The concept of home stay originated in the 1960s in the United Kingdom and saw widespread development in Taiwan and Japan during the 1980s. Internationally, “home stay” often refers to accommodations known as Bed and Breakfast (B&B) facilities, and terms such as Family Hotel, House Hotel, and Guesthouse likewise express the meaning of B&B (Zhang and Meng, 2017). In China, the term B&B is derived from the Japanese word “minshuku”, and the early development of the home stay industry began in Taiwan. Rural B&Bs, as described in this study, are small-scale accommodation facilities that integrate local natural amenities, cultural resources, and lifestyles, so they provide not just lodging but also immersive cultural experiences.

Scholars abroad have studied B&Bs since the 1980s, examining host-guest relationships, value perception, and cultural immersion (Stringer, 1981; Hultman and Cederholm, 2010; Gunasekaran and Anandkumar, 2012). In recent years, studies have extended into rural sustainability and emotional connection. For example, recent research has highlighted emotional support and destination attachment as core mechanisms driving tourist engagement in rural B&Bs (Li et al., 2024).

In China, research on the B&B industry has evolved from preliminary explorations to in-depth development. Early studies mainly focused on qualitative analyses involving basic concepts of B&B, market positioning, and development conditions. These studies were generally broad and lacked specificity and operability. However, research on B&Bs has deepened over time, especially since 2006, with more diverse and practical subjects and methods being employed. For example, Deng et al. (2014) analyzed B&B development in southwestern ethnic minority areas and proposed the extension of industry chains. Meanwhile, Li and Cheng (2017) used big data to analyze tourist reviews and behavioral tendencies. Based on Maslow's hierarchy of needs theory and using the IPA analysis method, Luo et al. (2014) summarized the importance and satisfaction levels of tourists regarding B&B service quality. However, there is limited research integrating modern approaches like fuzzy- set qualitative comparative analysis (fsQCA) to uncover the interactions of various factors influencing customer behavioral intentions in rural B&Bs. Recent research, such as that by Li and Phakdeephirot (2023) using TPB to study experiential consumption in rural homestays, marks an important methodological shift, but still falls short of examining combinatorial influences using fsQCA and NCA techniques.

Thus, although both domestic and international research on rural B&Bs has yielded rich results, in-depth analyses on the factors influencing the behavioral intentions of rural B&B customers and their combination pathways are still lacking. Particularly in the context of big data and the internet, obtaining real customer feedback through online reviews can provide a new perspective and method for understanding customer behavioral intentions.

Behavioral intention refers to an individual's propensity to engage in future actions. This concept is rooted in marketing theories, particularly the Theory of Planned Behavior (TPB), which posits that behavioral intentions are influenced by attitudes, subjective norms, and perceived behavioral control. Scholars, both domestically and internationally, largely concur in this definition. Most believe that behavioral intentions suggest the likelihood of an individual partaking in specific activities or behaviors. For instance, Feng and Cheng (2009) suggested that behavioral intentions are predictive of future behaviors, indicating the probability that an individual will undertake certain actions. Alternatively, some scholars argue that behavioral intentions represent a customer's predisposition to engage in specific actions towards a product or service post-consumption.

In the tourism context, behavioral intentions are often predictive of actual future behaviors, such as revisiting, recommending, or even paying a premium for a service (Yuksel et al., 2010; Hyun et al., 2011; Chen et al., 2020). Shu (2018) noted that tourist behavioral intentions are their tendencies to revisit and recommend a destination after their travel experience. In studies on customer environmental behavior intentions in ecological accommodations, Liu et al. (2009) defined behavioral intentions as the tourists’ willingness to repurchase ecological travel products, recommend them, or pay a premium. Lee et al. (2007) described behavioral intentions as a tourist's foresight into the likelihood of undertaking certain specific actions in the future. Zhang (2008) defined it as the attitude tourists develop toward a destination or host location after their experience. Hyun et al. (2011) further viewed behavioral intentions as the probability that tourists, after traveling, predict they might revisit the site, recommend it to others, or engage in positive publicity.

Recent studies have shown that behavioral intentions in rural tourism contexts are shaped by a combination of cognitive, emotional, and cultural factors. Research by Rasoolimanesh et al. (2022) identified memorable tourism experiences (MTE) as the key determinants of behavioral intentions and emphasized that satisfaction and destination attachment mediate these effects. Moreover, a study by Satar et al. (2024) demonstrated how customer engagement and co-creation influence behavioral intentions in tourism, which extended TPB by integrating emotional and experiential dimensions.

Despite the extensive use of TPB in behavioral intentions research, a gap remains in integrating modern techniques such as fsQCA to identify the multiple configurations of factors that lead to strong behavioral intentions. Current studies often overlook the complexity of behavioral intentions by reducing them to linear paths. However, fsQCA allows for a more nuanced understanding of the complex, heterogeneous combinations of factors that shape these intentions. This study addresses this gap by applying fsQCA and NCA to explore the necessary and sufficient conditions for predicting customer intentions in the rural B&B sector.

This study used Python to develop web-crawling scripts and collected travelogues that mention “rural B&B” from four representative platforms: Ctrip, Mafengwo, Qunar, and Xiaohongshu. As of June 15, 2024, a total of 268 travelogues were collected, comprising 100 from Ctrip, 50 from Mafengwo, 9 from Qunar, and 109 from Xiaohongshu, with a combined word count of 50943 words. During the data cleaning process, incomplete, irrelevant, duplicated, corrupted, and meaningless statements were removed, and statements relevant to the customer behavioral intentions in rural B&Bs were extracted. Synonyms were standardized to ensure consistency. Subsequently, the ROST CM 6 software was employed to conduct a frequency analysis of the text, which initially identified the customer behavioral intentions in rural B&Bs. Then, three experienced graduate and doctoral students specializing in rural B&B research were invited to categorize the identified high frequency terms. In addition, external professors and experts were enlisted to review the categorization results, thereby ensuring the surface and content validity of the content analysis.

High frequency terms refer to terms that appear frequently in user reviews and are deemed relevant (excluding, for example, punctuation marks). These terms significantly reflect the focal points of user interest. The ROST CM 6 software was used to extract the top 50 characteristic terms by frequency. These terms were then ordered from highest to lowest frequency, ultimately forming a high-frequency vocabulary table, as shown in Table 1.

Based on the analysis of high frequency terms, this study employed content analysis to classify them by grouping those that relate to the same theme into corresponding categories. Considering the need to control the number of categories, ensure mutual exclusivity among them, and fulfill the requirements for explanatory comprehensiveness, nine analytical categories were ultimately established: natural environment, material space, cultural space, quality of service, experience of home, entertainment education, culinary delights, surrounding environment, and value-added services. The initial concepts for each category were developed based on the original text of the customer travelogues, as shown in Table 2.

Fuzzy-set Qualitative Comparative Analysis (fsQCA) was developed in the early 2000s by Ragin (2008) to address the limitations encountered when analyzing social phenomena using traditional qualitative and quantitative methods. As a hybrid method that incorporates both quantitative and qualitative approaches, fsQCA is based on case-oriented research. In addition, this method has the advantages of being applicable to small samples, handling fuzzy data flexibly, identifying multiple causal paths, distinguishing between necessary and sufficient conditions, having strong interpretive results, and supporting theory testing and construction, which can help researchers to comprehensively and deeply explore and explain complex social phenomena.

Necessary Condition Analysis (NCA) is a method that specializes in analyzing the necessity relationships between variables, and it is used to quantify how the antecedents constitute the necessity of the outcome, which complements the qualitative focus of necessity analysis in the fsQCA method.

Based on a content analysis method, this study constructed a model comprising nine factors that influence the behavioral intentions of rural B&B customers. These factors include the natural environment, material space, cultural space, quality of service, experience of home, entertainment education, culinary delights, surrounding environment, and value-added services.

The survey questionnaire for this study was finalized by combining the specific context of rural B&Bs and drawing on relevant established scales, as well as similar research by the authors, and consultation with experts. The questionnaire was divided into two parts. The first part assessed the basic characteristics of the respondents, including gender, age, educational background, and income; while the second part addressed specific items across nine dimensions, namely natural environment, material space, cultural space, quality of service, experience of home, entertainment education, culinary delights, surrounding environment, and value-added services (Table 3). The questionnaire employed a 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree), with respondents rating the items based on their actual experiences.

One village from each of six districts listed in the “2022 Beijing National Key Rural Tourism Villages” was selected for this study. Before the formal survey commenced, a preliminary online questionnaire was distributed from June 15 to June 20, 2024, with favorable outcomes that required no further modifications. The formal survey was conducted from July 1 to July 8, 2024, using convenience sampling with the assistance of community workers to distribute questionnaires to rural B&B consumers. During this period, 180 questionnaires were distributed, and after discarding those with excessive duplications or incomplete data, 150 valid questionnaires were retained, which met the sample size requirements for fuzzy-set qualitative comparative analysis.

The demographic characteristics of the valid survey samples were as follows: 72 males and 78 females, constituting 48% and 52% of the sample, respectively. A majority had received higher education, with 28% holding junior college degrees and 37.3% holding bachelor's degrees. The largest age group was 26-35 years, accounting for 31.3% of the sample, followed by the 36-45 years group at 25.3%. The most common monthly income bracket was 10000 to 20000 yuan, representing 28% of the samples, followed by 5000 to 10000 yuan at 22%.

The questionnaire included 33 items in nine dimensions. The overall Cronbach's α coefficient was 0.953, significantly exceeding the accepted threshold of 0.7 and indicating high internal consistency of the scale. The Cronbach's α coefficients for each dimension were all above 0.8, also well above 0.7, confirming high reliability within each latent variable.

According to the results of the model fit test presented in Table 4, the CMIN/DF was 1.389, which was within the acceptable range of 1-3. The RMSEA was 0.051, approaching the excellent range. In addition, the ITI, TLI, and CFI indices all achieved outstanding levels above 0.9. Thus, the combined results from this analysis indicate that the CFA model demonstrates good fit.

Given the satisfactory model fit of the Confirmatory Factor Analysis (CFA) for the scale, further verification of each dimension's convergent validity (Average Variance Extracted, AVE) and composite reliability (CR) was conducted. The verification process involved calculating standardized factor loadings for each measurement item within its corresponding dimension through the established CFA model. Subsequently, the AVE and CR values for each dimension were calculated using their respective formulas. According to the analysis results in Table 5, this validity check of the scale yielded CR values for all variables above 0.7, AVE values above 0.5, and Variance Inflation Factor (VIF) values below 5, indicating that the scale does not suffer from multicollinearity issues. These data demonstrate that each dimension of the scale exhibits high reliability.

The items in this questionnaire were adapted from previous research by other scholars and modified according to the needs of this study, thus ensuring high content validity for the scale. According to the analysis results presented in Table 6, the discriminant validity test yielded standardized correlation coefficients between each pair of dimensions that were less than the square root of the AVE values corresponding to those dimensions. This indicates that each dimension demonstrates good discriminant validity.

In this study, variable calibration within the (0-1) range was conducted using Fiss's triadic calibration method in fsQCA 4.1, based on a 5-point Likert scale for both conditions and outcome variables (Getz and Petersen, 2005). Specifically, the 95th percentile was set as the point of full membership (Fully In), the mean as the crossover point (Crossover Point), and the 5th percentile as the point of non-membership (Fully Out) (Fiss, 2011). The data were then calibrated accordingly.

Generally speaking, when the consistency of a variable is not less than 0.9, the variable can be regarded as a necessary condition for the outcome variable. Table 7 shows that the consistency of each individual antecedent condition affecting positive or negative customer behavioral intention does not exceed 0.9, and therefore none of them constitute a necessary condition, which confirms the previous theoretical analysis that no single variable can explain customer behavioral intention in rural B&Bs. In addition, the coverage of each antecedent variable is relatively high, with most exceeding 0.5, indicating a better representation of the antecedent conditions.

Figure 1 shows the scatterplots of the necessity analysis. Each scatterplot determines the necessity of an antecedent variable by the space above the two ceiling lines (capping lines). The larger the space, the more likely the antecedent is to limit the outcome variable and the more likely it is to be necessary for the outcome. NCA generates the ceilings lines in two ways: ceilings-envelope-freedom-of-disposition (CE-FDH) and ceilings-regression-freedom-of-disposition (CR-FDH).

The results in Figure 1 show that for each of the antecedent variables, there is a very limited amount of blank area to the upper left of the upper limit line, which suggests that none of the antecedent variables is necessary for the customer behavioral intentions in rural B&Bs. The data in Table 8 shows that the effect values of all variables are less than 0.1. According to the NCA test rule, only when the effect value of the antecedent variable is greater than 0.1 while the significance is less than 0.05 can it be regarded as necessary for the outcome variable. In conclusion, the analysis did not find the necessity of any of the individual antecedent variables for the behavioral intention of rural B&B customers. This absence of necessary conditions highlights the importance of testing the sufficient conditions, i.e., examining the combined effects of different antecedent variables on customer behavioral intention in rural B&Bs.

Through counterfactual analysis, fsQCA 4.1 generates three types of solutions: Parsimonious, intermediate, and complex. Since the complex solution involves more logical remainders, this type of analysis typically employs a strategy where the intermediate solution is primary and the parsimonious solution is supplementary (Andrews et al., 2015). According to the analysis, a total of nine pathways were identified (Table 9).

The analysis results revealed multiple pathways to positive customer behavioral intentions in rural B&Bs. The overall consistency of these pathways reached 0.950, surpassing the consistency threshold of 0.7 and indicating that the explanatory model possesses considerable persuasive power (Cheng et al., 2019). This implies that in cases where customers in rural B&Bs exhibit positive behavioral intentions under the nine identified configurations, 95% of the intentions are positively manifested, demonstrating high consistency in the model. The nine combinations have original coverage rates ranging from 13.3% to 43.9%, indicating that any single configuration can explain approximately 13.3% to 43.9% of the cases. The unique coverage rates range from 0.004 to 0.216, suggesting that no single combination can fully explain all cases, and customer behavior intentions in rural B&Bs are influenced by various combinations (Gu, 2023).

Corresponding configurations were constructed according to the core conditions of each pathway. Configurations 1, 2, and 3, whose core conditions are Natural environment and Surrounding environment, are termed Environmental Immersion. Configuration 4, which includes Natural environment, Cultural space, Entertainment education, and Surrounding environment, is named Educative Entertainment. Configuration 5, which focuses on Cultural space, Quality of service, and Value-added services, is called Service Experience. Configurations 6 and 7, encompassing Service of quality, Experience of home, Entertainment education, Culinary delights, and Surrounding environment, are labeled Comprehensive Perception. Configurations 8 and 9 did not display core conditions, so they were not named or analyzed further.

In the Environmental Immersion configuration, Configuration 1 has a consistency of 0.968, indicating a 96.8% likelihood of positive customer behavioral intentions in cases with this configuration, with an original coverage rate of 20.7%. Configuration 2 has a consistency of 0.958, explaining 14.5% of positive intentions, and Configuration 6 has a consistency of 0.956, accounting for 15.9% of positive intentions. Customers in this category place high value on their natural and surrounding environments, seek deep engagement with nature and desire relaxation in tranquil settings (Liu and Pratt, 2017). They also tend to engage in various outdoor activities such as hiking, biking, and stargazing, so they seek an escape from urban life to enjoy quiet and beautiful scenery, with natural settings helping them to relax and reduce stress. These customers also often have interests in eco-tourism and sustainable development, so they wish to experience the beauty and harmony of nature through their travels (Liu, 2024).

In the Educative Entertainment configuration, Configuration 4 has a consistency of 0.959, explaining 13.3% of positive intentions. Customers of this type are keen on understanding and experiencing local culture and traditions. They particularly enjoy participating in cultural lectures, where they can gain in-depth knowledge about the history, customs, and values of the local community. Craft experiences provide them with hands-on opportunities to engage with traditional artisanal skills, thereby fostering a deeper appreciation for local craftsmanship. Historical site visits allow them to connect with the past, to visualize and understand the historical context that has shaped the region. These customers appreciate the harmonious blend of learning and entertainment in natural settings, as noted by Xiao and Ming (2003). The natural environment serves as a backdrop that enhances their overall experience, making the learning process more engaging and memorable. They are typically cultural enthusiasts with deep-seated interests in local culture, history, and traditions. Driven by a passion for knowledge and cultural exploration, they are willing to invest both time and money to fully immerse themselves in these experiences and expand their understanding, as supported by the findings of Lee et al. (2024). This suggests that rural B&Bs aiming to attract this customer segment should prioritize offering rich cultural and educational activities and integrate them seamlessly with the natural beauty of the surroundings.

In the Service Experience configuration, Configuration 5 has a consistency of 0.966, explaining 14.4% of positive intentions. Customers within this category place a premium on service quality and view it as a key determinant of their overall experience. They seek not only comfortable accommodation but also attentive, personalized, and professional service throughout their stay. Simultaneously, they have a strong desire to be enveloped in a local cultural atmosphere, which adds depth and uniqueness to their experience. As Schlesinger et al. (2020) pointed out, the combination of high-quality service and a rich cultural context can significantly enhance customer satisfaction. The value-added services provided by rural B&Bs play a pivotal role in meeting the personalized needs of these customers. These services could range from customized tour arrangements that showcase the best of the local area to in-house dining experiences featuring traditional local cuisine prepared with fresh, local ingredients. By tailoring services to individual preferences, rural B&Bs can create a sense of exclusivity and care that makes customers feel valued and special. This, in turn, enhances their satisfaction levels, as supported by the research of Eluwole et al. (2024). For rural B&Bs targeting this customer group, continuous improvement of service quality and the development of unique, culturally-rich, value-added services are essential strategies for fostering positive customer experiences and building long-term customer loyalty.

In the Comprehensive Perception configuration, Configuration 6 has a consistency of 0.951, explaining 19.1% of positive intentions, while Configuration 7 has a consistency of 0.972, accounting for 43.9% of positive intentions. Customers in this category demand high standards across service quality, home experience, entertainment and education, culinary delights, and the surrounding environment (Kumar et al., 2023). They also expect the surrounding environment of the B&B to provide a pleasant vacation atmosphere (Pappas and Glyptou, 2021b). Typically, they are family travelers who expect every family member to enjoy the trip, so they seek a comprehensive travel experience that allows them to enjoy a variety of experiences from dining to entertainment to learning in one trip. They are detail-oriented, and ensure that every aspect of the trip meets high standards to satisfy each family member's needs (Assiouras et al., 2019).

To ensure the robustness and reliability of the research findings, this study implemented a stringent verification procedure. Following recommendations in the existing literature, an analysis was conducted to adjust the consistency thresholds. Specifically, the initial consistency threshold was raised from 0.80 to a more stringent 0.90, and then the research data were reprocessed. This adjustment was made to test the sensitivity of the research findings to changes in the threshold. A comparative analysis showed that even under the higher consistency requirements, the causal configurations identified were essentially consistent with the original analysis. This high degree of consistency not only validates the rigor of the analytical process employed in this study but also significantly supports the reliability and stability of the conclusions.

This study contributes to our theoretical understanding of customer behavioral intentions in rural Bed & Breakfasts (B&Bs) by utilizing fsQCA (fuzzy-set Qualitative Comparative Analysis) and NCA (Necessary Condition Analysis). It reveals the impact pathways of multiple factor combinations on customer behavior, enriching the theoretical framework of behavioral intentions. Traditional customer behavior studies often employ quantitative methods and focus on the impacts of individual factors, while neglecting the complexity of multiple factor interactions. For example, Chen and Rahman (2018) demonstrated the significant role of service quality in customer behavioral intentions, but did not explore the interactions between service quality and other factors such as the natural environment or cultural space. In contrast, Pappas and Glyptou (2021a) examined how fsQCA and NCA could reveal risk factors that influence tourism purchase intentions in recessionary times, validating the effectiveness of these methods and highlighting the potential of multi-factor models in decision-making processes. This study adopted similar methods, further extending the application of multi-factor interactions.

The Theory of Planned Behavior (TPB) has long been used to examine customer behavioral intentions, and it emphasizes the roles of attitudes, subjective norms, and perceived behavioral control in predicting behavior. However, TPB traditionally overlooks the interactions between multiple factors, especially the interplay of environmental, cultural, and service-related factors. Our study bridges this gap by using fsQCA and NCA to explore how combinations of various factors interact to shape customer behavioral intentions. This was extended to the TPB framework by incorporating these multi-dimensional influences, thereby offering a more dynamic and comprehensive understanding of customer behavior in tourism contexts. Specifically, fsQCA reveals how different factor combinations influence customer behavior, providing a more nuanced perspective, while NCA quantifies the necessity of each antecedent condition, thereby enhancing the rigor of the study.

In tourism research, Restorative Theory (Kaplan and Kaplan, 1989) posits that exposure to natural environments facilitates psychological restoration, increasing customer satisfaction and loyalty. Our results validate this theory and further elucidate the relationship between the natural environment and customer behavioral intentions, especially for Environmental Immersion customers who prioritize natural surroundings and landscapes. This finding aligns with Environmental Psychology and Cognitive Appraisal Theory (Lazarus, 1991), which emphasizes the psychological benefits of interacting with nature.

Our study also reinforces the significance of factors such as cultural space, service quality, and value-added services in influencing customer behavioral intentions. This is consistent with Expectancy-Disconfirmation Theory (Oliver, 1999), which suggests that customer behavioral intentions are closely related to whether their expectations are met or exceeded. The results of this study show that Service Experience customers place more emphasis on service quality and value-added services, whereas Environmental Immersion customers focus more on the natural environment and cultural space. Abbas et al. (2023) further confirmed the relationship between Market Mavenism, co-creation experiences, loyalty, and vacation intention, which indicated how value-added services influence customer loyalty and future purchase intentions, and this aligns with our findings regarding the importance of service quality and value-added services for certain customer types.

First, operators should focus on the protection and enhancement of the natural environment. For environmental immersion customers, the natural environment and surrounding landscapes are key factors in their choice of rural B&Bs. Operators should adopt sustainable development measures to prevent environmental pollution and ecological damage, enhance the natural beauty through landscaping, and provide outdoor activities like hiking, biking, and stargazing to meet their customers’ needs for nature engagement (Choi and Sirakaya, 2006).

Second, enriching cultural experiences and educational entertainment programs is a key strategy for attracting educative entertainment customers. Operators can organize regular activities such as traditional handicraft making, cultural lectures, and historical site tours to enhance the cultural atmosphere (Park and Santos, 2017). Designing family-oriented educational activities like farming experiences and ecological education can meet the needs of family customers.

In terms of service quality and value-added services, operators need to strengthen staff training to enhance service awareness and professional skills, in order to ensure high service standards (Kandampully et al., 2015). Providing personalized services, such as customized breakfasts and exclusive tours based on customer needs, can significantly increase customer satisfaction.

Comprehensive perception customers have high demands across quality of service, experience of home, entertainment education, culinary delights, and the surrounding environment. Therefore, operators should strive to create a homey atmosphere, provide family rooms, parental facilities, and activities, and create a homey and comfortable feel. Enriching entertainment options such as adding children's playgrounds, family cinemas, and other facilities can meet the family members’ entertainment needs. Enhancing food services that offer a variety of culinary options, especially local specialties, can satisfy customer taste preferences. Also, optimizing the surrounding environment to ensure convenient transportation around the rural B&B and providing detailed travel guides and recommendations can enhance the overall travel experience (Kim, 2014).

This study has several limitations, and future research directions can be envisioned. First, the sample was regionally limited, so expanding it nationally would improve the representativeness of results. Second, while this study explored the factors affecting customer behavior, it did not deeply analyze the specific needs of different customer types. Future research could use market segmentation to better understand these needs and offer targeted advice for rural B&B operators. Lastly, although the findings support rural B&B operations, future studies should test these conclusions in real-world settings to refine the services and management strategies.