Complexities of trust building through sociomaterial arrangements of peer-to-peer platforms

ABSTRACT

Trading on peer-to-peer tourist accommodation platforms requires a sufficient level of trust between individual consumers and service providers. This is often achieved by using mutual consumer–provider evaluations, which are perceived as a trustworthy resource for information about upcoming stays. A variety of mechanisms and metrics are used to facilitate trust building on platforms; however, trust itself is being established by the platform’s users. This study investigates the case of Airbnb to show how arrangements of sociomaterial metrics and mechanisms are embedded in trust building. Findings from a virtual ethnographic study of the platform’s users show how trust building is performed through these arrangements. Based on organizational theories of trust and sociomateriality, the study suggests that establishing both attitudes of trust and distrust on peer-to-peer platforms are equally important. When the sociomaterial arrangement fails, trust may deteriorate outside of the platform organization’s control.

KEYWORDS:

• Trust
• distrust
• online evaluations
• sociomateriality
• Airbnb

Introduction

It is now common to organize peer-to-peer (P2P) exchanges of various assets in the tourism sector with the help of online platforms. Platforms offer a distinct form of organizing exchanges. Consumers and service providers are co-opted as platform’s members that trade resources, which belong to themselves rather than the platform organization (Stark & Pais, 2020). Exchanges themselves are mediated by an information and communication technology (ICT) platform (Perren & Kozinets, 2018). Platforms make organizing a large number of exchanges effective by using tools of algorithmic management, i.e. transforming ratings and activities of providers and consumers into quantifiable rankings and data that can be used to maintain control of the platform (Minca & Roelofsen, 2019; Stark & Pais, 2020). However, observing, evaluating and reviewing each other resembles surveillance, which implies a lack of trust in the individual users (Gössling et al., 2021). As platforms become a common way to organize trading of assets and services that range from accommodation to touristic activities, transportation, food or information (Gössling & Hall, 2019), few studies address how trust is built when such methods are employed.

It is more common to address trust building as a part of a business model that uses online evaluations and ratings for quality control (Baute-Díaz et al., 2019; Ert & Fleischer, 2019). Although reviewing of services itself is not a new phenomenon, the ways to build trust on P2P platforms differ from earlier applications of consumer reviews in marketing or quality control. Reviews and ratings are often provided mutually as both consumers and service providers evaluate each other. Reviews and ratings are then accumulated into a single record that forms a platform user’s online reputation, which becomes an excessively important resource that helps other users make decisions about the trustworthiness of service providers and consumers (Zloteanu et al., 2018). This helps reduce the complexity of trusting unfamiliar strangers, as being in close contact with unfamiliar others can create perceived risks for tourists and hosts (Lugosi, 2021). Tourism is also a specific context for trust building due to displacement – being in an unfamiliar setting and lacking local tacit knowledge (Williams & Baláž, 2021). This makes tourists reliant on their hosts, and trust becomes particularly important in their relationship. Vice versa, distrust or a lack of mutually beneficial service relationship can be detrimental to the tourist’s ability to manage uncertainty (Sthapit & Björk, 2019).

Trust is often discussed as desirable in order for P2P exchanges to happen, while distrust is considered undesirable and harmful to P2P platform use (Guttentag, 2015; Sthapit & Björk, 2019). However, earlier studies of trust and distrust hint that both phenomena can be viewed as functional equivalents in forming expectations about possible desirable or undesirable outcomes of a relationship (Lewicki et al., 1998; Luhmann, 1979). P2P platforms affect trust building by including sociomaterial mechanisms, in the form of digital devices, applications, networks, software and algorithms dedicated to building trust via reputation management systems. This can be perceived as relying on reputation-based trust, informed by previous experiences of other people and aggregated satisfaction scores. However, trust is interpersonal and built as interactions progress (Korsgaard, 2018), making it important to consider how a platform’s users practice trust building. This study addresses the question of how trust building is practiced on P2P platforms by applying the integrative model of organizational trust (Mayer et al., 1995) and analysing sociomaterial relations between the platform and its users.

The focus on sociomateriality highlights that trust building is embedded in both social and material platform’s elements, including devices, applications, algorithms, as well as rules and regulations for using them (Gössling et al., 2021). Studies of sociomateriality suggest that both the social and material aspects are integral to organizing, and the daily practices of everyday life are performed through them (Orlikowski, 2007). Accordingly, the social and material platform’s elements are integral to performing online evaluations. For example, Scott and Orlikowski (2012) suggest that online evaluations create a form of ‘quasi-formalised knowledge’ that is considered technical and unbiased, while in reality it is based on subjective opinions. The Mayer et al. (1995) model complements the analysis by bringing relational antecedents of trust, such as users’ ability, benevolence and integrity to the front as subconscious mediators between technology, tourists and service providers.

This study applies insights from the field of organizational studies to inform the empirical tourism context this way extending the discussion about trust in tourism (c.f. Ert & Fleischer, 2019; Williams & Baláž, 2021) with the notion of sociomateriality. The analysis is based on a virtual ethnographic study of Airbnb platform users (Hine, 2015), consisting of interviews with Airbnb users, go-along observations of their platform use (Jørgensen, 2016), and the platform’s content related to trust building. The approach applied in this study helps identify trust building as practiced through a platform’s sociomaterial arrangements and discuss how the use of these arrangements may lead to developing trust or distrust when applied in practice.

Complexities of trust building on P2P platforms

Trust is a necessary element of P2P exchanges, where both tourists and hosts do not have prior knowledge about each other (Guttentag, 2015). As a result, tourism studies about trust often investigate the effectiveness of systems for facilitating trust in relation to P2P business models. This can include looking into effects of reputation on purchase decisions (Chen & Chang, 2018), complaints relating to trust in a platform (Sthapit & Björk, 2019), consumers’ intentions to use or discontinue using a specific platform (Huang et al., 2020) or reviewing biases (Meijerink & Schoenmakers, 2020). Studies about Airbnb in particular report that review scores are often inflated, when compared to other accommodation rentals platforms (Bridges & Vásquez, 2018). Studies attributed this to various factors, such as herding behaviour, underreporting of negative experiences, self-selection and strategic manipulation of reviews (Meijerink & Schoenmakers, 2020; Zervas et al., 2021). In relation to such findings, Zloteanu et al. (2018) suggest that in P2P exchanges it is more accurate to describe trust as the perceived trustworthiness of other users based on their online reputation.

Although the literature discussed above provides valuable insights into the effectiveness of ways to facilitate trust, they offer a limited discussion of how specific sociomaterial mechanisms are used to facilitate trust. It is also commonly proposed that trust is a necessity for exchanges, while facilitating distrust among tourists and hosts should be avoided often limiting the discussion to facilitating trust and avoiding distrust (Sthapit & Björk, 2019; Tussyadiah & Pesonen, 2018). This gap is addressed by using organizational theories of trust to analyse interactions between Airbnb users and the ICT platform. The remaining section shows how theories of trust help address the complexity of these interactions.

Trust is a complex social phenomenon, as both parties in an exchange mutually influence the levels of trust in each other, also possibly developing distrust (Korsgaard, 2018). Trust is a psychological state involving a willingness to be vulnerable towards another person, based on positive expectations towards their motivations and actions (Mayer et al., 1995; Rousseau et al., 1998). Kramer (1999) proposes that trust is facilitated through either cognitive or affective means. Cognitive trust is primarily based on what is perceived as good reasons to trust others, rather than affective, emotional bonds (McAllister, 1995). This may include prior knowledge about the other individual, as well as an analysis of their motives and intentions. However, as Kramer (1999) observes, cognitive aspects are instrumental, and it is necessary to also consider affective or social elements that facilitate building of trust. This may include considerations of actors’ self-presentation, motivations or emotional bonds that actors’ build towards each other (Rousseau et al., 1998).

For a better understanding of both cognitive and affective aspects of trust, the analysis in this study is based on a model of trust proposed by Mayer et al. (1995). The model identifies actors’ ability, benevolence and integrity as the main factors that contribute to facilitating trust. Ability represents the skills and competences that allow the actors to have influence on the domain in which trust develops. Similarly, they can be defined as ability to provide core service (Sparks & Browning, 2011), such as hosting or being a good guest. Benevolence represents the actors’ altruistic motives, showing care for each other. Integrity shows the actors’ adherence to common norms and values (Mayer et al., 1995). Various combinations of these factors are often provided in user reviews on Airbnb, helping its users to form expectations about future encounters and facilitating trust or distrust.

In the service relation on P2P platforms, distrust also plays an important role and needs to be considered as a necessary outcome of the trust building process. Distrust involves pervasive negative expectations and signals intentions to protect oneself from the actions of another (Lewicki et al., 1998). This definition follows Luhmann’s (1979) notion that both trust and distrust function similarly by allowing individuals to manage uncertainty in social interactions. Trust achieves this by allowing specific desirable outcomes to be considered as certain, and distrust allows considering undesirable events as likely, thus allowing the individual to act and prevent such outcomes (Luhmann, 1979). Both phenomena are similar, as they involve management of expectations for reducing complexity by focusing on the likelihood of possible (un)desirable outcomes. Here trust is viewed not as a state or a characteristic of a relationship, but as a dynamic process in which both parties affect each other through their behaviour (Korsgaard, 2018). Relationships are complex, and individuals can simultaneously have both trusting and distrusting views of each other at different times and different contexts (Bies et al., 2018).

Trust building in sociomaterial assemblages

Trust building on P2P platforms is performed by producing evaluations of other users. It is now common to see evaluations of previous performance produced by a large number of anonymous and distributed consumers that use informal and individual criteria ‘grounded in personal opinions and experiences’ (Orlikowski & Scott, 2014, p. 864). Evaluations performed by consumers on platforms lack formal criteria, and instead refer to an arrangement of ‘infrastructure of networked computers, software code, databases, algorithms, and embodied consumer writing habits’ (Orlikowski & Scott, 2014, p. 888). Evaluations of an individual are accumulated to form their reputation – a collective measure of received evaluations on the platform, which indicates trustworthiness. The infrastructure used for collecting and accumulating them characterizes evaluations as sociomaterial.

On Airbnb, the sociomaterial arrangement is an assemblage of mechanisms that requires the guests and hosts to monitor each other’s actions and report them to the rest of platform users. The assemblage is characterized by intertwining mechanisms of communication, technology, platform’s services and policy (Gössling et al., 2021). Notably, this structure suggests how evaluations should be performed, according to norms implied as necessary for building trust among users (Roelofsen & Minca, 2018). Technological mechanisms, related services and policies mediate digital guest–host interactions and drive the user’s journey from reservation to evaluation (Gössling et al., 2021). The arrangements of the platform’s material elements create affordances to act in specific ways, e.g. by asking to provide specified types of information in user reviews. This shapes users’ practices of trust building, thus facilitating their outcomes.

Evaluations are performed as a coordinated action linked by common understandings of the platform’s rules and affordances. For example, rankings on Airbnb are produced in a quantifiable format that the platform requires (Roelofsen & Minca, 2018). Such affordances create a shared understanding of evaluating as a social practice and enact a particular structural order (Schatzki, 2002). The evaluators follow a shared understanding of what is expected from their evaluations, and how to produce them. From this perspective, trust building is seen as a set of practices entangled in social meanings of online evaluations and the platform’s sociomaterial arrangements. However, evaluations are still based on subjective opinions of the evaluators, who know that their opinions will be publicly visible (Scott & Orlikowski, 2012).

In a similar context of small tourism enterprises, Kelliher et al. (2018) suggest that trust can be built on the basis of positive interactions in mutual exchanges. This means that trust evolves as individuals reciprocate each other’s actions with cooperation and voluntary sharing of resources that benefits each other (Korsgaard, 2018). Over time, trust may become stronger; however, initially it is based on contextual information, such as role constraints, reputation or stereotypes (McKnight et al., 1998). However, in larger networks such as Airbnb, relations are often brief and rarely reoccur. On platforms, reviewing systems can replace mutual exchanges by preserving public testimonies about prior interactions. This has implication for how evaluations can be used for establishing contextual information necessary for building trust in the brief guest–host relations. Especially the reliability of used information, and a shared understanding of using it are important for facilitating trust and distrust.

Methodology

This paper draws on a virtual ethnography of tourists that use the Airbnb platform (c.f. Hine, 2015). Studying the case of Airbnb provides context-dependent knowledge, which allows further theorizing about trust building in similar contexts of users interacting with an ICT platform. The empirical design is based on go-along observations of online platform’s users, in this case either on Airbnb mobile application or website, in combination with semi-structured in-depth interviews. The data in this study were collected from 15 interviews paired with observations that lasted from 40 to 70 min, were audio-recorded and transcribed. The interviews explored themes related to trust building practices. Taking into account the performative nature of online evaluations, explored themes included common understandings of the platform, its different sociomaterial elements, engagement with them in previous experiences, and procedures followed when performing evaluations (Halkier & Jensen, 2011). Additionally, interviews showed participants’ perceptions of the platform’s functions, overall business model, and perceived roles of guests and hosts, their relation to trust building according to social roles.

The interviews were conducted in two stages. The first stage explored the themes related to trust building. For example, they include expectations towards stays, their sources, changing attitudes during stays. Afterwards, the interviews shifted to go-along observations of platform use. The platform has its verbal, and material components, and the go-along observations supply empirical material of how both the participant and the researcher are able to navigate those (Jørgensen, 2016). The participants were observed while using the platform’s search engine interface, structure of the hosts’ and listings’ profiles, Airbnb’s help centre, and other related elements. This showed how participants navigate the platform, allowing to compare their individual actions across the sample, and identify common understandings of the platform, and how its sociomaterial elements take part in the trust building.

The data collection was organized in two stages, along with consecutive data analyses. The initial five interviews were more open, aiming to gain initial familiarity with how trust building is practiced. These data were later used together with literature about trust, as well as sociomaterial elements of platforms to develop a more detailed interview guide. The initial participants were identified through snowballing from the researcher’s personal network, while later interview participants were identified through purposive sampling – selecting participants that fulfil specific criteria (Daniel, 2012). People that reside in Southern Sweden, have travelled and used Airbnb for arranging accommodation in the last six months were interviewed. Table 1 provides an overview of interview participants. Many of the participants were of international background, with university education, on average were around 30 years old, and have different amount of experience in using Airbnb. This corresponds to characteristics of Airbnb users identified in other studies (e.g. Guttentag et al., 2018; Lutz & Newlands, 2018). Therefore, although Airbnb’s user base is broad, participants still represent common users of this platform. The interviews were collected until data were saturated and meaningful comparisons between the different users could be made (Corbin & Strauss, 2012). Further data collection was also seen as unnecessary, due to ongoing COVID-19 pandemic, during which it became significantly more difficult to identify active recent platform users.

Table 1. Characteristics of interview participants.

Participant ID	Age	Gender	Nationality	Airbnb use experience
IP1	23	Male	UAE	Occasional user, for ∼1 year
IP2	38	Female	Denmark	Occasional user, for ∼4 years
IP3	26	Female	Ecuador	Frequent user, for ∼2 year
IP4	30	Female	USA	Frequent user, for ∼5 years
IP5	25	Female	USA	Frequent user, for ∼5 years
IP6	26	Female	USA	Frequent user, for ∼4 years
IP7	44	Male	Sweden	Frequent user, for ∼10 years
IP8	30	Female	Norway	Occasional user, for ∼1 year
IP9	31	Male	Canada	Frequent user, for ∼6 years
IP10	39	Male	Sweden	Occasional user, for ∼4 years
IP11	30	Female	Czech Republic	Occasional user, for ∼4 years
IP12	32	Male	Peru	Occasional user, for ∼1 year
IP13	29	Female	Mexico	Frequent user, for ∼6 years
IP14	26	Female	Mexico	Occasional user, for ∼2 years
IP15	23	Male	China	Occasional user, for ∼1 year

The analysis was carried out by following the main procedures of grounded theory (Matteucci & Gnoth, 2017). Interviews in all data collection stages were first coded with open coding, i.e. coding every line of the transcripts, to identify people’s understandings, possible objects of interests, tentative relationships between concepts that appear in the data. This stage allowed identifying the participant’s perception of platform’s functions, note down individual actions and compare them to identified sociomaterial practices, and eventually, identify the meanings of trust building and related platform’s elements before, during and after stays.

The coded interviews were later reviewed and coded again using axial coding – data were re-read while using theoretical constructs as categories and looking for connections with earlier codes (Matteucci & Gnoth, 2017). Klicka eller tryck här för att ange text. The trust building practices identified in the previous stage were compared with theoretical constructs, such as the role of ability, benevolence and integrity in trust. This helped identify categories that represent explicit acts important in performing evaluations and building trust, and the ones that represent more tacit guest–host interactions, e.g. subtle communicative acts that are hard to evaluate. The categories were further related to the platform’s affordances, e.g. considering the perceived use of available information, and making explicit their connections with trust building practices.

Analysis: trust building as the stay progresses

The Airbnb platform can be used for different purposes, including tourist accommodation rentals, and experiences such as tours and activities (Melián-González et al., 2019). The participants of this study use the Airbnb platform mainly for renting tourist accommodation. Therefore, the analysis focuses on trust building in different stages of tourist accommodation rentals: before, during and after stays. While before the stay, trust building is based on inferring the trustworthiness of hosts, during the stays trust building occurs in guest–host interactions, and after the trip guests focus on maintaining own positive reputation and evaluating. The analysis is structured around these three stages and related platform’s arrangements.

Perceiving hosts’ trustworthiness

During planning of a trip, guests need to develop a perception of trustworthiness towards their hosts. P2P platforms are a context where expectations for trustworthiness are informed by reputation (Williams & Baláž, 2021). A host’s reputation is created based on reviews and information on the listing and the host’s profile. The initial search for accommodation mostly focuses on finding this information and comparing it among different listings. The search is carried out on a search page structured by algorithms that determine how the search is structured on the platform’s interface (see Figure 1). This includes determining what information about hosts is discoverable and filtering information. One of the participants specifically notes importance of search filters at the initial information search:

And then you have to make sure that your filters are correct. And when you search through the listings, I usually look at the pictures first. Then price, and then you request to book sometimes, or you can book automatically. (IP6)

Figure 1. Airbnb accommodation search filters.

Ert and Fleischer’s (2019) notes that the structure of accommodation search in Airbnb is adapted to highlight what is perceived as objective information and obscure more subjective cues. This is the first information that guests see when searching and contributes to forming the initial basis for perceived trustworthiness. The information perceived as objective relates mostly to the hosts’ functional or interpersonal competences (Mayer et al., 1995), and is displayed according to the specified criteria used in search. Notably, although information at this point is structured according to the search algorithm, users rarely reflect on how the information is provided.

Search interface directs initial attention to showing aggregated rating scores, while more detailed qualitative information is found in user profiles. The go-along observations showed that while selecting a rental listing, not only the information about the listing, but also hosts’ profiles are evaluated. Sometimes this even includes available information about other people that have reviewed the place. This process is best explained by the following quote, noted when observing how a guest selects listings to research:

It would probably take around an hour to narrow it down. And then, when I have like two or three, I would go in depth for an hour. < … >

The guest then processes the information by reviewing both what is known about the host and previous reviewers:

So I'll read the reviews, all the reviews. And I’ll look at the host’s personal reviews < … > Also, I check if they have any other listings. And if they do I'll go and see what people are saying about the host specifically. < … > And then on top of that, I'll do a research into the area where it is. Sometimes, if I am very suspicious, but I need the place, I will look at the people who have reviewed, look at their profile. See what they’re saying about other people’s stuff. And if I find a pattern of them only talking negative, I’ll be – ‘ok, this person just highlights negative things, so maybe it’s not that bad’. (IP9)

The quote shows that an experienced Airbnb user can operate the system to form a detailed trust base, with information not only about the listing, but also the hosts, the area, profiles of previous reviewers. Being able to perceive the possibilities that the platform provides requires having prior experience with it (Dohn, 2009). Less experienced guests described a much shorter process of research, for example, relying on filters to find information for cognitive trust bases:

Typically, I think when I do the filters to search for like a Superhost and  …  Usually Superhosts have good reviews generally, so I don’t usually see when I am browsing that many places that have a lot of negative reviews. (IP5)

This is particularly important because the reviewing system and the information provided by hosts are not always trusted. As another guest explains:

The host can photoshop the pictures, take good angles from the pictures. Of course, in this place he will say wonderful things about it. But you will never know until you are really there. Also, probably best  …  You don’t know if the guy is crazy. At least with reviews you know people have been there, had seen him or her, he’s nice, normal guy. (IP3)

Information provided by hosts themselves is often approached with suspicion. This puts particular emphasis on the quality of reviews and other quality markings, for example the Superhost status. Airbnb explains that this certification marks ‘top-rated and most experienced hosts’ (Airbnb, 2022b). Guests often perceive Superhosts as ‘hosts that you can trust’ (IP3) or indicate that it ‘feels more safe’ and ‘comfortable’ (IP10) to stay in their accommodation. Search filters were also noted as means of simplifying search on the platform, when looking for specific amenities, type of places, prices or review scores. However, platform’s users rarely reflect on the way filters and quality markings structure their search. Perceptions of trustworthiness are thus formed in accordance with platform’s algorithms, reviewing system and quality markings.

Building trust in interactions

Earlier studies show that direct interactions help build trust (Baute-Díaz et al., 2019), however, even during stays, interactions happen both in interpersonal meetings and digitally. The platform is still used for its communication functions, other channels such as social media are also used, due to restrictions Airbnb places on users of its communication system. At this point, the guest and the host are two individuals with limited knowledge of each other, but who are interdependent and have clear role constraints. Blomqvist and Cook (2018) identify these characteristics as relevant to forming what they term ‘swift trust’, which is built over short term and is mainly based on cognitive cues about ability and integrity, rather than shown benevolence that tends to become more important over a longer period of time.

The ‘swift’ nature of trust may relate to guests’ preference for fully rented properties, thus avoiding direct contact with their hosts. As a result, longer-term trust relationships are rarely built. The study’s participants indicate that actually sharing space with the host ‘can get quite awkward’ and that ‘you’ve got to be more comfortable when you are on a trip’ (IP1). Most of them associate fully rented properties with more comfort, while the host mainly fulfils the role of a distant helper. This limits guest–host interactions to possible in-person greetings and communication via the platform, also limiting the affective indicators of trust. In such cases, trust building is based on cognitive cues, such as understanding of roles and a source of common values (Blomqvist & Cook, 2018).

Guests and hosts can refer to Airbnb’s community standards (Airbnb, 2022c) as a source of common values. It presents platform users with a set of norms to uphold safety and security, fairness, authenticity and reliability during stays (Gössling et al., 2021), which serves in setting expectations towards the roles of guests and hosts. However, when asked if they use such guidelines, most guests responded negatively. Instead, guests refer to previous reviews as a source of information about how to act in specific situations and communicate with their hosts. Although, guidelines are present, previous reviews and available information about the hosts are more important for setting guest and host roles. Clear information about the host’s role and identity is important:

Because you know the person since the beginning, and they have introduced themselves, and have taken the time to go and hand you the key, present their house and all the things they have – you have more of the sense that if something happens, something is missing, you don't know how to, even open the oven, or turn on the microwave, it is easier for you to communicate with these people, and send a message. Even though, it would be something that is annoying for him. (IP3)

The host’s identity and the distinct characteristics of their role as a host can be considered as a way to establish a framework of common values. However, such identification can be problematic as Airbnb hosts are free to set their own policies about how their rented place should be used and how hosting is carried out (Airbnb, 2022b). Notably, hosts’s practices of hosting can also contradict Airbnb’s policies or local laws. For example, they might not have a right to rent the property or use unauthorized people for performing the hosting services. This is against Airbnb community standards and policy, and was noted as one of the reasons to hinder trust building, for example:

I was talking, communicating with who I thought was the person I was gonna meet. And her husband showed up. And I never met her. And I had to drop the keys off at his parents’ house. And so, that one I was very suspicious about. Does this person who  …  is on the profile, on the photo exist? I didn’t understand. The place worked out, so I was ok with it. But I was like  …  that was pretty sketchy. (IP9)

Situations such as the one described above are common when renting on Airbnb. The platform allows co-hosting – hosting with someone that can help listing owners take care of their home. Co-hosts can be other authorized Airbnb members that are ‘someone the listing owner already knows’, such as a trusted friend or a family member or neighbour of the host (Airbnb, 2022d). Situations where this is noticed often foster distrust, as the integrity of hosts becomes questioned. Swift trust is fragile (Blomqvist & Cook, 2018), so setting standards and monitoring can often help maintain it. Yet, flexible platform policy does not ensure monitoring, and leaves a space for distrust to develop during the stay.

In other cases, the hosts’ policies might include the use of smart home devices, or cameras to ensure security on their premises. Airbnb does not directly prohibit the use of such devices; however, their use is regulated by its policies (Mare et al., 2020). The actual use of such devices can however result in infringement of privacy, which has also been noticed by one of the interviewed guests:

So what happened was that I found out that he had what I think was cameras in the flat, hidden. And … Which I noticed and which he removed. Anyway, I later reported both to Airbnb and the police. (IP8)

Early-stage trust based on role constraints and reputation diminishes in situations that deviate from the normal (McKnight et al., 1998). This poses a higher risk for tourists, who are in an unfamiliar context (Williams & Baláž, 2021). This can also lead to developing distrust towards the platform organization. Airbnb maintains a ‘Trust and Safety Division’ to deal with customer support and promises support in irregular situations. However, guests often noted not being able to receive help they expect or receiving it too late. For example, noting that the support services are ‘not always very helpful in the way that I want it to be, and they don't fully assume the risk for the consumer’ (IP5). More often it was stressed that ‘you have to look for a solution yourself’ (IP3), and the importance of reviews as a way to verify that their experience will be good. This suggests that Airbnb itself is not trusted to help, and as a guest’s experiences accumulate, more emphasis is placed on their own ability to correctly assume the possible risks from initially available information.

Maintaining a positive reputation

Reviewing on Airbnb is reciprocal, and reviews are a resource for information about guests as well. A guest’s reviews are collected in their profile on Airbnb platform. Mutual reviewing means that guests can also be sanctioned for misconduct and need to remain trustworthy. Also, when talking about reviews they have left in previous stays, the guests in this study reported that they also consider their own reputation when evaluating their hosts. For example, giving this explanation for not leaving negative feedback on the platform:

Because to be honest, I felt like I might get like a bad rap [reputation] in the Airbnb system. So I was a little bit worried about that. Because I know that sometimes the hosts will look at the reviews of the travellers. (IP9)

This reputational trade-off may allow for retaliation to bad reviews. Airbnb prevents that by keeping reviews hidden until both parties have written theirs (Ert & Fleischer, 2019). However, writing of reviews can have further implications for how trust building is practiced. Guests still indicate being doubtful about writing negative reviews, not only for avoiding reciprocal negative feedback, but also a possibility to be perceived as ‘negative’ by future hosts, who might check reviews they have left previously. Guests may also be doubtful about leaving negative feedback that can impact the business of another person, e.g.:

Because I feel uncomfortable leaving a negative feedback, criticizing someone. Also depending on what is the problem. If I feel that something was not the host’s fault, or not something intentional, a minor thing. I would feel better leaving that in private. (IP8)

This suggests that at this, more reflective, stage benevolence takes a more important role, whereas both parties take trust in the other’s intention to do good for them. This increases the complexity of writing reviews, as trust invariantly involves risk that such expectations will not be met (Mayer et al., 1995).

Airbnb platform employs a set of both social and material mechanisms to reduce this complexity. As an example, Airbnb’s community guidelines and Terms of Use specify various criteria for reviewing that both guests and hosts can refer to. They include cleanliness, accuracy, check-in, communication, location and value (Airbnb, 2022a). However, guests rarely refer to this information, instead using other platform’s resources to perform evaluations. For example, at this stage the earlier mentioned quality markings gained a new value – guests also referred to previous reviews, the scores left by previous guests, as an indication of how they should formulate their own feedback. For example:

I did read the reviews and the people who gave it two stars and three stars, they didn’t comment on what was wrong. < … > Two stars, three stars. Just like that, so we just left three stars. (IP1)

In this case, previously written reviews serve not only as ‘expert knowledge’ about upcoming service use, but also as knowledge about how to evaluate the stay in the same accommodation. This way guests are able to both reduce the complexity of reviewing and maintain a positive image of themselves (Orlikowski & Scott, 2014).

The reviewing system is designed to separate public reviews from private communication and feedback. These measures can be seen as an attempt by Airbnb to maintain a more objective reviewing system that is used to inform other users about actual encounters, while also leaving space for private feedback for matters that do not need to be disclosed publicly. Notably guests use this function when in doubt about the feedback they are leaving, disclosing more problematic issues privately. For example:

And if there’s something that I have specific, that I feel like would improve the person’s hosting experience, I would mention that privately. (IP5)

Leaving certain parts of feedback in private affords the guests to leave a positive review, while also giving the host useful, but more negative feedback in private. This also suggests that at this stage guests willingly provide helpful feedback to hosts without a motive for profit or an aim to incriminate the host’s business, showing a degree of benevolence. The participants’ testimonies suggest that at this stage, the material infrastructure that affords differentiating between the public and private reviews allows benevolence in evaluating. Guests find different ways to write reviews, such as highlighting the positive and negative aspects other people would want to know, providing recommendations, indicating whether the place felt safe, and the reviewing system allows highlighting such feedback in public, while leaving comments that can be seen as negative in private. This also allows performing evaluations to maintain good image of self, while maintaining an acceptable level of honesty in a public review.

Discussion

The complexity of trust building through sociomaterial arrangements is marked by prominence of both trust and distrust. This study supports the notion of Lewicki et al. (1998) that trust and distrust should be viewed as parts of a multidimensional relationship, where distrust can benefit Airbnb guests in building negative expectations about possible risks on a stay. As the results suggest, although Airbnb’s reviewing system is made to support trust building, forming negative expectations in the early stages of the stay fulfils this purpose. On the other hand, if trust deteriorates later on, the guest–host relationship suffers, and their willingness to cooperate when solving upcoming issues recedes (Korsgaard, 2018). Table 2 summarizes the findings of this study, highlighting how trust and distrust is built at different stages of stays and sociomaterial arrangements that facilitate such outcomes. Sthapit and Björk (2019) further explain that distrust formed during stays is a common barrier to the use of P2P accommodation platforms. Distrust at this stage mainly stems from poor customer service or host’s unpleasant behaviour towards guests.

Table 2. Trust building stages and related sociomaterial arrangements.

Stage and practice	Facilitating trust	Facilitating distrust	Sociomaterial arrangements in use
Before stay: Perceiving trustworthiness	Based on online reputation. Informs about trustworthiness, highlights ability and integrity.	Lacking ability or integrity shown in evaluations. Allows avoiding risky service relations.	Search page: structured with algorithms, search filters, structure of platform’s interface. ‘Superhost’ and other quality marks structure search for providers
During stay: Building trust in interactions	Swift trust – developed over a short period of time for a temporary relationship. Based on ability – skills and competence, integrity – a common value framework.	Violations of expected integrity. Distrust is detrimental to service relationship.	Policies, e.g. Terms of Service, establish how a service relationship should look. Information in profiles and reviews form expectations.
After stay: Maintaining a positive reputation	Guest and host provide reciprocal public and private feedback. Benevolence shows in mutual reviewing.	Earlier distrust prevents writing reviews. Guests have a propensity to trust rather than distrust.	Distinction between public and private feedback in platform’s design allows benevolence in reviewing.

Distrust was observed as originating from guests noticing a lack of trust indicators, especially the host’s integrity. Meanwhile, experienced guests were prone to look for cues that indicate reasons to distrust a host, especially cues for the lack of ability or integrity, in information available on the platform – reviews, hosts’ profiles and listings.

This suggests that trust building starts before the stay and is based on determining trustworthiness of the other party by considering their online reputation. Reputation is an important indicator of trustworthiness in an early stage of a relationship (McKnight et al., 1998), and is established based on the infrastructure for performing evaluations that platform maintains. Trust building before the stay happens in a sociomaterial assemblage of reviews and ratings, the search system which presents this information according to criteria that are considered ‘objective’, and users that provide and read reviews. Reputation helps form expectations for trust and reflects professional competence or ability (McKnight et al., 1998). However, since the platform presents reputation in a functional manner – highlighting ‘objective’ information, such as listing details, price and aggregated ratings, quality markings, it can be less informative about other aspects, such as benevolence and integrity.

Actual trust is interpersonal and built as the stay happens and guests interact with hosts. At this point, trust building is based on role constraints and interdependency, maintaining ability and integrity as the most important trust indicators. This fits well with the concept of swift trust, as trust developed over a short period of time for a temporary relationship (Blomqvist & Cook, 2018). Ability shows the host’s competence to carry out what the role of a host demands, and integrity testifies to having a common framework of values (Kramer, 1999; Mayer et al., 1995). Violations of expected integrity have shown to result in distrust towards hosts, which is detrimental to the service relationship at this stage. This can relate to guests being placed in an unfamiliar context, where they experience increased risk (Williams & Baláž, 2021), thus being dependent on hosts to preserve safety and security.

Finally, as the stay comes to an end, trust building shifts to maintaining reputation. As public feedback is reciprocal, this concerns both guests and hosts. Earlier trust building practices relied on ability and integrity as indicators that facilitate trust, but at this stage benevolence was also shown in reviewing. Even in cases, where negative evaluations could have been performed, guests have shown an inclination to leave negative aspects in private, out of concern for the host or themselves maintaining a more positive look. However, this has not been the case if the host’s integrity has been compromised prior to reviewing. This suggests that guests have a propensity to trust that even negative aspects of stays are not intended. Kramer (1999) suggests that such propensity to trust is an important antecedent to overall trust that indicates a state opposite to suspicion that the other party will act to harm the guest or host. Thus, as stays come to an end, if trust has not deteriorated before, guests show more willingness to trust and express that with positive reviews.

Notably, Airbnb’s reviewing system itself creates affordances for maintaining more positive overall results by separating feedback into private and public. However, there are no indication about how experiences noted in private feedback could be expressed publicly. This helps maintain higher overall platform’s ratings (Ert & Fleischer, 2019), and create a perception that Airbnb hosts offer hospitality of higher standard (Zervas et al., 2021). However, this makes it harder for guests and hosts to build reasonable trust expectations before the stay begins, which can lead to increased risks during stays.

Conclusions

This study presents a view of trust building on P2P platforms based on findings from virtual ethnography with Airbnb users. Guests practice trust building in three distinct stages – determining trustworthiness while planning their stay, developing trust in interactions with hosts during stays and maintaining own reputation while evaluating stays. The study shows sociomaterial arrangements for mutual evaluations are embedded in this process, facilitating both trust and distrust as an outcome. Understanding trust building as performed in a sociomaterial arrangement helps understand the role that a platform takes in trust building. This includes both the platforms technological and social mechanisms. While previous studies point to a positivity bias in online evaluations (Bridges & Vásquez, 2018; Meijerink & Schoenmakers, 2020), these findings suggest that evaluations result from building trust in a sociomaterial arrangement.

This study shows that tourism organized with platforms is embedded with sociomaterial arrangements such as the Airbnb’s reviewing system. The analysis shows that building trust with sociomaterial arrangements facilitate cognitive and affective antecedents of trust, however, as noted by Roelofsen and Minca (2018) equating ‘trustworthiness’ with an evaluation of past behaviour may overlook other ways of generating trust. This could be addressed in further studies into building trust on P2P platforms.

From a practical perspective, this study provides an insight into the functionality of online evaluation arrangements on P2P platforms. Perhaps the most important implication is the necessity to treat trust and distrust as functional equivalents, rather than opposites. Reviewing arrangements need to be designed in a way to afford building of both trust and distrust in trip planning, for tourists to establish realistic expectations and prevent uncertainty and risk noted by Williams and Baláž (2021). During the later stages of stays, the level of trust is more dependent on the interactions between guests and hosts and their purposeful use of the reviewing system. This can be hard to manage with the same mechanisms and leaves a possibility for trust to deteriorate. Guidelines for performing evaluations can be a particularly effective method for managing trust building, if they reflected actual sources of deteriorating trust and addressed possible related situations.

Ethical declaration

This study does not require an ethical approval. All participants were informed about the study’s details and data use via an informed consent form. All information about the participants in this study is anonymized.

Disclosure statement

No potential conflict of interest was reported by the author(s).