Holladay, 2002). As such, Coombs (2014) proposed 10 types of communication strategies under SCCT, which have been widely used in relevant empirical studies (e.g., Ki & Nekmat, 2014; Park, 2017). Referring to Coombs’ (2014) typology of SCCT communication strategies, we first coded the eight hotel chains’ social media messages into categories. Then, we assessed the hospitality companies’ crisis communication strategies by identifying what types of messages were shared most often by these hotel chains. We further analyzed the attention/reactions given by these hotel chains’ stakeholders on various types of COVID-19 crisis communication messages. Our findings extend SCCT and can help companies develop effective crisis communication strategies on social media.
The Research Background and Relevant Literature
Relevant Research on COVID-19
COVID-19’s devastating effects drew immediate attention from academia in various disciplines. First, a good number of conceptual reviews emerged in business, tourism, and hospitality literature, where researchers pointed out directions for future research on COVID-19 and identified possible solutions from existing literature to help organizations cope with such an unprecedented crisis. For example, Bapuji, de Bakker, et al. (2020) described the COVID-19 pandemic as an unprecedented crisis that would leave a long-term impact on businesses and society, which requires collaboration between organizations and a diverse set of stakeholders in responding to the pandemic. The authors expected three research themes to arise for business and society scholarship: lessons learned from the crisis, assessment of organizational responses, and organizational designs in the future. In another conceptual review, Bapuji, Patel, et al. (2020) proposed how organizational practices in corporate social responsibility (CSR), work design, recruitment, and selection, and compensation management may contribute to the normalization, reinforcement, and reduction of economic inequalities. Besides, relevant literature discussed the impact of COVID-19 on marketing philosophy and CSR (H. He & Harris, 2020), market-shaping strategies (Nenonen & Storbacka, 2020), retail business (Pantano et al., 2020), equal and sustainable society (van Barneveld et al., 2020), international business (Sharma et al., 2020), business education (Krishnamurphy, 2020), human rights of participating in hospitality and tourism activities (Baum & Hai, 2020), the transformation of tourism (Brouder, 2020), and tourism in the future (Higgins-Desbiolles, 2020).
Besides the inductive inquiries that examined the COVID-19 phenomenon through the lenses of business owners or executives (e.g., Kraus et al., 2020; Ritter & Pedersen, 2020), a few studies have adopted statistical modeling and survey methods in their inquiries. For example, Chinazzi et al. (2020) used a global metapopulation disease transmission model to estimate the impact of travel restrictions on the spread of the COVID-19 pandemic. Yang et al. (2020) used a dynamic stochastic general equilibrium model to estimate the coronavirus impacts on the tourism industry. Meanwhile, Gössling et al. (2021) compared the impact of COVID-19 on the global tourism industry with that of other pandemics in history.
Overall, the devastating impact of COVID-19 on the business world has been well recognized in the literature. Looking forward, we echo Tracey’s (2020) viewpoint about the future of hospitality research in the post-COVID-19 era, where more considered and creative inquiries are needed to address the problems faced by the industry. We then framed our study under SCCT to assess hospitality companies’ crisis communication strategies during the COVID-19 pandemic. Our work addresses an issue that has not been discussed in COVID-19 literature and falls into the area of assessing organizational responses to the COVID-19 crisis, one of the three key research streams recommended by Bapuji, de Bakker, et al. (2020).
Crisis Communication and the Research Questions
Crises are harmful to organizations and their stakeholders who might have conflicting interests and needs. Crisis management is the organizational actions and communication that are expected to lower the likelihood of a crisis, minimize the harm from a crisis, and re-establish orders after a crisis (Bundy et al., 2017). As a result, organizational research regarding crisis management usually addresses one or more of the event’s three areas/stages: (a) pre-crisis preparedness, (b) proposal for handling midst-crisis situations, and (c) post-crisis strategies to restore the corporate image (Omilion-Hodges & McClain, 2016). Then, from the external and internal perspectives, crisis management literature often involves the shaping of stakeholder perceptions during the three stages of a crisis, in addition to the leadership efforts of managing organizational dynamics and learning (Bundy et al., 2017). Bundy and Pfarrer (2015) explicated the socio-cognitive mechanism of how organizations and their external stakeholders perceive and respond to a crisis. That is, external stakeholders usually rely on two sources of information to evaluate a crisis and its attributions: the perceived responsibility of an organization in the crisis and the communication strategies adopted by the organization. Consequently, a great deal of research effort was put into studying crisis communication strategies, through which many typologies have also been introduced to capture organizational responses (Bundy et al., 2017).
For instance, Coombs (1995, 2007) drew from the logic of attribution theory and introduced the SCCT, which posits that the threat caused by a crisis on an organization depends on the level of responsibility that the public attributes to the organization (Coombs, 2007). SCCT was introduced to help organizations select the appropriate communication strategies and thus minimize a crisis’s threat to their reputations (Roshan et al., 2016). According to SCCT, crises can be grouped into three categories: victim, accidental, and preventable clusters (Ma & Zhan, 2016). Within the victim clusters, the public attributes a low amount of responsibility to an organization when the organization could also be a victim of a crisis, such as in a natural disaster (Coombs, 2007). Meanwhile, the public attributes a minimal amount of responsibility to an organization for accidental crises, which are unintentionally caused by the organization’s actions (Coombs, 2007). Preventable clusters include situations when an organization is fully aware of the risks and consequences of their actions resulting in a crisis, such as human-error accidents; the public will attribute a considerable amount of responsibility to a preventable crisis (Coombs, 2007). Although reputation-attribution is a more straightforward psychological process, crisis communication strategies are more complicated but can influence cognitions and help lessen a crisis’s negative impacts on an organization and its stakeholders (Coombs, 2016). Organizations are highly recommended to choose the type of crisis communication strategy based on the level of responsibility that the public attributes to the organization (Coombs & Holladay, 2002).
Coombs (2014) summarized 10 SCCT communication strategies under four crisis response strategy clusters. Table 1 provides the descriptions of the 10 SCCT communication strategies. Businesses usually adopt different types of SCCT communication strategies, depending on their belongingness to the victim, accidental, or preventable clusters (Roshan et al., 2016). For example, organizations may choose the strategies in the denial posture to avoid any connections with a crisis. They may also use the strategies in the diminishment cluster to lower a crisis’s negative effects. Rebuilding posture allows organizations to re-establish relationships with the stakeholder through apologies or compensations. Strategies in the bolstering posture enable organizations to build a positive relationship with the stakeholders. A large number of empirical studies used SCCT to categorize organizations’ communication strategies for analysis (e.g., Ki & Nekmat, 2014; Park, 2017; Roshan et al., 2016). In the situation when COVID-19 negatively affects all stakeholders of the travel and hospitality industry substantially, including businesses that provide services and consumers who receive the services, every stakeholder is a victim of the crisis. According to SCCT (Coombs 2014; Coombs & Holladay, 2002), because travel and hospitality businesses held no responsibility for the crisis, they will likely use the bolstering strategies, including Reminding, Ingratiation, and Victimage, more often than others to build a positive relationship with their stakeholders. It is unlikely for the victims to be in denial or apologize for the global pandemic.
| Crisis Response Cluster | Purpose/Description |
|---|---|
| Denial posture (cluster) | |
| 1. Attack the accuser | To confront those claiming that the crisis exists |
| 2. Denial | To deny the existence of the crisis |
| 3. Scapegoating | To blame someone outside the organization for the crisis |
| Diminishment posture | |
| 1. Excusing | To minimize the organization’s responsibility for the crisis |
| 2. Justification | To minimize the perceived damage associated with the crisis |
| Rebuilding posture | |
| 1. Compensation | To compensate the victims |
| 2. Apology | To apologize for the crisis |
| Bolstering posture | |
| 1. Reminding | To remind stakeholders about the organization’s past good work |
| 2. Ingratiation | To praise the stakeholders |
| 3. Victimage | To explain how the organization is a victim of the crisis |
Source. Adopted from Coombs (2014) and Roshan et al. (2016).
Today, social media has become a critical component of people’s life (Utz et al., 2013). Various hospitality and tourism businesses have also used social media as an essential tool for communications, promotions, marketing, brand management, service recovery, research, and other managerial implications (Mehraliyev et al., 2020). During the COVID-19 pandemic, people work remotely at home, and classes are taught online. Social media’s role in crisis communication can become even more critical for hospitality and tourism organizations. Nevertheless, social media is still underused in crisis communication by the hospitality and tourism industry (Ritchie & Jiang, 2019). Möller et al. (2018) assessed how hotels used Facebook in crisis communication following the tropical cyclone in Fiji. They also concluded that social media was underused before the natural disaster and that Facebook could be useful in helping hotels raise funds and donations in the recovery phase. Furthermore, a destination marketing organization’s positive, detailed, and timely responses during the tourism-related human-induced preventable crisis can effectively mitigate the public negation emotions on social media toward the destination (Zhai et al., 2019). The above scattered, relevant literature suggests that social media can be an effective tool in crisis communication. Studies on corporate crisis communication strategies in social media, while currently being underused and possibly under-researched, are timely and likely to help hospitality and tourism organizations effectively engage their stakeholders online. Moreover, social media captures real-time conversations among the stakeholders of a business; thus, researchers in crisis management can go beyond the quasi-experimental design and study people’s real reactions toward different types of messages (Spence et al., 2016).
Research in business communications reveals that different types of business-to-consumer (B2C) messages can have various impacts on internet users’ attention. For example, updates with photos instead of videos, hyperlinks, or pure text can receive more “Like” reactions or more comments from internet users on Facebook (e.g., Chao et al., 2014; Kwok et al., 2015; Kwok & Yu, 2013; Sabate et al., 2014) and more “Like” reactions or retweets on Twitter (e.g., Araujo et al., 2015; Li & Xie, 2020; Soboleva et al., 2017). Furthermore, if the B2C messages are categorized by content, such as sales/marketing messages (e.g., promoting a brand, a product, or a service offered by an organization) versus conversational messages (e.g., talking about the weather or other issues that are not associated with the organization), conversational messages can also receive more “Like” reactions or comments among Facebook users (Kwok & Yu, 2013, 2016)
Through a review of SCCT and the above empirical studies on B2C communications, we expect that hospitality businesses will use three SCCT strategies, namely Reminding (telling stakeholders the organization’s past good work), Ingratiation (praising the stakeholders), and Victimage (reminding the stakeholders that the organization is also a victim), more often than other SCCT strategies if they also see themselves as a victim of the COVID-19 crisis. We also speculate that these companies will share updates with photos rather than other media types since early research has shown that messages with photos tend to receive more attention from internet users. Last but not least, different message types, such as different SCCT messages and messages in different media types, may have different effects in getting internet users’ attention. Hence, we advance the following four specific research questions to guide our analysis:
Research Question 1 (RQ1): Do hospitality companies also use the strategies of Reminding, Ingratiation, and Victimage to communicate with their stakeholders about the COVID-19 crisis?
Research Question 2 (RQ2): Do hospitality companies share photos more frequently than other media types?
Research Question 3 (RQ3): What are the internet users’ reactions toward hospitality companies’ COVID-19 messages compared with other non-COVID-19 messages?
Research Question 4 (RQ4): What are the internet users’ reactions to hospitality companies’ social media messages in different media types?
By answering the above research questions, this study makes contributions to crisis management literature regarding the real-time interactions between an organization and its stakeholders (Bundy et al., 2017) and the incorporation of social media in crisis communication (Möller et al., 2018; Utz et al., 2013; Veil et al., 2011; Zhai et al., 2019). Practically, the findings can inform hospitality organizations how to communicate with their stakeholders during the COVID-19 crisis effectively.
Method
We answered the research questions using the publicly accessible data on Facebook and Twitter, enabling us to capture both the messages posted on the social media sites and the stakeholders’ real reactions toward the messages in the real-world setting. In 2019, 69% and 22% of U.S. adults reported that they used Facebook and Twitter, respectively (Perrin & Anderson, 2019). Moreover, companies often share similar content across different social media platforms (Kwok & Yu, 2013); therefore, insights drawing from Facebook and Twitter are applicable and highly relevant to crisis communication.
The Research Setting
We focused our analysis on the top eight hotel chains globally, as measured by the number of room counts in the market. As shown in Table 2, these eight hotel chains operated more than 5.2 million rooms as of June 2019. Marriott was the biggest hotel chain and had the highest number of followers on Twitter, and Best Western had the most likes and most followers on Facebook. We expect such a sample will allow us to capture a broad aspect of what a hospitality business usually entails because many full-service hotels also run various food and beverage outlets.
| Hotel Chain | No. of Roomsa | No. of People Liking the FB Pageb | No. of People Following the FB Pageb | Twitter Followersb | Twitter Followingb |
|---|---|---|---|---|---|
| Marriott | 1,332,826 | 418,938 | 426,460 | 430,700 | 426 |
| Hiltonc | 939,297 | 534,882 | 532,785 | 159,000 | 165 |
| IHG | 842,759 | 254,837 | 262,574 | 99,000 | 2,623 |
| Wyndham | 809,903 | 117,857 | 117,582 | 8,180 | 202 |
| Choice Hotels | 572,659 | 422,544 | 420,658 | 47,900 | 3,868 |
| Best Western | 371,871 | 1,162,645 | 1,117,705 | 85,100 | 3,937 |
| Hyatt | 213,173 | 678,250 | 679,720 | 78,400 | 3,517 |
| Radisson | 181,185 | 132,263 | 132,893 | 42,000 | 637 |
| Total | 5,263,673 | 3,722,216d | 3,690,377d | 950,280d | 15,375d |
a
As of June 2019; Reported at Statista.com via https://www.statista.com/statistics/197859/total-number-of-guestrooms-of-us-hotel-companies-worldwide. b The numbers were recorded on June 17, 2020, where Facebook provides a specific number. For accounts with a large number of followers, Twitter uses “K” denoting thousands. c The Twitter page of Hilton Newsroom instead of Hilton was used because of its large number of followers. d Total numbers without considering the fact that some people may follow multiple accounts.
Data Collection
We used a data-crawling program on June 24, 2020, to automatically retrieve the publicly accessible data from the eight hotel chains’ Facebook pages and Twitter accounts. Data crawling is a big data collection technique that has been widely used in similar studies that also analyzed the public data on the internet (e.g., W. He et al., 2013; Kwok & Xie, 2018, 2019). Facebook allows researchers to access its API (Application Programming Interface) through the FORT (Facebook Open Research & Transparency) platform and then publish research using the data on Facebook pages without Facebook’s permission.2 Likewise, Twitter allows researchers to assess and analyze the public data on Twitter through the Twitter API, such as doing research for scholarly purposes that may include work published in academic/scientific journals, news, and press.3
We focused on analyzing the messages shared between January 1 and June 19, instead of June 24, 2020, because (a) the first coronavirus was reported in China in December 2019 and quickly spread to other countries, and (b) it usually takes several days for the internet users to react to a newly initiated message. For example, a message published on the evening of June 23, 2020, is unlikely to receive a large number of Likes because we retrieved the data the next day. Our approach enables us to avoid the possibility of underestimating the attention that a new message would receive. From the eight hotel chains’ Facebook pages, we retrieved the field of post date, author (one of the eight hotel chains), the content of a message, the media type associated with a message (photo, video, hyperlink, or pure text), as well as the numbers of reactions (including like, love, care, ha-ha, wow, sad, and angry), comments, and shares of each message. From the eight hotel chains’ Twitter accounts, we retrieved the field of post date, author, message content, media type associated with a message, as well as the numbers of hearts/likes, replies (comments), and retweets of each message.
Because the fan base of our sample’s Facebook and Twitter pages vary from about 40,000 to over a million, simply using the numbers of reactions, comments, or shares/retweets that a message received may not reflect the true attention received by this message across different hotel chains’ fan pages. We then adopted the same approach taken in similar studies (e.g., Kwok & Yu, 2013, 2016), in which we covered the numbers of reactions, comments, and shares/retweets into a z score. Hence, the dependent variables become the z scores of reactions, comments, and shares/retweets, which measure the relative popularity level of a message within a fan page. A message with a positive z score means this message received above-average attention in terms of reactions, comments, and shares/retweets, from a fan page’s followers. Meanwhile, the message with a negative z score means this message received below-average attention from a fan page’s followers. In our sample, the z score of various measures ranges from −2.470 to 12.439.
Data Coding
First, the research team screened a random sample of 50 Facebook messages to explore the feasibility of classifying social media messages about COVID-19 according to the SCCT framework. Because hotels are also victims of the crisis, we found that hotel chains indeed used the Reminding, Ingratiation, and Victimage strategies under the Bolstering Posture from time to time, as suggested by SCCT (Table 1). Nevertheless, one message type did not fit into any communication strategies under the current SCCT framework. That specific message informed stakeholders about the proactive actions that the hotel has taken or will take to cope with the pandemic, such as enhancing the cleaning procedures across all properties. Thus, a new category under the Rebuilding Posture in SCCT was necessary. We then added the Prevention category to capture such a strategy adopted by the hotel chains and developed an expanded coding scheme for this investigation (Table 3), allowing us to systematically analyze hotels’ social media messages by the content conveyed in the messages during the COVID-19 crisis. As a result, we coded hotels’ social media messages during the COVID-19 pandemic as Prevention, Reminding, Ingratiation, Victimage, and (Other COVID-19) Updates, plus one type of Non-COVID-19 Messages.
| Message Type | Description | Example (Company, Month/Date in 2020) |
|---|---|---|
| Rebuilding posture | ||
| Preventiona | To inform stakeholders the proactive procedures taken or will be taken by the organization to cope with a crisis (i.e., stop the spread of the coronavirus) | Introducing Marriott’s Commitment to Clean our new initiative that aligns with expert health & safety protocols. Watch our video to learn more about how we’re working to provide a cleaner safer environment for our guests and associates. /URL/ (Marriott, 05/06) |
| Bolstering posture | ||
| Reminding | To remind stakeholders about the organization’s good work in the past | For nearly 75 years, Best Western Hotels & Resorts has been an industry leader in providing clean well-maintained hotel accommodations for our valued guests . . . (Best Western, 06/16) |
| Ingratiation | To praise the stakeholders, such as the staff, customers, and the management team. | Caring For Our Healthcare Workers: Párisi Udvar Hotel Budapest is supporting nearby hospitals by providing 50 lunches per day. (Hyatt, 05/29) |
| Victimage | To explain how the organization is a victim of the crisis | COVID-19 has impacted the travel industry in a multitude of ways—not least the employment of our talented hotel colleagues as guests press pause on their travel plans . . . (IHG, 03/30) |
| Updatesa (about COVID-19) | Updates about adjustments in operations due to COVID-19 (e.g., cancelation policy and use of reward points/status) or expressions of hopes about COVID-19 | Help those impacted by COVID-19 by donating your Radisson Rewards points to one of our charity partners: American Red Cross or SOS Children’s Villages International. /URL/ (Radisson, 05/14) |
| Non-COVID-19 Messagesa | Other updates irrelevant to COVID-19 | Save up to 25% on your next stay with #dreamdeals! Whether you want to stay close to home or explore somewhere new don’t miss this dream of an offer. Book by July 14 and stay before August 31. /URL/ (Radisson, 06/03) |
Source. Adopted from Coombs (2014) and Roshan et al. (2016).
a
Emerged from the preliminary coding process by the authors.
To further validate whether the newly developed coding scheme applied to this research setting, we built a training data set with Facebook messages initiated by the eight hotel chains in June for a test run. The training data set contains 41 messages. The first author, along with two graduate students, coded these 41 messages independently according to the expanded coding scheme in Table 3. In this training exercise, all three coders first coded the messages into two types: COVID-19 messages versus non-COVID-19 messages. The coders then coded the COVID-19 messages in one of the five categories of Prevent, Reminding, Ingratiation, Victimage, and Updates. We found that some COVID-19 messages on Facebook were lengthy, which the first few sentences can be coded with one of the five categories of COVID-19 messages, and the rest can be coded with another category (e.g., Reminding + Prevention or Victimage + Ingratiation). In the real-world setting, most social media platforms display only the first two to three lines (first few sentences) of a long paragraph. Hence, most internet users may not spend time reading lengthy messages unless they choose to expand them to read more. We then coded those COVID-19 messages that could be categorized into two message types with one priority type based on the information conveyed in the first few sentences. At the end of the coding exercise, the two graduate students reached 92.68% intercoder agreement when coding the messages as COVID-19 versus non-COVID-19 messages, with three messages that were coded differently. Among the COVID-19 messages, only four messages were coded differently under the five (priority) types of COVID-19 messages. Finally, the research team discussed the discrepancies and reached a consensus on the type(s) of all messages in the training data set.
Finally, the two graduate-student coders coded all 657 Facebook messages and 754 Twitter messages on their own. Through a comparison of their coding results, the intercoder reliability for COVID-19 versus non-COVID-19 messages reached 99.70% for the Facebook data set and 99.07% for the Twitter data set (two Facebook messages and seven Twitter messages were coded differently by the two coders), indicating an almost perfect agreement. Furthermore, the intercoder reliability for the five priority types of COVID-19 messages and the five secondary types of COVID-19 messages reached 76.09% for the Facebook data set and 67.65% for the Twitter data set (33 of 138 COVID-19 messages on Facebook and 55 of 170 COVID-19 messages on Twitter were coded differently by the two coders), suggesting an acceptable agreement between the two coders for exploratory analysis (Lombard et al., 2002).
During the coding process, we also found that these hotel chains sometimes share the same messages (content) on both Facebook and Twitter, confirming the suggestion of Kwok and Yu (2013) that businesses usually share similar content across different social media platforms. In the cases when the coders coded a message under different categories, a consensus was reached after the coders discussed the discrepancies with the research team.
Data Analysis to Answer the Three Research Questions
We performed a series of descriptive analyses to answer RQ1 and RQ2 and analysis of variance (ANOVA) and post hoc analysis to answer RQ3 and RQ4. Our results yielded a few intriguing findings that can provide valuable implications regarding crisis communication during the COVID-19 crisis.
Results and Discussion
The COVID-19 Messages Shared by the Hotel Chains
RQ1 aims to identify the social media communication strategies adopted by hospitality companies through an analysis of the types of messages shared by the companies. We found that hotel chains significantly reduced social networking activities in April. As shown in Table 4, the number of these hotel chain’s Facebook posts and tweets “suddenly” decreased from about 200 in January to less than 100 after April. The pandemic might have significantly slowed down the social media activities for these hotels as they were trying to figure out effective communication strategies to stay engaged with their stakeholders or understand/estimate the significant impacts of this pandemic on the business. Alternatively, hotel chains were more quiet than usual on Facebook and Twitter because they were making serious decisions concerning lockdowns, closing down, temporary furloughs, and other challenges.
| Month | ||||||
|---|---|---|---|---|---|---|
| COVID-19 Messages | All Messages | % | COVID-19 Messages | All Messages | % | |
| January | 0 | 182 | 0 | 1 | 224 | 0.45 |
| February | 0 | 170 | 0 | 0 | 200 | 0 |
| March | 26 | 115 | 22.61 | 36 | 133 | 27.07 |
| April | 44 | 67 | 65.67 | 51 | 66 | 77.27 |
| May | 37 | 82 | 45.12 | 58 | 84 | 69.05 |
| Junea | 16 | 41 | 39.02 | 27 | 47 | 57.45 |
| Total | 123 | 657 | 18.72 | 173 | 754 | 22.94 |
a
Until June 19, 2020.
Our findings also suggested that the world’s largest hotel chains failed to take a proactive approach in responding to such a devastating crisis before the WHO officially announced the already widespread coronavirus as a global pandemic. Despite the news media frenzy, communication between the hotel chains and their stakeholders about the pandemic before March 2020 was seemingly lacking. Nevertheless, although the total number of messages shared on Facebook and Twitter dropped substantially in April, the percentage of COVID-19 messages reached the highest level of 70% of all messages in April 2020. Such a trend continued in May and June 2020, where a large portion of Facebook and Twitter messages were about COVID-19.
Table 5 shows the breakdown of the COVID-19 messages shared on Facebook and Twitter by the eight hotel chains. IHG posted the most COVID-19 messages, with 40 on Facebook and 60 on Twitter, followed by Hyatt. Marriott is the largest hotel chain globally but only shared six COVID-19 messages on Facebook and Twitter, the lowest number among all hotel chains. Most hotel chains posted a similar number of messages on Facebook and Twitter, but Hilton and Radisson seemed to post more on Twitter, and Wyndham posted much more on Facebook. It was also probably because Wyndham does not maintain an active Twitter page, whose account has the lowest number of followers among all hotel chains (Table 1).
| Hotel Chains | Total | Prevention | Reminding | Ingratiation | Victimage | Updates | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ST | FB | TW | ST | FB | TW | ST | FB | TW | ST | FB | TW | ST | FB | TW | ST | FB | TW | |
| Marriott | 12 | 6 | 6 | 2 | 1 | 1 | 1 | 0 | 1 | 1 | 1 | 0 | 0 | 0 | 0 | 8 | 4 | 4 |
| Hilton | 21 | 7 | 14 | 5 | 3 | 2 | 3 | 0 | 3 | 10 | 4 | 6 | 0 | 0 | 0 | 3 | 0 | 3 |
| IHG | 100 | 40 | 60 | 9 | 4 | 5 | 0 | 0 | 0 | 51 | 18 | 33 | 11 | 5 | 6 | 29 | 13 | 16 |
| Wyndham | 15 | 14 | 1 | 5 | 5 | 0 | 0 | 0 | 0 | 3 | 3 | 0 | 0 | 0 | 0 | 7 | 6 | 1 |
| Choice Hotels | 28 | 13 | 15 | 7 | 3 | 4 | 0 | 0 | 0 | 12 | 6 | 6 | 0 | 0 | 0 | 9 | 4 | 5 |
| Best Western | 38 | 20 | 18 | 2 | 0 | 2 | 5 | 3 | 2 | 25 | 14 | 11 | 1 | 1 | 0 | 5 | 2 | 3 |
| Hyatt | 70 | 33 | 37 | 5 | 3 | 2 | 2 | 1 | 1 | 41 | 18 | 23 | 0 | 0 | 0 | 22 | 11 | 11 |
| Radisson | 24 | 5 | 19 | 7 | 2 | 5 | 0 | 0 | 0 | 5 | 0 | 5 | 0 | 0 | 0 | 12 | 3 | 9 |
| Total | 308 | 138 | 170 | 42 | 21 | 21 | 11 | 4 | 7 | 148 | 64 | 84 | 12 | 6 | 6 | 95 | 43 | 52 |
Note. ST is the sum of FB and TW. ST = subtotal; FB = Facebook; TW = Twitter.
Looking at the types of COVID-19 messages shared by different hotel chains, we found that Ingratiation, with which organizations praised their stakeholders during the COVID-19 crisis, was the most frequently used strategy by the sample. It accounted for 48.06% (148 of 308) of the total COVID-19 messages on both platforms. In the cases of IHG, Best Western, and Hyatt, Ingratiation even accounted for more than 50% of all COVID-19 messages they shared. Updates, where hotel chains informed their stakeholders about the changes in operations or shared relevant information about COVID-19, was the second most-used message type, with 95 posts or 30.84%. Meanwhile, Prevention was the message type that showed hotels had taken or would take proactive procedures to prevent the spread of the coronavirus, which accounts for 13.63% (42 of 308) of all COVID-19 messages. The crisis communication strategies of Reminding and Victimage appeared to be underused among the hotel chains. Only Best Western, for example, reminded its stakeholders of its past good work about five times. Some other hotel chains never used the Reminding strategy. Victimage was the message type used least often by all hotel chains, of which IHG used it 11 (of 100) times, and Best Western used it once only. Possibly, not every hotel chain’s content managers were fully aware of the crisis communication strategies available to organizations. Alternatively, the hotel chains in the sample might purposefully choose to avoid relating their past good work to the COVID-19 or admitting that they were also the victim of the crisis.
The Media Types Used by the Hotel Chains
RQ2 asked whether hospitality companies share photos more frequently than other media types, as suggested in previous studies (e.g., Kwok et al., 2015; Kwok & Yu, 2013; Sabate et al., 2014). We also discovered in this investigation that visual content had become the dominant form of online communications. From the web operators’ perspective, we observed that Facebook and Twitter displayed a screenshot of the video for updates with a video or a preview photo from a web page if an update included a hyperlink directing to the web page. Such features made these types of updates look the same as a post with a photo. Even for a message with pure text, if it had fewer than 130 characters, Facebook (not on Twitter) would display the message with a colorful background by default, also making such a post with pure text look like a photo instead. As a result, almost all Facebook posts, except for those in pure text of 130+ characters, looked visual, as if they were messages with at least one photo. From the marketing communication perspective, we found that hotel chains shared messages in a combination of two or more media types (e.g., photos + hyperlinks, hyperlinks + videos, photos + videos, or photos + hyperlinks + videos). Of the 657 Facebook messages, the raw data yielded no messages with pure text on Facebook, but 506 messages with photo(s), 479 messages with hyperlink(s), and 124 messages with video(s). Of the 754 tweets, the data reported 426 messages with photo(s), 544 messages with hyperlink(s), 142 messages with video(s), and only 20 messages with pure text.
Videos can be perceived as a collection of (moving) images; therefore, we expected the following hierarchical order from the richest to the least amount of information carried by the four media types: video, photo, hyperlink, and pure text. For additional comparisons, we then went one step further to count a message’s media type exclusively with one category only in the order of video, photo, link, and pure text, allowing us to categorize the messages based on the level of richness of information that a media type (message) carries. In the end, we identified the following numbers of Facebook messages and Twitter messages, respectively: 124 and 142 with videos, 505 and 426 with photos, 28 and 166 with hyperlinks, and 0 and 20 with pure texts (Table 6).
| Hotel Chains | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Link | Photo | Video | Sub-Total | Photo (%) | Video (%) | Text | Link | Photo | Video | Sub-Total | Photo (%) | Video (%) | |
| Marriott | 2 | 152 | 5 | 159 | 95.60 | 3.14 | 1 | 91 | 60 | 5 | 157 | 38.22 | 3.18 |
| Hilton | 10 | 28 | 11 | 49 | 57.14 | 22.45 | 0 | 13 | 78 | 11 | 102 | 76.47 | 10.78 |
| IHG | 11 | 57 | 19 | 87 | 65.52 | 21.84 | 14 | 18 | 67 | 18 | 117 | 57.26 | 15.38 |
| Wyndham | 1 | 34 | 2 | 37 | 91.89 | 5.41 | 0 | 0 | 1 | 0 | 1 | 100.00 | 0.00 |
| Choice Hotels | 0 | 26 | 5 | 31 | 83.87 | 16.13 | 0 | 1 | 26 | 4 | 31 | 83.87 | 12.90 |
| Best Western | 0 | 115 | 17 | 132 | 87.12 | 12.88 | 4 | 8 | 98 | 11 | 121 | 80.99 | 9.09 |
| Hyatt | 3 | 55 | 63 | 121 | 45.45 | 52.07 | 0 | 5 | 67 | 80 | 152 | 44.08 | 52.63 |
| Radisson | 1 | 38 | 2 | 41 | 92.68 | 4.88 | 1 | 30 | 29 | 13 | 73 | 39.73 | 17.81 |
| Total | 28 | 505 | 124 | 657 | 76.86 | 18.87 | 20 | 166 | 426 | 142 | 754 | 56.50 | 18.83 |
Looking at how each hotel chain used visual content on Facebook, we found that Marriott (152 of 159) and Best Western (115 of 132) seemed to be well versed in using images in communications. Meanwhile, although Wyndham, Choice Hotels, and Radisson did not share many Facebook messages, the majority of their posts were attached with photo(s). For sharing messages with videos, Hyatt topped the list, recording 52.07% or 63 messages. Likewise, Best Western tweeted 98 messages with photos (80.90%), and Hyatt tweeted 80 messages with videos (52.63%). Although most hotel chains relied on sharing visual content on Twitter, Marriott tweeted 91 of 157 messages with hyperlinks, which was different from Marriott’s photo-focused approach on Facebook. Such a difference might come from the distinct features on Facebook and Twitter. Both Facebook and Twitter have the preview function to display some visual content of the messages with hyperlinks. However, Twitter does not have the “album” function that allows companies to organize the photos they shared, as they would do on Facebook. Such a difference between Facebook and Twitter might discourage Marriott from attaching photos to the tweets with hyperlinks. Meanwhile, the fact that Hilton’s Twitter account was specified as “Hilton Newsroom” may explain why Hilton shared 102 tweets, as compared with the 49 messages shared on Hilton Hotels’ Facebook page. In addition, our results show that Wyndham did not tweet much and that Choice Hotels and Radisson were not as active as other hotel chains in terms of sharing information on Twitter or Facebook.
Internet Users’ Reactions to the COVID-19 Messages Shared by the Hotel Chains
RQ3 posits that internet users will react differently toward COVID-19 versus non-COVID-19 messages. First, we conducted a series of independent t tests to answer RQ3. As shown in Table 7, COVID-19 messages, as compared with non-COVID-19 messages, received higher numbers of reactions/Likes, comments/replies, and shares/retweets across both platforms. Moreover, the large gap in mean differences indicates that hotel stakeholders paid great attention to the COVID-19 messages. Our results confirmed that social media, although being underused, can be an effective means for businesses to distribute information about a crisis, echoing other scholars’ recommendations about crisis communication (Möller et al., 2018; Utz et al., 2013; Veil et al., 2011; Zhai et al., 2019).
| Dependent Variables | Independent Variables (Facebook) | Independent Variables (Twitter) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| COVID-19 Messages | Non-COVID-19 Messages | Mean Difference | t-Score | p Value | COVID-19 Messages | Non-COVID-19 Messages | Mean Difference | t-Score | p Value | |||||
| M | SD | M | SD | M | SD | M | SD | |||||||
| Reactions/Likes | 0.462 | 1.444 | −0.123 | 0.793 | 0.585 | 6.324 | <.000 | 0.604 | 1.641 | −0.176 | 0.602 | 0.780 | 9.516 | <.000 |
| Comments | 0.553 | 1.892 | −0.298 | 0.894 | 0.851 | 7.561 | <.000 | 0.361 | 1.832 | −0.105 | 0.508 | 0.466 | 5.474 | <.000 |
| Shares/Retweets | 0.616 | 1.781 | −0.164 | 0.535 | 0.779 | 8.626 | <.000 | 0.562 | 1.682 | −0.164 | 0.583 | 0.726 | 8.781 | <.000 |
Note. M = mean; SD = standard deviation.
It is noteworthy to mention that through our observations, #BLM or #BlackLivesMatter messages seemed to receive significant attention on Facebook and Twitter as well, even though such a topic was not the focus of this study and was coded as non-COVID-19 messages. Such a finding suggested that hotels should go beyond the mindset of using social media as a sales marketing tool and learn how to carry a conversation about the concurrent issues with their stakeholders online, as suggested in the relevant literature (Kwok & Yu, 2013, 2016).
Furthermore, we conducted a series of ANOVA analyses to test whether a specific type of COVID-19 message would receive significantly more attention from internet users. Our results showed significant differences among the five specific COVID-19 message types (Prevention, Reminding, Ingratiation, Victimage, and Updates) and the non-COVID-19 messages as measured by the number of reactions received on Facebook, F(5, 655) = 44.581, p < .001, and Twitter, F(5, 752) = 92.272, p < .001, the number of comments received on both Facebook, F(5, 655) = 64.076, p < .001, and Twitter, F(5, 752) = 117.223, p < .001, and the number of shares/retweets received on both Facebook, F(5, 655) = 185.603, p < .001, and Twitter, F(5, 752) = 143.321, p < .001. The descriptive statistics of the five COVID-19 message types versus the non-COVID-19 messages are presented in Table 8.
| Message Type | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Reactions | Comments | Shares | Likes | Comments | Retweets | |||||||
| M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | |
| Prevention | 0.809 | 1.812 | 1.223 | 2.050 | 1.660 | 2.383 | 1.121 | 1.209 | 0.696 | 2.777 | 1.434 | 1.709 |
| Reminding | 0.470 | 0.620 | 1.903 | 2.231 | 4.609 | 2.855 | 1.911 | 2.963 | 1.105 | 1.805 | 1.694 | 2.955 |
| Ingratiation | 0.264 | 1.055 | 0.011 | 1.621 | 0.106 | 1.052 | 0.419 | 1.410 | 0.028 | 1.060 | 0.250 | 1.211 |
| Victimage | 0.224 | 0.316 | –0.066 | 0.259 | –0.131 | 0.115 | –0.108 | 0.200 | –0.373 | 0.469 | –0.201 | 0.210 |
| Updates | 0.621 | 1.836 | 0.993 | 2.066 | 0.597 | 1.691 | 0.601 | 1.901 | 0.747 | 2.310 | 0.650 | 2.022 |
| Non-COVID-19 | –0.123 | 0.793 | –0.298 | 0.894 | –0.164 | 0.535 | –0.176 | 0.602 | –0.105 | 0.508 | –0.164 | 0.583 |
Note. M = mean; SD = standard deviation.
In the categories where the F score reported a significant value, we performed additional Bonferroni post hoc analyses. As reported in Table 9, Prevention, Reminding, Ingratiation, and Updates seemed to be the types of COVID-19 messages that received more attention from Facebook and/or Twitter users, contingent on the outcome variables. Referring to Table 5, which shows the breakdown of how often the different types of COVID-19 messages were used, Reminding and Victimage were the two types of communication strategies barely used. These two types of messages could be as effective if they were used more often and by more hotel chains. Future research with additional data can help confirm the effectiveness of Reminding and Victimage strategies in crisis communication.
| Message Type (I) | Message Type (J) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Reactions | Comments | Shares | Likes | Comments | Retweets | ||||||||
| MD (I – J) |
p Value | MD (I – J) |
p Value | MD (I – J) |
p Value | MD (I – J) |
p Value | MD (I – J) |
p Value | MD (I – J) |
p Value | ||
| Prevention | Reminding | 0.339 | 1.000 | −0.680 | 1.000 | −2.949 | <.000 | −0.791 | .765 | −0.409 | 1.000 | −0.260 | 1.000 |
| Prevention | Ingratiation | 0.545 | .374 | 1.212 | <.000 | 1.554 | <.000 | 0.701 | .030 | 0.668 | .068 | 1.184 | <.000 |
| Prevention | Victimage | 0.585 | 1.000 | 1.289 | .234 | 1.791 | <.000 | 1.229 | .064 | 1.069 | .248 | 1.635 | .002 |
| Prevention | Updates | 0.189 | 1.000 | 0.231 | 1.000 | 1.063 | <.000 | 0.520 | .457 | −0.051 | 1.000 | 0.784 | .016 |
| Prevention | Non-COVID | 0.932 | <.000 | 1.521 | <.000 | 1.824 | <.000 | 1.296 | <.000 | 0.801 | .003 | 1.598 | <.000 |
| Reminding | Ingratiation | 0.206 | 1.000 | 1.893 | .022 | 4.503 | <.000 | 1.492 | .001 | 1.077 | .068 | 1.444 | .001 |
| Reminding | Victimage | 0.246 | 1.000 | 1.969 | .121 | 4.739 | <.000 | 2.019 | .001 | 1.478 | .088 | 1.895 | .004 |
| Reminding | Updates | −0.151 | 1.000 | 0.911 | 1.000 | 4.012 | <.000 | 1.310 | .007 | 0.358 | 1.000 | 1.043 | .078 |
| Reminding | Non-COVID | 0.593 | 1.000 | 2.202 | .002 | 4.772 | <.000 | 2.087 | <.000 | 1.210 | .015 | 1.858 | <.000 |
| Ingratiation | Victimage | 0.040 | 1.000 | 0.076 | 1.000 | 0.237 | 1.000 | 0.528 | 1.000 | 0.401 | 1.000 | 0.451 | 1.000 |
| Ingratiation | Updates | −0.356 | .920 | −0.982 | <.000 | −0.491 | .056 | −0.181 | 1.000 | −0.719 | <.000 | −0.401 | .215 |
| Ingratiation | Non-COVID | 0.387 | .038 | 0.309 | .639 | 0.270 | .264 | 0.595 | <.000 | 0.133 | 1.000 | 0.414 | .002 |
| Victimage | Updates | −0.396 | 1.000 | −1.058 | .522 | −0.728 | .770 | −0.709 | 1.000 | −1.120 | .105 | −0.852 | .495 |
| Victimage | Non-COVID | 0.347 | 1.000 | 0.233 | 1.000 | 0.033 | 1.000 | 0.068 | 1.000 | −0.268 | 1.000 | −0.037 | 1.000 |
| Updates | Non-COVID | 0.743 | <.000 | 1.291 | <.000 | 0.761 | <.000 | 0.777 | <.000 | 0.852 | <.000 | 0.814 | <.000 |
Note. MD = mean difference.
Internet Users’ Reactions to the Messages in Four Media Types
RQ4 was set to determine whether internet users would react to hospitality companies’ social media messages differently based on the four media types associated with the messages. Table 10 shows the descriptive statistics of the four media type usages in the sample. None of the hotel chains shared any Facebook messages in pure text; hence, we reported no value under that category. The results of ANOVA analyses reported the following F scores for the numbers of reactions, F(2, 654) = 1.475, p = .230, comments, F(2, 654) = 4.488, p = .012, and shares, F(2, 654) = 5.946, p = .003, received on Facebook, and for the numbers of Likes, F(3, 750) = 4.646, p = .003, comments/replies, F(3, 750) = 1.000, p = .392, and retweets, F(3, 750) = 4.340, p = .005, received on Twitter.
| Media Type | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Reactions | Comments | Shares | Likes | Comments | Retweets | |||||||
| M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | |
| Text | N/A | N/A | N/A | N/A | N/A | N/A | −0.473 | 0.338 | −0.288 | 0.482 | −0.330 | 0.369 |
| Link | −0.263 | 0.314 | −0.765 | 1.623 | −0.210 | 0.145 | −0.117 | 0.794 | 0.006 | 0.884 | −0.121 | 0.615 |
| Image | −0.007 | 0.955 | −0.070 | 1.184 | −0.054 | 0.888 | −0.005 | 0.892 | −0.019 | 0.917 | −0.017 | 0.924 |
| Video | 0.088 | 1.223 | −0.175 | 1.252 | 0.268 | 1.394 | 0.219 | 1.429 | 0.090 | 1.338 | 0.238 | 1.471 |
Note. M = mean; SD = standard deviation.
We then performed additional Bonferroni post hoc analyses for the categories whose F score has a p value lower than .05. We reported the results for the post hoc analysis in Table 11. The results support the proposition suggested in the media richness theory (Daft & Lengel, 1986) that internet users process the information presented in various types of media differently because of the unique information cues associated with the medium that carries the information. Rich media with visual content, such as a video and a picture, may carry more information than pure text or hyperlink and hence receive more attention from internet users. Such a finding also agreed with some empirical studies (Kwok et al., 2015).
| Message Type | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Comments | Shares | Likes | Retweets | ||||||
| (I) | (J) | MD (I – J) |
p Value | MD (I – J) |
p Value | MD (I – J) |
p Value | MD (I – J) |
p Value |
| Text | Link | N/A | N/A | N/A | N/A | −0.356 | .769 | −0.209 | 1.000 |
| Text | Photo | N/A | N/A | N/A | N/A | −0.468 | .233 | −0.313 | 1.000 |
| Text | Video | N/A | N/A | N/A | N/A | −0.693 | .021 | −0.567 | .099 |
| Link | Photo | −0.696 | .010 | −0.156 | 1.000 | −0.112 | 1.000 | −0.104 | 1.000 |
| Link | Video | −0.590 | .063 | −0.477 | .063 | −0.337 | .018 | −0.358 | .009 |
| Photo | Video | 0.105 | 1.000 | −0.322 | .004 | −0.225 | .115 | −0.254 | .049 |
Note. MD = mean difference.
Implications and Conclusion
Businesses must maintain continuous communications with their stakeholders during the pandemic. They may share various types of information/messages with their stakeholders on social media. The messages receiving more reactions and comments and getting more shares from the internet users can create a more significant electronic word-of-mouth (eWOM) effect (Kwok et al., 2019; Soboleva et al., 2017).
This study identifies effective social media communication strategies in an unexpected crisis and which strategies are actually working. Guided by SCCT, we first analyzed the types of Facebook and Twitter messages shared by the top eight hotel chains. Then, we compared what type of Facebook and Twitter messages received more attention from Facebook and Twitter users as measured by the number of reactions/likes, comments, and shares/retweets. Our analysis reveals some valuable insights that make significant contributions to theory and practices.
Theoretical Implications
First, we extended the SCCT framework (as in Table 2; Coombs, 2014; Roshan et al., 2016) with the creation of the Prevention strategy under the Rebuilding Posture. The COVID-19 pandemic is still an ongoing global crisis, and nobody can precisely predict how and when the spread of coronavirus would end. Thus, businesses of all sectors are still finding ways to identify solutions to cope with the event’s devastating effects. The current SCCT framework could not capture the proactive actions taken or will be taken by a business in responding to events like the COVID-19 pandemic. The introduction of the Prevention strategy is helpful in future studies in a similar setting.
Second, we contribute to the crisis management literature by responding to the call for more analyses of real-time information exchange among an organization and its stakeholders (Bundy et al., 2017). This study joins relevant studies in crisis management (e.g., Möller et al., 2018; Zhai et al., 2019; on crisis communication on Facebook and in a Chinese social media context, respectively) to narrow such a research gap. Our analysis of internet users’ real reactions toward different types of COVID-19 messages on Facebook and Twitter adds valuable insights to the literature.
Third, the coding scheme reported in Table 3 can be a reference in developing an algorithm for machine learning. Currently, we tested the intercoder reliability of the coding scheme with a small data set of the Facebook and Twitter messages initiated by eight hotel chains. Such a coding scheme can be further used to test an algorithm that can eventually be applied to big data analytics, such as machine learning (e.g., Antonio et al., 2019; Kwok et al., 2020). The machine learning approach can be used in an automatic analysis with a much larger data set to assess more companies’ crisis communication strategies during the COVID-19 pandemic.
Finally, to our best knowledge, this research is the first study in hospitality and tourism management that addresses the urgent need to identify effective communication strategies about COVID-19. Responding to Tracey’s (2020) suggestion, this investigation helps us advance several actionable recommendations in crisis communication, providing alternative solutions to the hospitality industry as the COVID-19 crisis evolves.
Practical Implications
Our findings revealed that hotel stakeholders responded well to companies’ COVID-19 messages on Facebook and Twitter. Nevertheless, hospitality companies do not seem to have fully used social media’s power in sharing COVID-19-related information with their stakeholders. When a B2C message receives more likes, comments, and shares/retweets, it creates a higher eWOM effect for the business (Kwok et al., 2019). We recommend hospitality companies use social media as an essential channel to distribute relevant information about COVID-19. Table 12 summarizes this study’s significant findings and the specific recommendations to hospitality companies regarding the available communication strategies during the COVID-19 crisis.
| Message Type | Reactionsa | Commentsa | Shares/Retweetsa | Recommendations |
|---|---|---|---|---|
| COVID-19 messages | F + T | F + T | F + T | Use social media to communicate with the stakeholders about the crisis. |
| Preventionb | F + T | F + T | F + T | Let people know what the organizations are doing proactively to cope with the crisis. |
| Remindingb | T | F + T | F + T | Although underused, try this message type out to test its effect in attracting stakeholders’ attention. |
| Ingratiationb | F + T | — | T | Highlight the good work that the stakeholders (e.g., consumers, employees, and managers) are doing to cope with the crisis. |
| Victimageb | — | — | — | Although underused, try this message type out to test its effect in attracting stakeholders’ attention. |
| Updatesb | F + T | F + T | F + T | Share changes or updates of operation issues on social media. |
| Non-COVID-19 messages | — | — | — | Do not use social media only for sales and marketing purposes. Stakeholders want to hear about COVID-19 and concurrent issues (e.g., #BLM). |
| Media type | ||||
| Video | T | — | F + T | Create short videos to communicate with the stakeholders. |
| Photo | — | F | — | Continue using photos as the visual content. |
| Hyperlink | — | — | — | Select an appealing image as a preview of the hyperlinks. Attach a video or photo to the message. |
| Pure text | — | — | — | Use 130 or fewer characters to make the update look like a “picture” on Facebook. Attach a video or photo to the message. |
a
F and T indicate that this message type is likely to get more reactions, comments/replies, or shares/retweets on Facebook and Twitter, respectively. bAs the message type compared with non-COVID-19 messages.
We first recommend that hospitality companies actively use social media in communicating with their stakeholders about the COVID-19 crisis. Our results suggest that people want to hear the information about COVID-19 on social media. They can further help businesses spread the COVID-19 information in their networks through Likes, writing comments/replies, and/or shares/retweets. Although Prevention, Ingratiation, and Updates were used more often by hotels and seemed to be effective in attracting internet users’ attention, we highly encourage more businesses to also incorporate Reminding and Victimage in their crisis communication strategies. Reminding and Victimage may still be influential among a specific group or segment of stakeholders.
Then, we encourage hospitality companies to refer to the crisis communication strategies presented in Table 3 before sharing a COVID-19 message with their stakeholders. Taking the Ingratiation strategy as an example, hospitality companies can find creative ways to praise the good deeds of their stakeholders, such as the management team, the frontline employees, the customers, and other business partners, in response to the COVID-19 crisis. Tagging or mentioning those stakeholders in a message can also be helpful.
Finally, we praise the fact that the hotel chains in our sample have widely used visual content in social media communication. We believe hospitality companies should continue using videos and photos as the essential media to communicate with their stakeholders online. In our analysis, however, we still found a few hotel chains often sharing lengthy messages in paragraphs (about 200 words). Most social media platforms only display the first two to three lines of a long message to their users. Therefore, we strongly recommend that companies shorten their social media messages, ideally within 130 characters (i.e., in one or two simple sentences) because not every internet user has the patience to read long paragraphs of information.
Limitations and Future Research
This study is not without limitations. First, we focused our analysis on the top eight hotel chains. Running a hotel is very different from operating restaurants, casinos, and other hospitality businesses. Although we believe our implications can benefit more than just hoteliers, we suggest that other businesses should test out the effectiveness of different crisis communication messages as outlined in Table 3 for developing the right social media strategies to engage their stakeholders during COVID-19. Second, we only assessed hotels’ crisis communication strategies on Facebook and Twitter. The active users on other social media sites, such as LinkedIn, Instagram, and YouTube, might have very distinct tastes or interests than those on Facebook and Twitter. We encourage practitioners to also test different crisis communication strategies in other social media sites. Third, to a large extent, our analysis relied on coding accuracy. Although we tried our best to minimize human errors through intercoder reliability testing and finalized the coding results based on the whole research team’s collective decision, this study may not be completely free of errors. Fourth, our analysis excluded other possible influential factors in our estimations, such as the length of messages or usage of emojis and punctuation marks. We encourage further studies to also assess the effects of those factors in crisis communication. Last, we highly recommend future studies to go beyond SCCT and embrace the big data analytic techniques in their investigations when analyzing businesses’ crisis communication strategies during the COVID-19 pandemic. On one hand, studies using different theories, frameworks, or methods can usually enrich our understanding of a very complex phenomenon (Xie et al., 2020). On the other hand, studies using big data analytics techniques can be more exploratory and might not necessarily follow the deductive tradition of inquiry, such as in Antonio et al. (2019). Through the exploration of an immense amount of data on the internet, best practices of how businesses engage their stakeholders during the COVID-19 pandemic can also emerge.
Acknowledgments
The authors highly appreciate Hao Wang and Yuan Zhuang’s valuable contributions in coding the data, both of whom are students in the Kyoto-Cornell Double Degree Opportunity Program in the Graduate School of Management at Kyoto University in Japan.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, or publication of this article: This research was supported by an internal fund at California State Polytechnic University Pomona.
ORCID iDs
Linchi Kwok https://orcid.org/0000-0001-8838-2485
Jungwoo Lee https://orcid.org/0000-0002-4130-4970
Spring H. Han https://orcid.org/0000-0002-0465-4437
Footnotes
1. The State Public Health Officers refer essential workers as the “essential critical infrastructure workers” who work to protect communities and ensure the continuity of functions that are critical to public health and safety, as well as economic and national security, such as those in emergency services, energy, and critical manufacturing. A full list of the essential workers is available at https://covid19.ca.gov/essential-workforce/.
2. More detailed descriptions about research using Facebook (public page) data can be accessed on https://research.fb.com/blog/2021/03/new-analytics-api-for-researchers-studying-facebook-page-data/.
3. More detailed descriptions about using Twitter API can be accessed on https://developer.twitter.com/en/developer-terms/commercial-terms and https://developer.twitter.com/en/products/twitter-api/academic-research.
References
Antonio N., de Almeida A., Nunes L. (2019). Big data in hotel revenue management: Exploring cancellation drivers to gain insights into booking cancellation behavior. Cornell Hospitality Quarterly, 60(4), 298–319.
Araujo T., Neijens P., Vliegenthart R. (2015). What motivates consumers to re-tweet brand content? Journal of Advertising Research, 55(3), 284–295.
Bapuji H., de Bakker F. G., Brown J. A., Higgins C., Rehbein K., Spicer A. (2020). Business and society research in times of the corona crisis. Business and Society, 59(6), 1067–1078.
Bapuji H., Patel C., Ertug G., Allen D. G. (2020). Corona crisis and inequality: Why management research needs a societal turn. Journal of Management, 46, 1205–1222. https://doi.org/10.1177/0149206320925881
Baum T., Hai N. T. T. (2020). Hospitality, tourism, human rights and the impact of COVID-19. International Journal of Contemporary Hospitality Management, 32(7), 2397–2407.
Brouder P. (2020). Reset redux: Possible evolutionary pathways towards the transformation of tourism in a COVID-19 world. Tourism Geographies, 22, 484–490.
Bundy J., Pfarrer M. D. (2015). A burden of responsibility: The role of social approval at the onset of a crisis. Academy of Management Review, 40(3), 345–369.
Bundy J., Pfarrer M. D., Short C. E., Coombs W. T. (2017). Crises and crisis management: Integration, interpretation, and research development. Journal of Management, 43(6), 1661–1692.
Chao R. M., Chang C. C., Chang W. Y. (2014). Exploring the antecedents of trust from the perspectives of uncertainty and media richness in virtual community. International Journal of Web Based Communities, 10(2), 176–187.
Chinazzi M., Davis J. T., Ajelli M., Gioannini C., Litvinova M., Merler S., Piontti A. P., Mu K., Rossi L., Sun K., Viboud C., Xiong X., Yu H., Halloran E., Longini I. A., Vespignani A. (2020). The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak. Science, 368(6489), 395–400.
Coombs W. T. (1995). Choosing the right words: The development of guidelines for the selection of the “appropriate” crisis-response strategies. Management Communication Quarterly, 8(4), 447–476.
Coombs W. T. (2007). Protecting organization reputations during a crisis: The development and application of situational crisis communication theory. Corporate Reputation Review, 10(3), 163–176.
Coombs W. T. (2014). Ongoing crisis communication: Planning, managing, and responding. SAGE.
Coombs W. T. (2016). Reflections on a meta-analysis: Crystallizing thinking about SCCT. Journal of Public Relations Research, 28(2), 120–122.
Coombs W. T., Holladay S. J. (2002). Helping crisis managers protect reputational assets: Initial tests of the situational crisis communication theory. Management Communication Quarterly, 16(2), 165–186.
Daft R. L., Lengel R. H. (1986). Organizational information requirements, media richness and structural design. Management Science, 32(5), 554–571.
Fernandes N. (2020). Economic effects of coronavirus outbreak (COVID-19) on the world economy. https://doi.org/10.2139/ssrn.3557504
Gössling S., Scott D., Hall C. M. (2021). Pandemics, tourism and global change: A rapid assessment of COVID-19. Journal of Sustainable Tourism, 29, 1–20.
He H., Harris L. (2020). The impact of covid-19 pandemic on corporate social responsibility and marketing philosophy. Journal of Business Research, 116, 176–182.
He W., Zha S., Li L. (2013). Social media competitive analysis and text mining: A case study in the pizza industry. International Journal of Information Management, 33, 464–472.
Higgins-Desbiolles F. (2020). Socialising tourism for social and ecological justice after COVID-19. Tourism Geographies, 22, 610–623.
Ki E. J., Nekmat E. (2014). Situational crisis communication and interactivity: Usage and effectiveness of Facebook for crisis management by Fortune 500 companies. Computers in Human Behavior, 35, 140–147.
Kraus S., Clauss T., Breier M., Gast J., Zardini A., Tiberius V. (2020). The economics of COVID-19: Initial empirical evidence on how family firms in five European countries cope with the corona crisis. International Journal of Entrepreneurial Behavior & Research, 26(5), 1067–1092.
Krishnamurthy S. (2020). The future of business education: A commentary in the shadow of the COVID-19 pandemic. Journal of Business Research, 117, 1–5.
Kwok L., Mao Z., Huang Y. K. (2019). Consumers’ electronic word-of-mouth behavioral intentions on Facebook: Does message type have an effect? Tourism and Hospitality Research, 19(3), 296–307.
Kwok L., Tang Y., Yu B. (2020). The 7Ps marketing mix of home-sharing services: Mining travelers’ online reviews on Airbnb. International Journal of Hospitality Management, 90, Article 102616.
Kwok L., Xie K. L. (2018). Buyer–seller similarity: Does it lead to a successful transaction of peer-to-peer home-sharing services? International Journal of Contemporary Hospitality Management, 30(9), 2925–2944.
Kwok L., Xie K. L. (2019). Pricing strategies on Airbnb: Are multi-unit hosts revenue pros? International Journal of Hospitality Management, 82, 252–259.
Kwok L., Yu B. (2013). Spreading social media messages on Facebook: An analysis of restaurant business-to-consumer communications. Cornell Hospitality Quarterly, 54(1), 84–94.
Kwok L., Yu B. (2016). Taxonomy of Facebook messages in business-to-consumer communications: What really works? Tourism and Hospitality Research, 16(4), 311–328.
Kwok L., Zhang F., Huang Y., Yu B., Maharabhushanam P., Rangan K. (2015). Documenting business-to-consumer (B2C) communications on Facebook: What have changed among restaurants and consumers? Worldwide Hospitality and Tourism Themes, 7(3), 283–294.
Li Y., Xie Y. (2020). Is a picture worth a thousand words? An empirical study of image content and social media engagement. Journal of Marketing Research, 57(1), 1–19.
Lombard M., Snyder-Duch J., Bracken C. C. (2002). Content analysis in mass communication: Assessment and reporting of intercoder reliability. Human Communication Research, 28(4), 587–604.
Ma L., Zhan M. (2016). Effects of attributed responsibility and response strategies on organizational reputation: A meta-analysis of situational crisis communication theory research. Journal of Public Relations Research, 28(2), 102–119.
Mair J., Ritchie B. W., Walters G. (2016). Towards a research agenda for post-disaster and post-crisis recovery strategies for tourist destinations: A narrative review. Current Issues in Tourism, 19(1), 1–26.
Mehraliyev F., Choi Y., King B. (2020). Theoretical foundations of social media power in hospitality and tourism: A hierarchical model. Cornell Hospitality Quarterly, 62(2), 215–231.
Möller C., Wang J., Nguyen H. T. (2018). #Strongerthanwinston: Tourism and crisis communication through Facebook following tropical cyclones in Fiji. Tourism Management, 69, 272–284.
Nenonen S., Storbacka K. (2020). Don’t adapt, shape! Use the crisis to shape your minimum viable system–and the wider market. Industrial Marketing Management, 88, 265–271.
Omilion-Hodges L. M., McClain K. L. (2016). University use of social media and the crisis lifecycle: Organizational messages, first information responders’ reactions, reframed messages and dissemination patterns. Computers in Human Behavior, 54, 630–638.
Pantano E., Pizzi G., Scarpi D., Dennis C. (2020). Competing during a pandemic? Retailers’ ups and downs during the COVID-19 outbreak. Journal of Business Research, 116, 209–213.
Park H. (2017). Exploring effective crisis response strategies. Public Relations Review, 43(1), 190–192.
Perrin A., Anderson M. (2019, April 10). Share of U.S. adult using social media, including Facebook, is mostly unchanged since 2018. Pew Research Center. https://www.pewresearch.org/fact-tank/2019/04/10/share-of-u-s-adults-using-social-media-including-facebook-is-mostly-unchanged-since-2018/
Ritchie B. W., Jiang Y. (2019). A review of research on tourism risk, crisis and disaster management: Launching the annals of tourism research curated collection on tourism risk, crisis and disaster management. Annals of Tourism Research, 79, Article 102812.
Ritter T., Pedersen C. L. (2020). Analyzing the impact of the coronavirus crisis on business models. Industrial Marketing Management, 88, 214–224.
Roshan M., Warren M., Carr R. (2016). Understanding the use of social media by organisations for crisis communication. Computers in Human Behavior, 63, 350–361.
Sabate F., Berbegal-Mirabent J., Cañabate A., Lebherz P. R. (2014). Factors influencing popularity of branded content in Facebook fan pages. European Management Journal, 32(6), 1001–1011.
Sharma P., Leung T. Y., Kingshott R. P., Davcik N. S., Cardinali S. (2020). Managing uncertainty during a global pandemic: An international business perspective. Journal of Business Research, 116, 188–192.
Soboleva A., Burton S., Mallik G., Khan A. (2017). “Retweet for a chance to . . .”: An analysis of what triggers consumers to engage in seeded eWOM on Twitter. Journal of Marketing Management, 33(13–14), 1120–1148.
Spence P. R., Lachlan K. A., Rainear A. M. (2016). Social media and crisis research: Data collection and directions. Computers in Human Behavior, 54, 667–672.
Tracey J. B. (2020). Take care of yourself and look out for others. Cornell Hospitality Quarterly, 61(3), 236.
Utz S., Schultz F., Glocka S. (2013). Crisis communication online: How medium, crisis type and emotions affected public reactions in the Fukushima Daiichi nuclear disaster. Public Relations Review, 39(1), 40–46.
van Barneveld K., Quinlan M., Kriesler P., Junor A., Baum F., Chowdhury A., Junankar P. N., Clibborn S., Flanagan F., Wright C. F., Friel S., Halevi J., Rainnie A. (2020). The COVID-19 pandemic: Lessons on building more equal and sustainable societies. The Economic and Labour Relations Review, 31(2), 133–157.
Veil S. R., Buehner T., Palenchar M. J. (2011). A work-in-process literature review: Incorporating social media in risk and crisis communication. Journal of Contingencies and Crisis Management, 19(2), 110–122.
Xie K. L., Kwok L., Heo C. Y. (2020). Are neighbors friends or foes? Assessing Airbnb listings’ agglomeration effect in New York City. Cornell Hospitality Quarterly, 61(2), 128–141.
Yang Y., Zhang H., Chen X. (2020). Coronavirus pandemic and tourism: Dynamic stochastic general equilibrium modeling of infectious disease outbreak. Annals of Tourism Research, 83, Article 102913.
Zhai X., Zhong D., Luo Q. (2019). Turn it around in crisis communication: An ABM approach. Annals of Travel Research, 79, Article 102807.
Biographies
Linchi Kwok, Ph.D., is an associate professor in the Collins College of Hospitality Management at California State Polytechnic University Pomona (Cal Poly Pomona). He received his M.S. and Ph.D. degrees in Hospitality Administration at Texas Tech University and his MBA at Syracuse University. His research interests include information technology and service operations.
Jungwoo Lee, Ph.D., is the Director of Center for Work Science and a professor of smart technology management in the Graduate School of Information at Yonsei University, Seoul, Korea. He received his M.S. and Ph.D. degrees in Computer Information Systems at Georgia State University, and his MBA at Sogang University. His research interests include the impacts of smart technologies on work and services.
Spring H. Han, Ph.D., is an associate professor in the Graduate School of Management at Kyoto University. Her current research interests include sensory marketing, emotions and service experience, and health care service management. She has participated in various industry and academic projects in Korea, USA, China, Japan, and Russia.
Cite article
Cite article
Cite article
OR
Download to reference manager
If you have citation software installed, you can download article citation data to the citation manager of your choice
Information, rights and permissions
Information
Published In
Article first published online: June 30, 2021
Issue published: November 2022
Keywords
Authors
Metrics and citations
Metrics
Journals metrics
This article was published in Cornell Hospitality Quarterly.
VIEW ALL JOURNAL METRICSArticle usage*
Total views and downloads: 11761
*Article usage tracking started in December 2016
Articles citing this one
Receive email alerts when this article is cited
Web of Science: 37 view articles Opens in new tab
Crossref: 13
-
Exploring Secondary Crisis Response Strategies for Airlines Experienci...
-
Heterogeneity of crisis communication practices in hotels: Anti-COVID-...
-
Geo-sentiment trends analysis of tweets in context of economy and empl...
-
Determinants of Digital Transformation in the Hospitality Industry: Te...
-
Influence of chatbots on purchase intention in social commerce
-
Visual Social Media Communication of Italian Luxury Hotels and the Pan...
-
‘Riding out the pandemic’: The role of brand message appeals on social...
-
Strategic responses of chain hotels to COVID-19 from a situational cri...
-
Crisis communication strategies for health officials
-
Antecedents and consequences of new technology application behavior on...
-
Research on the Public Opinion Guidance Mechanism of Major Public Heal...
-
How Official Social Media Affected the Infodemic among Adults during t...
-
Examining restaurant purchase intention during crises: the role of mes...
Figures and tables
Figures & Media
Tables
View Options
View options
PDF/ePub
View PDF/ePubGet access
Access options
If you have access to journal content via a personal subscription, university, library, employer or society, select from the options below:
loading institutional access options
Alternatively, view purchase options below:
Purchase 24 hour online access to view and download content.
Access journal content via a DeepDyve subscription or find out more about this option.
