Previous study has indicated that information seeking encompasses various aspects, including product-related information, pre-purchase inquiries, and the search for product reviews ( Whiting & Williams, 2013 ). Notably, on Facebook brand pages, each post containing detailed brand or product information has the potential to stimulate consumer engagement with the brand on SNSs. Furthermore, information seeking behavior is notably prominent on Facebook and is positively correlated with customer engagement ( Metallo et al., 2020 ). Consider a scenario where a consumer is browsing through a brand’s Facebook page in search of information about a new product release. They come across a detailed post containing specifications, features, and user reviews. Intrigued by the comprehensive information provided, the consumer engages with the post by liking it, leaving a comment to inquire further, and sharing it with their friends who might also be interested. In this example, the consumer’s information-seeking behavior

on the SNS directly influences their engagement with the brand’s content, demonstrating the significant impact of information seeking attributes on consumer engagement behavior in Cambodia.

Hypothesis 1 (H1): Information seeking has a positive influence on consumer engagement behavior.

Research indicates that SNSs function primarily as entertainment platforms, with engaging content significantly motivating user interaction ( Pletikosa Cvijikj & Michahelles, 2013 ). In terms of entertainment, four main activities have been identified: relaxation or escapism, inspiration and mood management, enjoyment, and time-passing ( Muntinga et al., 2011 ). Users typically engage in these activities by watching funny videos, seeking inspirational content, enjoying local music and dance, and passing time during commutes or breaks. Additionally, empirical studies have shown a positive correlation between entertainment and customer engagement, identifying entertainment as a key factor influencing individuals to follow brand pages on SNSs ( André, 2015 ). Therefore, the importance of entertainment in enhancing user engagement on SNSs is evident in Cambodia, mirroring global trends noted in the cited research.

Hypothesis 2 (H2): The positive effects of entertainment on consumer engagement behavior are significant.

Previous research findings indicate a positive correlation between economic benefits and customer engagement ( André, 2015 ). As noted by Kujur and Singh (2018) , users often anticipate some form of compensation in exchange for engaging in activities such as liking, reacting, commenting, and sharing brand posts on Facebook brand pages. This underscores the significance of economic benefits as a primary motivator for brand engagement behavior among users ( Jayasingh, 2019 ). For instance, a user on Facebook sees a sponsored post from an online retailer offering a limited-time discount code. Intrigued by the savings opportunity, they click the link to explore further. After finding a product they like, they make a purchase using the discount code. They engage with the post by leaving a comment expressing satisfaction and sharing it with friends to inform them of the offer. In this scenario of Cambodia, the discount code’s economic benefit drives user engagement, potentially boosting sales for the retailer.

Hypothesis 3 (H3): Economic benefits have a positive impact on consumer engagement behavior.

In social media research, numerous studies have consistently indicated that one of the primary reasons for individuals refraining from engaging with Facebook brand pages is their apprehension regarding trustworthiness and privacy concerns, particularly concerning the handling of personal data ( Kleine-Kalmer, 2016 ). Consumers tend to perceive SNSs as more credible and expedient sources of information regarding products and services compared to information directly disseminated by companies ( Mangold & Faulds, 2009 ). Consequently, it is anticipated by previous studies that trust plays a pivotal role in influencing consumer engagement ( Jayasingh, 2019 ). For instance, consider a user browsing through Facebook and encountering a sponsored post from a travel agency showcasing glowing testimonials from satisfied customers. Trusting the authenticity of these testimonials, the user in Cambodia explores further into the advertised vacation package and ultimately decides to make a booking. Subsequently, they engage with the post by expressing enthusiasm about their upcoming trip. In this scenario, the user’s trust in customer testimonials directly impacts their engagement behavior and contributes to driving sales for the travel agency.

Hypothesis 4 (H4): Trust positively influences consumer engagement behavior.

Gironda and Korgaonkar (2014) and Muntinga et al. (2011) have highlighted the significant influence of friends, family, and social pressure on individuals’ decisions to join and actively participate in brand-related communities. Additionally, Wang and Sun (2016) observed that social influence plays a crucial role in shaping intentional customer engagement behavior. Based on these findings, Bolton et al. (2013) assert that social influence directly impacts customer engagement, emphasizing the tendency for individuals to align their attitudes, behaviors, and beliefs with their social environment. In line with these research insights, it can be inferred that social influence exerts a direct influence on customer engagement behavior ( Jayasingh, 2019 ). For instance, consider a scenario where a Cambodian customer comes across a post featuring a popular influencer endorsing a fashion brand in Facebook. Influenced by the influencer’s style and influence, the user engages with the post by liking it, leaving a comment expressing admiration, and sharing it with their friends. In this example, the influencer’s endorsement significantly influences the user’s engagement behavior and prompts their friends to explore the brand as well.

Hypothesis 5 (H5): Social influence positively affects consumer engagement behavior.

The concept of brand love varies among scholars and contexts, often referring to the intense emotional attachment consumers feel towards a particular brand or trademark ( Carroll & Ahuvia, 2006 ). Fetscherin and Heinrich (2014) characterizes brand love as extreme emotions, whether positive or negative, that consumers harbor towards brands. This relatively recent marketing concept has garnered attention due to its significant influence and strong correlation with customer engagement and brand loyalty on social networking sites ( André, 2015 ; Kleine-Kalmer, 2016 ). Following previous studies, it can be inferred that brand love exerts a positive influence on customer engagement. For instance, imagine a Facebook user in Cambodia encountering a post from their favorite clothing brand showcasing a new collection. Motivated by their deep affection for the brand, the user engages with the post by liking, commenting, and sharing it with friends. In this scenario, brand love serves as the driving force behind user engagement, reinforcing their bond with the brand and potentially inspiring others to participate as well.

Hypothesis 6 (H6): Brand love positively impacts consumer engagement behavior.

Numerous prior studies have conceptualized that highly engaged consumers tend to demonstrate increased brand loyalty ( France et al., 2016 ). Research into online brand communities reveals that customers express their loyalty, satisfaction, and sentiments towards a brand by recommending it to their social circles ( Brodie et al., 2011 ). Engaging in brand community activities on platforms like Facebook enhances consumers’ perceived value, ultimately leading to higher levels of satisfaction and loyalty ( Vivek et al., 2012 ; Jayasingh, 2019 ). In other words, imagine a Facebook user who regularly interacts with posts from a local coffee shop, sharing photos and leaving positive comments. The coffee shop reciprocates by expressing gratitude and offering exclusive promotions. Over time, the user develops a sense of loyalty to the coffee shop, consistently choosing it over competing establishments. This also applies to Cambodia. This demonstrates how engagement on Facebook contributes to building and reinforcing brand loyalty.

Hypothesis 7 (H7): Consumer engagement behavior positively influences brand loyalty towards the brand.

From the literature reviewed and the formulated hypotheses, this study investigates the connections among several factors. These encompass independent variables such as information seeking, entertainment, economic benefit, trust, social influence, brand love, and customer engagement, alongside dependent variables, namely customer engagement and brand loyalty. Consequently, the research develops and validates a conceptual model (as shown in Figure 1 ) that explains the influence of social networking sites on brand loyalty in Cambodia through customer engagement, incorporating elements from the new Technology Acceptance Model (TAM) and the Uses and Gratifications Theory (UGT).

To validate our proposed research framework concerning brand loyalty, we adopted a quantitative methodology, encompassing both data collection and analysis. Initially, we conducted an exploratory research design to gain insights into brand loyalty on brand pages and to delineate constructs and dimensions. The findings derived from this study are applicable to a broader context owing to the collection of data from a substantial sample size. This research aligns with exploratory objectives, seeking to explore the impact of social networking sites on brand loyalty. In terms of the quantitative approach, we targeted respondents who have either purchased or engaged with products advertised on Facebook’s brand pages. The selection of respondents for the survey questionnaire was based on the study’s locale situated in Phnom Penh.

Following the recommendation from Bowerman et al. (2016) , for an unknown population, at least 196 samples were required for a quantitative study to represent the whole population. To ensure the representativeness of findings, the research attempted to collect 400 samples for observations. In this paper, several research techniques can be utilized for data collection, including convenient sampling, snowball sampling, and purposive sampling. Convenient sampling is utilized to collect data from readily available population members, snowball sampling involves participants identifying others to join the sample, and purposive sampling selects individuals based on population characteristics and study objectives. Data collection involves gathering information from target respondents on relevant variables. Both primary and secondary data collection techniques are employed, with primary data gathered through a survey questionnaire distributed to target respondents, and secondary data collected from previous research studies, journal articles, working papers, and relevant books.

In this study, a self-administered survey questionnaire was developed and distributed through social media platforms like Messenger and Telegram to 400 target respondents, who are customers with experience purchasing products from brand pages on Facebook. The study’s survey instrument comprising 34 measurement items was created. Information seeking was evaluated using 4 items (IS1:IS4) adapted from Asghar (2015) . Entertainment was assessed with 4 items (EN1:EN4) sourced from Asghar (2015) and Gummerus et al. (2012) . Economic benefit was measured using 3 items (EB1:EB3) adapted from Azar et al. (2016) and Gummerus et al. (2012) . Trust was evaluated with 4 items (T1:T4) obtained from André (2015) . Social influence was assessed using 4 items (SI1:SI4) adapted from Azar et al. (2016) . Brand love was measured with 5 items (BL1:BL5) sourced from André (2015) . Customer engagement was measured using 7 items (CE1:CE7) obtained from André (2015) and Schivinski et al. (2016) . Finally, brand loyalty was assessed with 3 items (BRL1:BRL3) sourced from Jahn and Kunz (2012) . All questionnaire items were evaluated using a five-point Likert scale, spanning from “Strongly Disagree” to “Strongly Agree.” In addition, the study conducted a pilot study on 40 targeted respondents to ensure the validity of questionnaire items used in the study.

A total of 400 survey questionnaires delivered to targeted respondents who are Facebook users and have experienced of buying products online via Facebook’s band pages. However, only 302 responses were received at the end of data collection. After data cleansing, 250 samples were feasible for data analysis. Among them, 168 respondents identified as female, constituting 67.1% of the sample, while 82 respondents identified as male, representing 32.9%. This suggests a higher representation of women in the study compared to men. Regarding income distribution, 17.5% of respondents reported earning more than $501 - $1000+ per month, followed by 19.5% earning between $301 - $500; 7% earned between $201 - $300; and 56% of them got earning less than or equal to $200 per month. In terms of age distribution, the majority of respondents fell within the 21 to 25 age range (50.2%), followed by those aged 18 to 20 (38.5%), and 26 to 30 (11.3%). Furthermore, the education level distribution of the respondents revealed that 86.7% had a bachelor degree, 8.1% got a high school certificate, and 5.2% graduated master’s degree.

Standard deviations were computed to evaluate the extent of variability, while mean scores were computed for each statement corresponding to the eight research constructs. As displayed in Table 2 , mean scores varied from 3.145 to 3.948, indicating diverse levels of concurrence among participants. This variance may stem from specific statements carrying favorable implications, prompting respondents to express agreement. Additionally, standard deviation values ranged from 0.752 to 1.033, indicating subtle differences in responses among participants.

To ascertain the reliability and dimensionality of the research constructs, the collected data underwent analysis using SPSS 25 software. Specifically, the researcher utilized this software to conduct purification processes, including factor analysis and reliability testing. In Table 3 , the study comprised 34 questionnaire items, of which 32 had a reliability coefficient equal to or greater than 0.60. However, two questionnaire items (IS1 and EN4) with reliability coefficients below 0.60 were excluded from the analysis.

A correlation matrix was employed to determine the association between the mean value of each construct. The analysis result in Table 4 revealed significant positive correlation coefficients ranging from 0.330 to 0.734. Notably, medium-strength correlations were observed between information seeking and customer

Continued

Continued

**Correlation is significant at the 0.01 level (2-tailed).

engagement (r = 0.504, p < 0.01), entertainment and customer engagement (r = 0.551, p < 0.01), economic benefit and customer engagement (r = 0.483, p < 0.01), trust and customer engagement (r = 0.610, p < 0.01), social influence and customer engagement (r = 0.743, p < 0.01), brand love and customer engagement (r = 0.641, p < 0.01), and customer engagement and brand loyalty (r = 0.681, p < 0.01). These findings indicate a positive correlation among all variables, and no high inter-correlation among research variables is found. Therefore, the obtained data is favorable for further analysis.

Confirmatory factor analysis was conducted using AMOS software to evaluate model fit and confirm the research variables, specifically examining questionnaires to ensure reliability coefficients exceeded 0.60. Initially, the study encompassed 34 questionnaires; however, two (IS1 and EN4) were eliminated during the reliability test due to reliability coefficients of Cronbach’s alpha below 0.60 (0.481 and 0.489, respectively). Subsequently, confirmatory factor analysis was performed on the remaining 32 questionnaires, revealing reliability coefficients (or the Standardized loading), shown in Table 5 , ranging from 0.603 to 0.922, all surpassing the threshold of 0.60, suggested by Fornell and Larcker (1981) . Furthermore, the model fit indices showed that a χ²/d.f. ratio was 1.412, goodness of fit (GFI) was 0.920, adjusted goodness of fit (AGFI) was 0.903, normed fit index (NFI) was 0.901, comparative fit index (CFI) was 0.968, the root mean square error of approximation (RMSEA) was 0.041 with significance (p = 0.1000). Note that all of these criteria meet the acceptable thresholds and passed the CFA guidelines of Hair et al. (2019) .

In addition, composite reliability (CR) and the average variance extracted (AVE) were analyzed and the results not shown here. These two index values passed the thresholds (CR > 0.7) and (AVE > 0.5) respectively. These findings indicate that the measurement model has good composite reliability and convergent validity. Furthermore, the discriminant validity (using heterotrait-monotrait ratio (HTMT)) was also applied. All the HTMT values (not shown) are also satisfied achieving below the threshold value of 0.85 ( Henseler et al., 2015 ). Subsequently, all of the constructs are confirmed and ready for structural modeling analysis.

Structural equation modeling was conducted to evaluate the research hypotheses spanning from H1 to H7. The SEM analysis yielded a model achieved with a χ²/d.f. ratio of 1.770. Additionally, goodness of fit index (GFI) was 0.941, adjusted goodness of fit index (AGFI) was 0.912, normed fit index (NFI) was 0.920, comparative fit index (CFI) was 0.939, all meeting acceptable thresholds (above 0.7). Also, the root mean square error of approximation (RMSEA) was 0.056, and the p-value was 0.120. Utilizing AMOS software for SEM, all seven hypotheses were fully supported. As per the guidelines of Hair et al. (2019) , the proposed hypotheses

Note: A is fixed at regression weight 1. ***p < 0.001, **p < 0.05, *p < 0.1.

received positive support, with critical ratios (t-values) of all constructs exceeding |1.96| and p-values below < 0.05, meeting the established criteria. In conclusion, the proposed conceptual model investigating the effect of social networking sites on brand loyalty via customers’ engagement demonstrated a favorable model fit. Furthermore, based on the hypothesis testing results in Table 6 , all variables of social networking sites (information seeking, entertainment, economic benefit, trust, social influence, and brand love) have a significantly positive impact on customer engagement. Meanwhile, customer engagement also yields a positive effect on brand loyalty of customers. Therefore, all hypotheses (H1, H2, H3, H4, H5, H6, and H7) are significantly accepted. Figure 2 shows the confirmed structural model estimation.

Note: A is fixed at regression weight 1. ***p < 0.001, **p < 0.05, *p < 0.1.

The researcher aims to offer theoretical and practical insights along with recommendations to stakeholders, particularly those involved in similar research endeavors and business owners in Cambodia. Utilizing the findings of this study, adhering to the conceptual framework outlined herein is crucial for devising effective marketing strategies. This is especially important for entrepreneurs looking to establish or strengthen their ventures by understanding the factors influencing brand loyalty. Cambodian entrepreneurs should prioritize evidence-based marketing strategies to enhance brand loyalty, considering the unique and evolving landscape of the country.

Entrepreneurs are encouraged to utilize SNSs as platforms for promoting and advertising new products or services. Among SNSs, Facebook stands out as an effective channel due to its extensive user base. With approximately 13,575,700 users in Cambodia, constituting 76.6% of the total population, Facebook offers a significant avenue for businesses to engage with their target audience. Establishing dedicated brand pages on Facebook is advisable to cultivate brand identity and foster community engagement. Utilizing Facebook for marketing not only reduces costs but also facilitates targeted outreach to potential customers in specific geographical areas. Cambodian businesses should invest in developing dynamic and captivating Facebook pages to connect with local consumers, thereby enhancing brand loyalty and market presence.

Given that social influence emerges as the primary driver influencing brand loyalty, businesses should prioritize initiatives that leverage social influence. Cambodian enterprises should focus on building strong relationships with social media influencers to leverage social influence and amplify brand loyalty among local consumers. The inclination to seek information plays a pivotal role in consumer engagement. Consumers turn to brand Facebook pages not only for official product details but also for user-generated content, including experiences and reviews. Encouraging interactions such as likes, comments, and opinions is crucial for brands. Regular consumer promotion campaigns on brand pages can foster engagement and expand the fan base. These campaigns may include sales offers, discounts, freebies, competitions, and interactive games. Furthermore, the study highlights the significance of entertainment in driving consumer engagement on SNSs. Brand pages should regularly feature diverse and visually appealing content, such as jokes, puzzles, games, humorous videos, and cartoons, to captivate SNS users.

This study faces several limitations that warrant further exploration. Firstly, it predominantly focuses on Facebook as the primary social networking platform, reflecting its popularity in Cambodia. Thus, caution is advised when generalizing findings to other platforms or contexts. Additionally, the study employs non-random sampling methods like convenient and snowball sampling due to the unknown population size, potentially impacting result representativeness. Moreover, the sample size of 250 respondents from Phnom Penh may not adequately represent the entirety of customers in Cambodia who purchase products in Facebook. Therefore, future research should consider larger and more diverse samples from various locations and employ robust data collection methods. Furthermore, the study primarily examines consumer-related determinants, neglecting brand-centric factors. Future investigations could focus on these unexplored areas. Additionally, the study fails to differentiate between positive and negative engagement, offering an opportunity for further investigation. Moreover, it adopts a broad approach without focusing on specific brands or product categories, aiming to comprehend engagement motivations and their influence on brand equity more broadly. Future studies could narrow their focus to specific brands or industries to gain nuanced insights. Additionally, exploring potential variations in engagement motivations across different product categories could yield valuable insights.

Over the past decade, the rapid advancement of information technology has reshaped global economies and introduced new societal challenges. SNSs have become indispensable tools for interpersonal communication and connectivity. Among various SNSs, Facebook stands out as the most popular platform, boasting over 2.6 billion users worldwide and more than 13 million users in Cambodia. Subsequently, this study underscores the importance of understanding the factors influencing customer engagement and brand loyalty, particularly within the context of Facebook usage. The SEM results demonstrated significant positive influences of information seeking, entertainment, economic benefit, trust, social influence, and brand love on customer engagement. Additionally, customer engagement exhibited a significant positive influence on brand loyalty of users in Cambodia. The findings also underscore the significance of social influence as the most influential factor shaping brand loyalty among Facebook users in Cambodia.