Sources of distrust: Airbnb guests' perspectives

The present study explores sources of distrust in the Airbnb context. The study focuses on Airbnb customers' negative reviews posted in English on Trustpilot's website. The search for posts was employed with the keyword ‘trust’ to find online narratives from customers who had negative experiences of trust with Airbnb. Of the 2733 online reviews screened, the study concentrated on 216 negative reviews. The data analysis followed the grounded theory approach, which resulted in two themes that reflect the sources of distrust: Airbnb's poor customer service and the hosts' unpleasant behaviour. The managerial implications are that Airbnb should invest additional resources into minimising customers' negative experiences by focusing on trust-based relationships and maintaining quality in its core service elements. When customers report their complaints, their concerns should be addressed with prompt apologies, positive actions, and a willingness to compensate these customers to neutralise their distrust in the company.


Introduction
Recently, the sharing economy has been rapidly permeating the lodging industry (Sthapit & Jiménez-Barreto, 2018a). 'Sharing economy' is an umbrella term that covers the sharing of consumption through online platforms (Hamari, Sjöklint, & Ukkonen, 2016), and Airbnb is the world's largest accommodations provider in the sharing economy (Mody, Suess, & Lehto, 2019). On the supply side, Airbnb enables private owners or tenants of houses, apartments, and spare rooms to offer their premises to visitors for short-term rentals (Sigala, 2017). On the demand side, Airbnb fulfils travellers' needs, such as accommodations with lower prices and opportunities to interact with the local community (Guttentag, 2015). Existing research has examined brand personality (Lee & Kim, 2018), consumer experience (Pappas, 2019), value co-destruction (Sthapit, 2018a;Sthapit & Jiménez-Barreto, 2018b), memorability (Sthapit & Jiménez-Barreto, 2018a), and sharing in the Airbnb context (Sthapit & Jiménez-Barreto, 2018c). The accommodation service and its delivery process are dynamic, intertwined with the consumption process, and vulnerable to various factors. Thus, the actual accommodation experiences tend to be different from those described by peer customers (Zekanovic-Korona & Grzunov, 2014) and may result in consumers' distrust in Airbnb, which is the primary focus of this study.
Customers of the sharing economy-for example, Airbnb-are exposed to risks other than monetary loss (Ert, Fleischer, & Magen, 2016) and untrustworthy strangers; guests may be faced with the risk of unreliable hosts or even personal security (Huurne, Ronteltap, Corten, & Buskens, 2017). In most cases, the host rents rooms to strangers (Ert et al., 2016), and the quality of the accommodation service is highly dependent upon the host's competence in hospitality (Zhang, Yan, & Zhang, 2018). As a result, many unforeseen incidents may occur, as guests cannot determine one another's reliability in advance (Sun, Liu, Zhu, Chen, & Yuan, 2019). For example, a recent unfortunate incident involved the sexual assault of a nineteen-year-old boy by his Airbnb host during his stay in Madrid (Lieber, 2015). Although the notion of sharing presumes trust between parties (Lee, 2015;Parigi & Cook, 2015), such unpleasant experiences may certainly occur, subsequently lead to distrust, and furthermore discourage travellers from choosing Airbnb as an alternative to the conventional accommodation types (So, Oh, & Min, 2018). Recent studies have identified distrust as one of the major barriers surrounding consumers' use of Airbnb (Tussyadiah & Pesonen, 2018) and, in some cases, the only constraint that significantly predicts the overall consumer attitude towards Airbnb (So et al., 2018).
Although the trust between providers and consumers forms the basis of a successful transaction in the sharing economy (Chenga, Fua, Sunb, Bilgihanc, & Okumu, 2019;Ert et al., 2016;Guttentag, 2015;Zervas, Proserpio, & John, 2015), the consumer-centric (tourist) perspective is lacking in the underlying sources that cause distrust among guests during their Airbnb lodging experiences. To bridge the current gap in the existing literature on the sharing economy, the present study aims to explore the sources of distrust as experienced by Airbnb guests. This study is based on tourists' negative online reviews of their Airbnb experiences. A total of 2733 review posts were manually extracted from Trustpilot's website, wherein the keyword 'trust' was included in the search to find online narratives from customers who had endured negative experiences of trust with Airbnb. This study contributes to the literature on the sharing economy (particularly Airbnb) and, more specifically, the study proposes a conceptual framework geared towards a more holistic understanding of the sources of distrust in the Airbnb context that comprises two components: Airbnb's poor customer service and the hosts' unpleasant behaviour. In addition, although it is generally presumed that experiences are positive encounters, negative experiences are certainly possible (Sthapit, 2013). Interestingly, when experiences are described and defined, researchers generally imply positive or pleasant events or feelings (Oh, Fiore, & Jeoung, 2007;Pine & Gilmore, 1998). According to Pine and Gilmore (1998), an instance of poor service easily converts into an experience, thus creating a memorable encounter of a negative kind. Therefore, the study responds to the recent call from academics to additionally focus on Airbnb guests' negative experiences (Sthapit, 2018a;Sthapit & Jiménez-Barreto, 2018b) despite some studies indicating that Airbnb possesses remarkable customer satisfaction levels, which is also evidenced by its user reviews (Ert et al., 2016). In the same vein, some recent studies have identified an extremely positive bias among Airbnb's ratings (Bridges & Vásquez, 2018;Zervas et al., 2015). Moreover, identifying the sources of distrust in the present study is important because it covers new theoretical ground as well as poses practical value by both providing new insights into Airbnb's hosts and management concerning what causes distrust among Airbnb guests and exploring ways to minimise its customers' negative experiences of trust.

The sharing economy and Airbnb
The sharing economy is associated with the peer-to-peer (P2P) business model, which allows for the shared creation, production, distribution, and consumption of products and services among individuals (Tussyadiah & Pesonen, 2015). In other words, the sharing economy connects individuals who have excess property capacity to tourists who require accommodation using an online platform maintained by a thirdparty company (Botsman & Rogers, 2010). Alongside the rise of the sharing economy, sharing private space via platforms, such as Airbnb (Guttentag, 2015), became popular. Airbnb, founded in 2008, is a sharing economy platform for the short-term exchange of rooms and apartments for a fee (Airbnb, 2017). Airbnb's accelerated growth represents a 'disruption' as an innovative business model that is currently drawing at least some segment of the travel market away from hotels (Guttentag, 2015). In addition, Airbnb operates in most locations with minimal regulatory controls (Ert et al., 2016).

Trust, distrust, and Airbnb
Trust refers to one's willingness to rely on an exchange partner-that is, a reliable individual who keeps promises (Wang, Law, Hung, & Guillet, 2014). Trust helps decrease the anxiety, uncertainty, and vulnerability related to transactions, thus resulting in greater satisfaction-especially in cases of quite complex and experiential services (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004). Trust in the Airbnb context means accepting a position of vulnerability and trusting that the exchange partner will fulfil his or her part (Satama, 2014). However, trust on the Airbnb platform-that is, institution-based trust-should not be confused with disposition to trust, which is defined as trust in hosts or guests (Liang, Choi, & Joppe, 2018). Institutionbased trust can be defined as a buyer's perception that appropriate conditions are in place to facilitate transaction success among the marketplace's sellers (Pavlou & Gefen, 2004). McKnight, Choudhury, and Kacmar (2002) defined disposition to trust as 'the extent to which a person displays a tendency to be willing to depend on others across a broad spectrum of situations and persons ' (p. 339). In addition, as trustworthy behaviours are repeatedly demonstrated, trust levels in the relationship between the service provider and the customer are elevated (Coulter & Coulter, 2003).
Distrust is an intuitively negative feeling about another individual's conduct (Tussyadiah & Pesonen, 2018), defined as an unwillingness to become vulnerable to the trustee based on the belief that the trustee will behave in a harmful, neglectful, or incompetent manner (Benamati, Serva, & Fuller, 2010). In the Airbnb context, distrust is defined as the lack of interpersonal trust between the guest and the host, lack of trust towards technology, and lack of trust towards Airbnb (Tussyadiah & Pesonen, 2018). Distrustful relationships are associated with caution, defensiveness, and vigilance (Lewicki, Mcallister, & Bies, 1998). The sources of distrust in relationships are linked to a lack of cooperation (Cho, 2006), avoidance of interaction (Bies & Tripp, 1996), unwillingness to share views and preferences (Bijlsma-Frankema, 2004), reduced information sharing (Gillespie & Dietz, 2009), and intergroup conflicts (Fiol, Pratt, & O'Connor, 2009;Tomlinson & Lewicki, 2007). Distrust blocks business exchanges, especially those occurring in online businesses wherein transactions are not interpersonal (Komiak & Benbasat, 2008). In addition, distrust creates conditions of consumer ambivalence, including insecurity and anxiety, which convert active consumers to passive consumers and negatively affect their behaviour (Elbeltagi & Agag, 2016;Moody, Galletta, & Lowry, 2014).
Several authors have found that trust and distrust are composed of the same three facets: benevolence, integrity, and competence (Casaló, Flavián, & Guinalíu, 2007;Cho, 2006). First, according to Mayer, Davis, and Schoorman (1995), benevolence is the belief that the trustee wants to do good to the trustor aside from an egocentric profit motive. Benevolence additionally conveys the altruistic notion that one party wants to do good to the other party (Jarvenpaa, Knoll, & Leidner, 1998), especially when one party feels cared for by the other party (Mayer et al., 1995). The level of benevolence perceived by the potential guests reflects service providers' quality (Vázquez-Casielles, Suárez-Álvarez, & del Río-Lanza, 2013). On the other hand, benevolence distrust is the trustor's (the Airbnb guest) belief that the trustee (the Airbnb host) does not care about the trustor and is not motivated to act in the trustor's best interest (McKnight & Chervany, 2001).
Second, integrity is the belief that the trustee adheres to a set of principles that the trustor finds acceptable (Mayer et al., 1995). Integrity stresses the trustee's responsibility to follow the norms or rules of an organisation and possess a core set of values that guide his or her behaviours. In other words, the trustor believes the trustee will always keep good-faith agreements, tell the truth, act ethically, and fulfil his or her promises (McKnight & Chervany, 2001). Integrity is a key element when building successful relationships, particularly in the hospitality sector, which encompasses a consistent and harmonised approach to guests in all aspects of service delivery (Janowicz-Panjaitan & Krishnan, 2009). In addition, benevolence and integrity are more deeply concerned with the trustee's responsibilities than with the trustful relationship (McKnight & Chervany, 2001).
Third, competence refers to the trustee's ability to perform as expected by the trustor (Pavlou & Fygenson, 2006) as well as the trustee's skills, competencies, and abilities to exert influence within some specific domain or context (Mayer et al., 1995). Competence is domain specific; for example, a user who provides satisfying recommendations for cars that are for sale may not be an expert at buying clothes (Fang, Guo, & Zhang, 2015). When a guest recognizes competence and expertise, the perceived risk associated with a provider is reduced (Vázquez-Casielles et al., 2013).
According to Ert et al. (2016), sharing economy platforms such as Airbnb include additional risks aside from monetary aspects. The authors stress that 'the mere act of sharing a home with a stranger can be E. Sthapit and P. Björk Tourism Management Perspectives 31 (2019) 245-253 risky' (Ert et al., 2016, p. 63). Such risks and unpleasant experiences may lead to distrust and discourage travellers from choosing Airbnb as an alternative to conventional accommodation types (So et al., 2018). In fact, a commonly cited constraint factor with respect to Airbnb adoption is perceived risk (So et al., 2018)-the general belief in potentially negative results due to a specific purchase (Kim, Ferrin, & Rao, 2008). In addition, Mao and Lyu (2017) describe the perceived risk associated with Airbnb as a subjective expectation of a potential loss when pursuing a desired result. Some studies have identified that perceived risk negatively affects consumer behaviour (Chiu, Wang, Fang, & Huang, 2014;Yang, Liu, Li, & Yu, 2015), which-in this context-represents distrust. Given that Airbnb is a third-party platform that offers online-matching accommodation services between sellers and buyers, risk is a very important factor that influences its behavioural intention. In addition, Airbnb consumers have no choice but to estimate the risk of their transactions from the available information and communications because they cannot experience the actual service prior to their arrival at their chosen properties (Liang et al., 2018). In the same vein, most P2P accommodation platforms are suffering scanty trust, which has largely inhibited many potential customers' participation in their services (Ert et al., 2016;Wu & Zeng, 2017). Trust has been identified as a prerequisite for the creation and preservation of longterm relationships between companies and consumers (Morgan & Hunt, 1994), especially in the context of service markets (Martinez & del Bosque, 2013). Recent studies indicate that distrust is the most frequently cited barrier to P2P accommodation in a sharing economy, which includes the basic mistrust among strangers and privacy concerns (So et al., 2018;Tussyadiah & Pesonen, 2018).

Data collection
The researcher collected and analysed data from February to June 2018. For the data collection method, the study employed non-participant observations in the form of netnography. In today's service industry, customers are becoming increasingly active online before, during, and after their interactions with service providers, thus creating large masses of information concerning their activities and experiences (Berthon, Pitt, Kietzmann, & McCarthy, 2015;Wuenderlich et al., 2015). As customers share more of their experiences online, researchers are finding netnography to be increasingly useful for exploring these consumption-related experiences. Review sites for tourism, travel destinations, restaurants, products, and services are popular domains for netnographers' data collection purposes (Heinonen & Medberg, 2018). Netnography has gained popularity more recently in tourism studies (Sthapit, 2018b;Sthapit & Björk, 2018), and Kozinets (2015) defines netnography as a 'more human-centered, participative, personally, socially and emotionally engaged vector' (p. 96). The method was developed in the context of the increasing popularity of virtual communities wherein people share their interests and build social ties (Kozinets, 1999).
Netnography was considered appropriate for this study because it is relatively rapid, simple, and inexpensive; it allows access to naturalistic, unprompted insider experiences, perspectives, and reflections as well as captures the exchange of tourism information on the Internet (Mkono & Markwell, 2014). In addition, especially in the case of sensitive research topics-in this context, distrust-netnography's unobtrusiveness might be necessary for eliciting relevant data (Keeling, Khan, & Newholm, 2013;Langer & Beckman, 2005). Moreover, netnography is faster, simpler, and less expensive than traditional ethnography and more naturalistic, objective, and unobtrusive than focus groups or interviews (Wu & Pearce, 2014). Furthermore, this online approach facilitates access to emic voices and the studying of a larger number of individuals (Kozinets, 2010). According to Rageh, Melewar, and Woodside (2013), netnography 'excels at telling the story, understanding complex social phenomena and assists the researcher in developing themes from the respondents' point of view ' (p. 130). Data can be collected from numerous sources, such as online tourist reviews, and are analysed thematically (Catterall & Maclaran, 2001). Given that a significant amount of data collection occurs through the data shared freely on the Internet, Kozinets (2010) suggests pure netnography is entirely complete within itself and requires no offline ethnographic research. Many additional studies acknowledge the importance of this methodology (Bartl, Kannan, & Stockinger, 2016).
This study follows Kozinets's (2002) guidelines for conducting netnography. The first step involves identifying potential communities and then carefully selecting either one or several communities for data gathering based on predetermined criteria. The researcher then enters the community (with or without introducing his or her presence), gathers data by observing and participating in the community's interactions, and develops an insider understanding of the community's culture. A netnography in the form of non-participant observation is used in this study and is based on online customer reviews that contain detailed information about users' experiences with Airbnb. The reason for choosing non-participant observation is the undesirable influence of outsiders on the group (Elliott & Jankel-Elliot, 2003). The researcher intensively reviewed websites that offer online consumer reviews that detail users' Airbnb experiences. The search for a convenient website for this study was conducted on Google using combinations of the keywords 'tourist Airbnb experience' and 'visitor Airbnb experience'. Furthermore, the Trustpilot website was identified as relevant to this study's purposes.
The second step in netnography involves data collection. This research employed one keyword ('trust') when searching Trustpilot fora to avoid generating overwhelming amounts of data, and each narrative comprises one entry. Of 2733 online posts screened, the analysis focused on 216 negative Airbnb reviews. Irrelevant reviews and messages were omitted from the analysis to ensure analytical depth as well as a fixed focus on the topic.
The third step in netnography is linked to the ethics of the researcher's role. When researchers enter an active online community as participant-observers querying and directing communication, they should fully disclose their identities and motives, obtain informed consent, and conduct member checks with key informants (Wu & Pearce, 2014). However, when accessing blogs or review sites as nonparticipant-observers, there exists no compelling need to communicate research objectives or obtain consent, as the data are available on public (sometimes anonymous) Internet fora and the entries were often posted months or years in the past (Mkono, 2012). The covert netnographic approach applied here supports a highly personal and social distance between the researchers and subjects (Arsal, Woosnam, Baldwin, & Backman, 2010). In addition, according to Kozinets (2010), when conducting netnography in open online fora, it is not necessary that a researcher fully disclose his or her presence to the online community members with whom he or she is studying; this optional concealment guarantees the informants' confidentiality and anonymity or seeks and incorporates feedback from the online community members. To ensure the study's trustworthiness, we have explained all the phases of our research in detail (how the data were collected, categorised and analysed), and the website utilised in this study is an established public forum of communication, meaning consent was unnecessary for analysing those public postings.

Data analysis
A grounded theory research design (Glaser & Strauss, 1967) was used to analyse the collected data. This approach involves a continual interplay between data collection and theoretical analysis to examine causal factors and patterns of experience (Riley, 1995). In other words, the approach enables that understandings be formed into concepts and theories without an a priori definition. The concepts, theories, or models were thus developed from the participants' socially constructed knowledge (Daengbuppha, Hemmington, & Wilkes, 2006).
The collected data were first filtered through a process of open coding to identify discrete concepts, which are the basic unit of grounded theory analysis (Daengbuppha et al., 2006). This first step involves the breakdown of data into distinct units of meaning (Sthapit, 2018b) by naming words, lines, and segments of data. Charmaz (2006) suggests selecting the most useful analytical codes-a procedure that was guided in this study by analytical questions such as 'What do the data suggest?' (Charmaz, 2017). Each negative Airbnb review post was read and analysed separately to identify the emerging ideas and views mentioned by each user (Nunkoo & Ramkissoon, 2016) as well as extract specific information and each participant's views (Sthapit, 2018b). This process was important because it helped identify the equivalent meanings tourists attributed to the phenomenon under study (Mehmetoglu & Altinay, 2006). Table 1 illustrates how coding was performed in practice; the first column illustrates the raw data extracted from the Trustpilot website, while the second column conveys the initial codes extracted from the raw data through open coding.
This procedure was followed by axial coding to find relational patterns between the concepts and reduce the database to a small set of themes or categories that subsequently characterised the process under study (Matteucci & Gnoth, 2017). The axial coding process involves developing an understanding of the conditions that give rise to the categories and their contexts, interactions, and consequences. Then, by relating categories to one another, researchers are able to elaborate substantive theoretical propositions (or concepts) (Matteucci & Gnoth, 2017). In this stage, data are compared, while similar incidents are grouped together and assigned the same conceptual label (Nunkoo & Ramkissoon, 2016); for example, initial codes derived from the line-byline coding process, such as 'no easy phone/email support', 'can't refund the money', 'customer service is a nightmare', and 'getting a straight answer might just be impossible' were grouped together to form a subtheme labelled as 'poor customer service'. This process helped more thoroughly describe Airbnb guests' views of sources of distrust. As described in Table 1, two subthemes were identified and subsequently categorised into two main themes: (1) Airbnb's poor customer service and (2) the hosts' unpleasant behaviour.
Selective coding was performed to forge emerging structures and elaborate upon the sources of distrust. This coding process involves identifying core categories that represent the main research theme (Daengbuppha et al., 2006) by integrating the categories derived from the open and axial coding processes to form a conceptual framework. The codes and categories are explored further by rereading the coded statements. During the data analysis, the concepts and relationships revealed by the coding processes were compared with the extant literature. This stage involved noting consistencies and identifying research ideas or concepts (Sthapit, 2018b). In other words, the researcher reviewed the collected data by determining whether or not the newly developed categories remained constant when the data were analysed specifically for these categories (Elliott & Lazenbatt, 2005). In this stage, the transcripts analysis identified two sources that characterised distrust in the Airbnb context. Overall, throughout the data collection and analysis processes, the codes' validity was maintained by interpretively coding the text and constantly checking the codes and exemplars as the study was developed. As such, the iterative nature of the coding process promoted consistent treatment.
In accordance with the logic of a grounded theory approach (Glaser & Strauss, 1967), the identified factors were constantly compared (Maxwell, 1996;Miles & Huberman, 1984). Ethnography and grounded theory are methodological complements, as ethnographic studies can provide a thick description that constitutes useful data for a grounded Table 1 The coding process in practice.

Reviewers views (extracted from the analysed review posts)
Open coding (line-by-line coding) Subthemes (axial coding) Main themes (selective coding) "…Helpdesk (there is no easy phone/email support). This has caused stress and inconvenience … I would strongly advise that anyone using this service do not do so unless they have had correspondence with owners that they are 100% sure they can trust"; "We rented an apartment on Mykonos for 10 days … the place was horrible … On the fourth day, they replied … they can't refund the money, even if we didn't stay … our vacations were screwed up. Don't trust Airbnb for your holidays …"; "The customer service is a nightmare, and getting a straight answer might just be impossible … Airbnb should not be trusted' 'The building was not maintained. The beds were horrible, the water cuts off and the TV never worked … The host should not be trusted…"; "Absolutely no accountability from the host … worst service experience I have ever had … I will never trust Airbnb for accommodations again"; "The host, Diletta, advertised a remodelled apartment in Rome with a private court. In reality, it was a basement-level dump … We cancelled on arrival. After several days of exchanges with Airbnb, they refunded part of the money, but $101 of the cleaning and service fees are lost. Do not trust Airbnb …" "Absolutely no accountability from host. My family and I booked a house for the Fourth of July back in May … reservation confirmed by owner through emails, and on June 30, I get an automated email from Airbnb saying the host has cancelled our reservation …" no easy phone/email support, can't refund the money, customer service is nightmare, getting a straight answer might just be impossible Building was not maintained, the water cuts off, TV never worked, no accountability from host, basementlevel dump, reservation cancellation by the host theory analysis (Glaser & Strauss, 1967). Part of this compatibility is derived from the similarities of the characteristics between the two methods. As a naturalistic form of inquiry, ethnography entails observing and analysing behaviour in naturally occurring conditions (Belk, 1988), while grounded theory similarly performs most efficiently with data generated in natural settings (Robrecht, 1995). In addition, the iterative process-a comparison between and across data and theory (data comparison) as a form of informant triangulation-served to ensure the study's quality (Decrop, 2004). Moreover, to ensure the findings' credibility, this study included participants' actual testimonies during the coding process (Chiovitti & Piran, 2003).

Findings and discussion
The findings from the grounded theory research design were analysed to determine customers' negative experiences with Airbnb and identify the common sources of distrust. The coding process inductively identified two emerging sources of distrust as experienced by these Airbnb guests.

Poor customer service from Airbnb
Of the 216 negative reviews posted, 194 (89.81%) were associated with poor customer service experiences that provoked distrust in Airbnb. The identified interpretive codes demonstrate the role of poor customer service as a source of distrust as experienced by Airbnb guests, including: 'there is no phone/email support', 'simply ignored', 'can't get any help from Airbnb', 'contact with this organisation is virtually impossible', 'Airbnb failed to help us in any useful way', 'Airbnb made no effort to make things right for me', 'can't refund the money', 'customer service is a nightmare', 'wait for a response…usually weeks', 'I no longer can place any trust in…even their customer service', 'Airbnb… they do not care', 'customer service team just doesn't exist', 'no one from Airbnb contacted me', 'was not resolved until 26 hours after my initial submission', 'just denied that was their problem', 'Airbnb will not help you', 'dismissed my claims', 'worst customer service', 'bad customer service', and 'customer service is poor'. These codes are highlighted by the following responses from four reviewers.
We tried to book an apartment through the Airbnb website…The apartment was not available, but Airbnb held and then refunded approximately $700. On enquiring about a second apartment without entering credit-card details, Airbnb held off approximately $850 …It is now seven days, and the funds have not been released despite corresponding with their 'Help' desk. This has caused stress and inconvenience…I would strongly advise that anyone using this service do not do so unless they have had correspondence with owners that they are 100% sure they can trust…I would not use Airbnb or recommend Airbnb to anyone else…(review published on 10 June 2013).
We rented an apartment on Mykonos for 10 days…the place was horrible. The photos were different, and there were half the amenities available. The place was not cleaned properly. We didn't check in and called Airbnb. We also sent them photos and several emails. We rented another room for two days, waiting for some reply, and we left the third day. On the fourth day, they replied…they can't refund the money, even if we didn't stay…our vacations were screwed up. Don't trust Airbnb for your holidays… (review published on 19 August 2013).
The customer service is a nightmare. My account had an issue while over $3000 of my money was tied to my account. I called customer service, and they refused to talk to anyone for anything other than general questions. Instead, I had to email their team and wait for a response, which in my case, usually weeks, and it was always a generic-response email…my credit-card company stepped in and got my money back. Save yourself a future headache…Airbnb should not be trusted (review published on 18 January 2018).
I was tricked by a[n] Airbnb host into reserving a place which was absolutely nowhere it said it was on the map…The host agreed to refund my money but later refused and this is where is really gets ugly from the Airbnb's side…they only contacted the host to hear their side of the story…I thereupon rang Airbnb 4 times and have sent them countless emails though I am simply ignored…I therefore wrote a negative feedback to warn other users about the listing but my feedback was simply not published…I am so sincerely dis-appointed…I was tricked by a false listing and subsequently Airbnb made my experience hell. I no longer can place any trust in their maps, listings or even their customer service (review published on 11 April 2018).
After experiencing a service failure, customers usually complain to the service provider to mitigate their stress and protect themselves. In this context, many Airbnb guests adopted this strategy by contacting the company's customer service department; however, Airbnb's inadequate responses and poor interactions with its customers indicates a low level of benevolence from the customer service personnel towards their customers, which generated high levels of ambivalence and uncertainty among these customers that represent psychological distress (Moody et al., 2014).
The lack of cooperation from Airbnb's customer service personnel aroused doubts among guests about the personnel's willingness to share information and act honestly and ethically (integrity distrust), which included their ability to perform as expected (competence distrust). This further negatively influenced customers' self-esteem, and their lack of success in obtaining answers led to confusion and their perceived reduction in self-efficacy. Self-esteem has been defined as an overall appraisal of one's self-worth (Rosenberg, 1965), and studies indicate that the extent to which service providers are concerned about consumer esteem in trust recovery initiatives signals corporate sincerity in consideration of consumers' interests (Xie & Peng, 2009). Self-efficacy is defined as 'one's capability to perform a task' (Gist, 1987, p. 472) and is considered a critical motivational construct that influences individual emotional reactions, effort, coping mechanisms, and persistence (Gist, 1989;Gist & Mitchell, 1992). As a result, guests' distrust in Airbnb was amplified, and some noted they would stop using the service. This decision can be linked to institution-based distrust; in other words, these guests experienced what they viewed as unsatisfactory outcomes from a consumer transaction made under presented conditions, including its structures and regulations.

Hosts' unpleasant behaviour
Of the 216 negative reviews posted, 22 (10.19%) emphasised Airbnb hosts' unpleasant behaviour as the cause of their distrust in those hosts. The following four reviews describe negative experiences that can be linked to Airbnb hosts' unpleasant behaviour.
The host, Diletta, advertised a remodelled apartment in Rome with [a] private court. In reality, it was a basement-level dump, one small dirty window in the kitchen, second-floor ceiling six feet high, missing electric outlets, mouldy bathroom, the private court turned out to be an interior dilapidated service yard where garbage cans usually are. We cancelled on arrival. After several days of exchanges with Airbnb, they refunded part of the money, but $101 of the cleaning and service fees are lost. Do not trust Airbnb. (review published on 11 October 2016).
The building was not maintained. The beds were horrible, the water cuts off and the TV never worked. The rooms were dusty and dirty. One of the rooms had bed bugs. The beds are small full-size. The owner scammed up by marking up fares on tour prices. No Wi-Fi connection as advertised. The elevator did not work. They did not E. Sthapit and P. Björk Tourism Management Perspectives 31 (2019) 245-253 replace toilet paper, nor cleaned the rooms unless asked… The host should not be trusted. The place is unsuitable; don't be deceived… (review published on 27 February 2018).
Absolutely no accountability from host. My family and I booked a house for the Fourth of July back in May…reservation confirmed by owner through emails, and on June 30, I get an automated email from Airbnb saying the host has cancelled our reservation. When trying to call and email the host, she would not respond. Finally, she responded through a text message, saying she had double-booked by accident and she was giving the house to the other party. There are literally no other available houses now, and Airbnb…the worst service experience I have ever had…I will never trust Airbnb for accommodations again (review published on 1 July 2014).
Our trust in what Airbnb advertises…is forever gone…the host, who works for…states: Potential for noise -The Metro Light Rail runs near the property and is heard several times during the day onlyreality is that there is constant noise all day and night long. Resultno sleep … We left two days early…when the hosts never return my calls, I went to the apartment complex office to see if I could get another contact number to call. I found out there is another rental agency that subleases the apartment in this complex…I contacted them and they said the apartment was not one of theirs nor was the host one of their employees. We asked the host for a refund and were turned down (review published on 14 April 2018).
Hosts' unpleasant behaviour, such as their failure to treat guests with respect, last-minute reservation cancellations, and dishonesty, can be linked to integrity distrust, which is a trustor's (the Airbnb guest) belief that the trustee (the Airbnb host) has failed to uphold a goodfaith agreement that a promised service will be provided. This failure occurs when the trustee acts dishonestly or cheats the trustor out of the promised services (McKnight et al., 2002). Studies indicate that a single act of dishonesty is strong evidence of an individual's lack of integrity because only those who deeply possess this attribute will act in such a way (Kim, Dirks, Cooper, & Ferrin, 2006). In addition, some of the posted reviews mentioned the host's lack of communication and maliciously withholding of information related to booking, thus causing the guest to further question the host's integrity and competence with the Airbnb process. The literature recognizes the importance of (timely) communication as an effective approach to remove mutual suspicion, unify expectations, and subsequently facilitate trust between the host and guest (Moorman, Zaltman, & Deshpande, 1992;Yousafzai, Pallister, & Foxall, 2005). In the same vein, one experiential driver that fuels the sharing economy's growth-particularly that of Airbnb-includes the customer's desire for social interactions with the host (Guttentag, 2015). Moreover, hosting on Airbnb involves communicating with guests (Sthapit & Jiménez-Barreto, 2018b), and communication can establish trust between the host and guest as well as minimise uncertainty (Guttentag, 2015).
The lack of a mutually beneficial relationship between a host and guest, a host's opportunistic behaviour, and a host's lack of concern for his or her guests generated a low level of benevolence among the reviewed guests. As a result, the review posts reported losses of guests' physical, financial, and temporal resources; for example, interpretive codes such as 'lost about $130 in value in my reservation', 'no refund', 'lost enough money with them', 'overcharged', 'agreed to partial refund on difficult terms', and 'refused a refund' are indicative of these guests' loss of financial resources. These losses can be further linked to service failure that results in customers' economic and/or social losses through exchanges (Smith, Bolton, & Wagner, 1999). Service failure can be defined as service performances that fall below customer expectations (Sparks & Fredline, 2007) and can occur within the process and outcome of service delivery (Lewis & McCann, 2004). In this study's context, the level of service failure was serious and included a host's lastminute reservation cancellation, a booked apartment's poor condition, and a host's sharing of false information.
These review posts can be linked to hosts' poor communication skills, poor service quality, and lack of hospitableness. First, interpretive codes such as 'host did not respond and cancelled my reservation', 'host did not respond', 'never answered', 'host isn't replying', 'never got a response', and 'host never returned my calls' indicate poor communication that the reviewed guests endured during their Airbnb experiences. Recent studies indicate that poor communication leads to service failure in the Airbnb context, thus consequently making guests feel devalued as customers (Sthapit & Jiménez-Barreto, 2018b). Second, certain keywords can be linked to the poor service quality the reviewed guests experienced, including 'misrepresentation of the apartment', 'no private entrance', 'no Wi-Fi connection as advertised', 'beds were horrible', 'building was not maintained', 'rooms were dusty and dirty', 'the place was filthy', and 'had no windows'. In the Airbnb context, the service quality cannot be determined until it is experienced (Sthapit & Jiménez-Barreto, 2018b), which results in greater uncertainty among prospective guests (Wu, Ma, & Xie, 2017;Zhang et al., 2018), as has been identified in this study. According to Parasuraman, Zeithaml, and Berry (1985), service quality is 'a global judgment, or attitude, relating to the superiority of the service' (p. 16). Offering quality service is an essential strategy for success and survival in a competitive environment (Zeithaml, Parasuraman, & Berry, 1990) and is judged by factors such as tangibles, reliability, responsiveness, assurance, and empathy (Parasuraman et al., 1985). In addition, service-outcome quality strongly influences distrust (McKnight et al., 2002). Third, interpretive codes such as 'was tricked by a host', 'host was abusive', 'bad hosts', 'unacceptable behaviour of host', 'owner would not help', 'selfish and inconsiderate host', 'no one cared', and 'no accountability from host' are indicative of Airbnb hosts' lack of hospitality towards their guests. Tussyadiah and Pesonen's (2015) study indicates the significance of the host's hospitality in the Airbnb context from the guest's perspective, given that the sharing economy challenges the hotel industry along experiential lines (Mody et al., 2019).

Conclusion
This study contributes to the existing literature on the sharing economy, particularly Airbnb, and links the concept of distrust. In addition, the current study provides comprehensive analysis of the sources of distrust based on negative reviews that Airbnb guests posted online regarding their experiences: Airbnb's poor customer service and the hosts' unpleasant behaviour. Airbnb's poor customer service generated distrust in the company (institution-based distrust), while the hosts' unpleasant behaviour resulted in a disposition of trust (distrust in the host). Airbnb's response to service failures was lacking and generated trust violation. Customers felt cheated and experienced delayed service recovery after contacting the company's customer service department. In other words, the service recovery process was a failure and evoked psychological discomfort as well as negative experiences of trust. The interpretive codes, particularly, 'stress', 'totally unacceptable', 'inconvenience', 'headache', 'deceived', 'screwed', 'worst service experience', and 'unsafe' suggest that the guests experienced psychological discomfort alongside losses in self-esteem and self-efficacy. In addition, hosts' unpleasant behaviour evoked guests' distrust in their hosts due to discomfort. The hosts were unable to meet guests' expectations, thus leading to a service failure that resulted in guests' unexpected resource losses of time and money. Moreover, the identified sources can be linked to a lack of interactional justice-that is, how an individual is treated while a procedure is being enacted (Kickul, Gundry, & Posig, 2005).
Airbnb's marketing pitches that promised its customers would 'feel at home anywhere you go in the world' by 'offering a unique experience' did not hold up for these reviewed guests. Given that distrust-which creates anxiety and stress-can lead to severe behavioural responses (Kramer, 1998), some disappointed and irritated guests who vowed never to trust Airbnb again engaged in negative electronic word of mouth and chose other online travel agencies when searching for and booking future accommodations (e.g., Booking.com). This aspect is highlighted by the following four reviews: 'I will not be referring them to anybody, in fact I will be warning everyone I know and meet from now on' (review published on 27 October 2016); '…after 3 years of loyalty good bye. You have no right to stay in business' (review published on 12 January 2017); 'Save yourself a future headache, try Homeaway. They are a much better company with great customer service. Airbnb should not be allowed to operate' (review published on 28 February 2017); and 'I'm very disappointed in a company built on trust dismissing my claims so easily. What a bummer...good bye Airbnb, you've lost a customer' (review published on 3 April 2017). This finding supports studies that indicate distrust predominantly creates negative consumer outcomes by increasing consumer engagement in negative electronic word of mouth and lowering repeat purchase intention (Ahmad & Sun, 2018). Such negative experiences of trust suggest serious ramifications for Airbnb and a competitive advantage for hotels that have standardised procedures for dealing with such issues. In today's conversations of booking for holiday travel, individuals often ask one another: 'Did you book a room via Airbnb?' However, distrust in Airbnb may alter that question to: 'At which hotel did you stay during your holiday?' From a managerial perspective, Airbnb should first invest more resources into minimising its customers' negative experiences of trust by clearly defining the hosts' responsibilities. Second, hosts should engage in active communication with their guests, such as by clarifying facts related to the booking and disclosing of updated information about the lodging's condition in the pre-trip booking process both online and face to face. Positive online and offline communication may help develop trust between the host and the guest, as reciprocal interactions strengthen closeness and trust between two individuals (Reis & Shaver, 1988). The host's online information, such as his or her profile, public pictures, and accommodations, should be credible in order to demonstrate the host's integrity, competence, and benevolence during the booking process as well as build trust between the host and the guest. Third, hosts should focus on remaining well-mannered when welcoming guests to their rentals; in other words, hosts should treat guests in a friendly manner, which includes resolving any problems they face in relation to the accommodations. This signifies the assertion of the host's willingness to assume relevant responsibility. Fourth, given that benevolence plays a vital role in the development of distrust (Kim et al., 2006), when customers report their complaints, they should be addressed with prompt apologies by the company's customer service department, which may lead to the favourable impression that Airbnb orients itself towards solving problems. According to both social exchange theory and justice theory, an apology reallocates esteem as an important social resource in an exchange relationship (Walster, Berscheid, & Walster, 1973); for example, studies indicate that an apology issued by a provider following a service failure can enhance consumers' perceptions of interactional justice and improve post-recovery satisfaction (Smith et al., 1999). Fifth, after a service failure is exposed, the provider's willingness to provide financial compensation to remedy what has occurred to a certain extent (e.g., loss and suffering)-in the form of a refund or premium package, for instance-may be an effective trust-repair measure to neutralise the distrust Airbnb guests feel towards the company and may further lead to service recovery, consumer forgiveness, the rebuilding of consumers' overall trust, and guests' greater satisfaction. Sixth, Airbnb should focus on training customer service personnel to upgrade their skills and abilities in relation to handling complaints.
This study possesses some limitations. The search for online narratives was restricted to one keyword, all reviews were written in English, and this research study exclusively analysed the content of comments posted on Trustpilot.com. Future studies should consider other websites, such as TripAdvisor.com and Booking.com. In addition, given the significance of offering memorable experiences in a tourism context to further influence consumers' behavioural intentions (Coudounaris & Sthapit, 2017) and subjective well-being (Sthapit & Jiménez-Barreto, 2018a), future studies should examine tourists' memorable Airbnb experiences and their impact on those tourists' revisit intentions and subjective well-being. More research is needed to further clarify the facets of distrust in the context of an Airbnb hospitality experience. Lastly, future studies should examine the influence of consumer distrust in the Airbnb context alongside its influence on repeat purchase intention and negative electronic word of mouth.