Keywords

1 Introduction

At present people are getting easier to communicate with each other by the computer-mediated communication (CMC) technologies called social media [1]. Due to the growing amount web traffic, the social network sites (SNS) such as Facebook, Friendster, CyWorld and MySpace are the crucial web applications to access the various visitors in rapid time now a day. Under the statistics collected by Alexa, which is a web traffic data analytics company, Facebook obtained the third rank of the most viewed websites in 2018 after Google and YouTube [2]. With the benefit of friend recommendations, widespread sharing and dynamic responses, Facebook become one of the most powerful tools for today’s’ social networking based e-commerce. Facebook commerce (F-commerce) prescribes to the purchase and selling of products through Facebook [3]. Furthermore, Facebook announced unique features within the site. To illustrate, the famous “like” button, rapid “share”, easy “tag”, dynamic “comment” and “friend” relationships. These features could positively affect the customers’ purchase intention by obtaining more confident after friend recommendation. Social Networking Services (SNSs) consent the people and online sellers to engage in consumer-to-consumer social commerce (C2C s-commerce) [4,5,6]. Nowadays, Facebook users are cumulatively using the Facebook website to conduct the trade operations, by posting advertisements and buying or selling products from each other. This type of conception is called a Facebook C2C social commerce [7].

In Bangladesh the popularities of F-commerce are increasing, almost two thousands of Facebook pages are available [8,9,10]. These survey flourished manuscript has developed based on 5attributes (See Table 1). There are Argument Quality (AQ) [11], Tie Strength (TS) [12], Source Credibility (SC) [13], Product Usefulness Evaluation (PUE) [14] and Purchase Decision (PD) [15]. We have utilized these attributes because these attributes are very helpful and mostly used for finding the information quality in e-commerce. The literature review section has been demonstrated the importance of these five attributes. Our exploration develops a research model to investigate the influence of Facebook C2C communication on consumers’ purchase decision. It also compares the differences between real and virtual relationship environments to better understand consumers’ purchase decision in Facebook. This model has revealed which is based on the Information Adoption Model (IAM). IAM is a relatively new, effective and practical approach [16, 17]. Our findings can provide the better understanding of the role of C2C communication in Facebook communities for consumer behavior research in Bangladeshi domain.

Table 1. Five utilized attributes and their abbreviations.
Full size table

2 Literature Review

A recent study proposed a model which based on the Information Adoption Model (IAM) [18]. The results of this research have shown that AQ, SC, and TS positively influence the purchase decision, where used Partial Least Squares (PLS) technique [19, 20]. We applied this research model in Bangladeshi domain and shown more specific consumers behavior analysis in F-commerce.

Another study analyzed 297 effective data for understanding the influences of system quality, information quality, and service quality for a website to satisfy their consumers [21]. The results of the conception to evaluate the trustworthiness of online sellers showed that people are the key factors that make consumers trust towards online sellers [22]. The social media survey of consumer has shown that trust is more usefulness to make the intention of consumers to purchase through social networking sites [23].

An investigation managed to the effects of F-commerce browsing and usage intensity in predicting urge to purchase and impulse purchase behavior among the consumers [24]. An approach used 180 data that directly investigate and evaluate the success of e-commerce for any small and medium-sized companies [25]. E-commerce service based another research have measured the requirements feasibility for B2C environment [26].

A social learning theory has developed to examine customers’ learning behaviors [27]. A decision support model which has exhibited to manage the buyer risk and improves the quality of supplier development [28]. A theory evaluates using 307 effective data and reveals that cognitive and relational capital positively affects buyers’ loyalty [29]. One team researchers established a casual model to investigate review quality effect on product and purchase intention [30]. In this study, using 349 experimental data that proved demographically similar reviewers enhanced the effect of review quality.

In the above literature review, we observe that most of the research work focused on information quality which makes an intention for the consumer to purchase a product. We have divided the C2C information quality [31] of F-commerce into AQ, TS, SC, PUE and PD sections.

3 Research Model and Hypothesis Development

The intention of a consumer to purchase a product depends on information quality shown by the information adoption model [32, 33]. C2C communication refers to the transfer of the variety of information from one customer to another customer in a way that has the potential to change their preferences, actual purchase behavior, or the way they further interact with others [34]. In line with the information adoption model, the present study proposes that PUE (Product Usefulness Evaluation) has a positive influence on the PD (Purchase Decision) of consumers. Hypothesis 1 is shown as follows:

H1: PUE is positively associated with PD in F-commerce for Bangladesh domain.

According to the dual action models user decision making depends on persuasive information [35]. In accordance with information adoption model, the AQ (Argument Quality) data [11] is a momentous middle sign which exerts an essential influence on information. The Bangladeshi consumers are always eager to know about the product which is suitable for their personal usage [36]. So that, in the online marking, the information quality and consultation of related products is highly appreciated. In the context of AQ information is very meaningful because this makes a clear consciousness about the selected product. Hypothesis 2 is shown as follows:

H2: AQ is positively related to PUE in the Bangladesh region.

Another models user decision making depends on persuasive information [37]. The most common feature of C2C communication is the consumers easily get the advice from other consumers. Based on this shared knowledge the consumer can easily make a judgment which helps those to make the future prediction about consumer’s behaviors [38]. The judgment provides a cue that SC (Source Credibility) is very important to judge product usefulness evaluation (PUE). According to information adoption model, the SC of informational data is a potential peripheral cue which exerts a significant impact on information usefulness. Hypothesis 3 is shown as follows:

H3: SC is positively related to PUE.

Hence, in line with the information adoption model, the present study proposes that SC has a positive impact on the PUE. Members can communicate with others without time and space limitation in online communities [39]. Product messages come from stronger ties may more effective for consumers with a shopping demand to judge the usefulness of a product involved in C2C communications. The TS (Tie Strength) exerts an impact on the effectiveness of word-of-mouth communications in the online context [40]. The individual-level tie strength is an important antecedent of PD in online peer communications [41]. The present study proposes that TS has a positive impact on the PUE at Facebook communications in Bangladesh. Hypothesis 4 is shown as follows:

H4: TS is positively related to PUE.

In the online community contexts, consumers can choose to mainly communicate with real or virtual relationships by participating in different communities [42]. A study pointed out social ties play a different role in consumer decision making in different communication contexts [43]. The strong ties information are more influential in consumer decision making than that from weak ties in the context of offline word of mouth (WOM) communications or C2C interaction [44].

Hypothesis process and proposed research model has been displayed in Fig. 1. H2, H2, H3 hypothesis are connected with H1 where Product Usefulness Evaluation (PUE) is connected with PD for purchasing product in Facebook market place.

Fig. 1.
figure 1

Our research model.

Full size image

4 Methods

4.1 Data Collection and Measurement

A survey was utilized to examine the impact of C2C communication on customer purchase choice. In this research, we did 120 surveys among the Bangladeshi Facebook users who are university students [45]. These survey data has been taken from the Bangladesh Facebook users and strongly involved in product buy-sell at F-Commerce. The analysis of surveys is based on PLS (Partial least squares), where we have used SmartPLS3.0 software [46]. The questionnaire survey consists of five attributes, which are AQ, TS, SC, PUE and PD (See Table 1.). These attributes are focused on purchase decision making to measure the constructs of our research model. Likert scale approach [47] have used for ranging from “1 = strongly disagree” to “7 = strongly agree”, that presented with the set of attributes to measure the theoretical concepts. Following the recommended two-stage analytical procedures [48], we tested the measurement model (validity and reliability of the measures) followed by an examination of the structural model (testing the hypothesized relationship). To test the significance of the path coefficients (p) and the loadings a bootstrapping method was used [49].

5 Sample Profile

The demographic information of the respondents tabulated in Table 2, where derived from the descriptive analysis [50]. The majority of the age group (43.3%) was in the category of 22–25 years old. Most of the respondents were Male which is 76.7% and the frequency of the females is 23.3%. Most of the people are directly or indirectly connected to the internet as well as Facebook. In Table 1, we have analyzed our survey based on gender. Moreover, we separate the males and females and found that the male percentage is 23.3% and the female percentage is 76.7% (Table 3).

Table 2. Demographic information of different genders.
Full size table
Table 3. Demographic information of several ages.
Full size table

6 Result and Data Analysis

6.1 Model Measurement

To assess the measurement model, two types of validity will be examined, first the convergent validity and then the discriminant validity [51]. The convergent validity of the measurement is usually ascertained by examining the loadings, average variance extracted and also the composite reliability [52]. The composite reliabilities were all higher than 0.7 and the AVE (Average Variance Extracted) was also higher than 0.5 as suggested in the literature (see Table 4). The discriminant validity of the measures (the degree to which items differentiate among constructs or measure distinct concepts) was examined by following criterion of comparing the correlations between constructs and the square root of the average variance extracted for that construct (see Table 5). All the values on the diagonals were greater than the corresponding row and column values indicating the measures were discriminant.

Table 4. Convergent validity
Full size table
Table 5. Discriminant validity
Full size table

6.2 Structural Model

To assess the structural model (Structural equation modeling) [53] suggested looking at the path coefficients(p), effect sizes (f2), beta (β) and the corresponding t-statistic [54] via a bootstrapping procedure with a resample of various data or information. They also suggested that in addition to these basic measures researchers should also report the effect sizes (f2) [55].

The outcomes of structural model test display that the path coefficients (p) are statistically significant (p < 0.01), barring for the coefficients of control attributes to PD. The f2 of PD is 0.215, and the f2 of PUE is 0.235 (Table 6).

Table 6. Structural model result
Full size table

Our investigation shows that PUE (β = 0.201, p < 0.01) is a potential predictor of PD. As we hypothesized that PUE would positively effect on PD, hypothesis H1 is susceptible. The AQ (β = −0.152, p < 0.01), SC (β = 0.319, p < 0.01), and TS (β = 0.307, p < 0.01) are predictors of PUE. As we hypothesized that AQ, SC, and TS would positively affect PUE; hypotheses H2, H3, and H4 are adopted (Table 7).

Table 7. Effect size of calculation result
Full size table

7 Discussion

The research goal of this study was to investigate the role of information quality in increasing purchase intention in F-commerce for Bangladesh aspects. The significant impact of perceived reciprocal benefit on knowledge share indicates consumer will share knowledge based on their expectation of future benefit. If the consumer feels that sharing their knowledge in Facebook will be acknowledged by the consumers who sell their product, then the consumers might have a good impression on that their expected consumers. Hence PUE has a positive effect on the PD of consumers. Our second hypothesis H2 implies that if consumers enjoy being on the social network or if they enjoy helping other consumers then they are motivated to post knowledge related data on Facebook. We identified the objectives that increase the PD of F-commerce. AQ, TS, SC, PUE has a positive impact on consumers’ actions.

8 Conclusion and Future Work

With the integrated model of our own concept for the role of information quality in F-commerce for Bangladesh region, we proposed a theory for making a prediction on F-commerce purchase decision-making approach. First, we tested the relationship between product usefulness evaluation and purchase decision making. After that, we tested the relationships of AQ, SC, TS and PUE with the dependent variable purchase decision (PD). Overall, these three arguments are found strong predictors of consumer purchase decision making. Although the findings provide meaningful implications, future studies should address several limitations. First, AQ, SC, and TS may not fully reflect the overall central and peripheral cues. Further study could add other cues. Second, the present study found that age and gender are not significant control variables of the purchase decision. The future study will be conducted on more features for better understanding by integrating the predictors of techno stress [56], perceived enjoyment of purchasing online [57] and security issues [58].