Factors Influencing the Purchase Intention in Online Organic Fruit and Vegetable Stores

The increasing number of organic agricultural production and public awareness toward healthy living provide the opportunities in trading for the products itself. Meanwhile, the development of information technology produces innovation such as electronic commerce (e-commerce), and in agribusiness, it is known as e-agribusiness. E-agribusiness allows organic groceries trade activities to be conducted online through a website. However, that is a new and rare activity in Indonesia, therefore the management of company needs to develop the strategy to enhance the customers and survive in the industry. The purpose of this study is to analyze the factors influencing organic fruit and vegetable purchase intention through a website based on technology perspective and consumer behavior. WebQual and Theory of Planned Behavior (TPB) were used in this study. This study was conducted on website of Sayur Box, one of e-agribusiness in Indonesia. Face-to-face survey technique was used to collect the data from 113 respondents. PLS-SEM was used for data analysis. The result shows that online purchase intention is affected by attitude and perceived behavioral control. All of WebQual's instruments that used in this study affect user satisfaction toward the website except for the usefulness variable. User satisfaction toward the website has an indirect effect on online purchase intention through attitude variable.


INTRODUCTION
Food needs are increasing along with population growth. Therefore, the government prioritizes the development of the agricultural sector to meet the people's food needs while facing the challenges of global economic development (Sitanggang, 2015). One of the government programs in agriculture is the program of 1,000 organic farming villages. The program increases organic agricultural production, in line with the increasing global organic market reaching 20% per year (Sari and Setiaboedhi, 2017). Community interest in organic agricultural products is influenced by trends in healthy lifestyles, especially in big cities that begin to understand the negative effects of using chemicals on agricultural products (Marsyaf, 2017).
On the other hand, current global developments prove that Information and Communication Technology (ICT) plays a very important role for the success of development (Hanuranto, 2011). One form of information technology development is trading through electronic media, or better known as e-commerce. E-commerce developments affect a variety of industrial sectors including the food agriculture sector (referring to the sectors of agriculture, food and beverage products), which is one of the main business sectors throughout the world (Manouselis et al. 2009). As a result, the term e-agribusines appears and becomes a part of e-commerce or e-business; i.e. business activities through electronic media focusing on agricultural commodities (Soekartawi, 2007).
Based on the potential of selling organic products and the trend of online shopping in Indonesia, businesses in this sector are starting to emerge, one of which is Sayur Box which sells organic fruits and vegetables online. Sayur Box is a business to customer (B2C) based e-agribusiness established in July 2016. The business concept of Sayur Box can cut the product distribution chain, thereby allowing higher farmer selling prices and lower consumer purchase prices than those in traditional markets. However, the business of buying and selling organic fruits and vegetables online is not yet popular in Indonesia.
A number of internet users have started to be interested in shopping for agricultural products online, yet the results are not very significant because they tend to choose to visit traditional stores when buying fresh products so they can touch and choose the product themselves (Rizkia, 2015). This is evident from Sayur Box's web traffic which was still ranked at 20,320 as of June 2017 after almost a year of establishment, distantly behind other e-commerce, such as Lazada, ranked fourth after a year of standing based on Indonesian web traffic (Similar, 2017;Ayunita, 2017). The significantly far comparison of the company's growth is the basis of the importance of doing this research, so that Sayur Box can grow and survive amid tight business competition, considering that the government is also currently active in implementing 1,000 digital startup programs. The business of buying and selling business can be seen from the increase in the number of consumers. Therefore, the purpose of this research was to identify the factors that influence consumer intention in purchasing organic fruits and vegetables online in the Sayur Box website. Soekartawi (2007) states that one of the determinants of the success or failure of e-agribusiness depends on the quality of the tool (the software used and the interface and completeness of the information available). In online trading like Sayur Box, it refers to the dimensions of website quality. The website is the main marketing tool; therefore, it is very important for businesses to increase customer satisfaction on the website (Pane, 2009). Websites that are difficult to use and less interactive will project a bad image among internet users and weaken the competitiveness of organizations in the industry (Saputro, 2013). The quality of a website is very important in attracting and retaining consumers. The level of website quality is defined as the extent to which the features on the website meet user needs and reflect the overall superiority of the website (Hsin and Su, 2008). Website features that can meet user needs and highlight website excellence can be the basis of user satisfaction (consumers). WebQual is a method that can be used to evaluate the quality of a website. In the context of using technology for consumers, WebQual is measured by dimensions of ease of use, usefulness, entertainment and complementary relationship (Loiacono et al. 2002).
In addition to technological aspects, management must be able to assess consumer behavior to determine the right strategy for the company. Theory of Planned Behavior (TPB) is a method that is often used to predict behavior. Maichum et al. (2016) used TPB to examine the intention in purchasing agricultural products for consumers in Thailand. Thus, the results of the study showed that consumer attitudes, subjective norms, and The study used primary and secondary data. The primary data was obtained by conducting face-to-face surveys to respondents in accordance with the predetermined criteria. Meanwhile the secondary data were obtained based on the results of a literature study of a number of scientific publications such as books, research reports, research journals, theses, and dissertations. Articles in relevant print and electronic media were also used to support research.
The primary data collection was carried out using the face-to-face survey method of 113 respondents. The respondents were randomly selected with the criteria that they were domiciled in the Jabodetabek area, aged over 17 years and had been shopping for organic fruit or vegetables traditionally. The respondents were met directly and then asked to access and surf the Sayur Box website, such as shopping, but not until the payment process. This was carried out so that the respondents were confirmed to have surfed on the Sayur Box website and the experience of respondents could be directly observed. After that, they were asked to fill out the questionnaire completely, if in one questionnaire there were questions that were not answered then the questionnaire was considered invalid. The form of the question posed was a closed question with five Likert scales so that the respondent can choose one of the answers that best suits him, that is, strongly disagree, disagree, neutral, agree and strongly agree.
Determination of hypotheses based on previous researches was to examine the relationships among variables used in this study. The hypotheses built were as follows:  Lodorfos and Dennis (2008) examined consumer intention in purchasing organic food using TPB, the results showed that attitudes, subjective norms and perceived behavioral control have positive influences on purchase intention. Wong and Aini (2016) used TPB to examine the factors that influence intention in purchasing organic meat in Malaysia, and the results of the study show that consumer attitudes, subjective norms and perceived behavioral control influence the purchase intention with a moderate level of significance. Meanwhile, in the research conducted by Irianto (2015), consumer attitudes and subjective norms are factors that positively influence the intention in purchasing organic food. Consumer attitude is a key factor that most influences the intention in purchasing organic food (Mhlophe, 2016).
Based on the explanation of the background of the problem, TPB was used in this research to measure the intention in purchasing organic fruits and vegetables. The purchase intention, specifically online, is because the object being examined is an e-agribusiness site. In addition, this research also measured the quality aspect of the website used to conduct buying and selling. Based on TPB theory, variables used in the research on aspects of consumer behavior covered attitudes, subjective norms and perceived berhavioral control. Meanwhile, the variables examined from the technological aspect included ease of use, usefulness, entertainment, and complementary relationship which are dimensions of WebQual. The user satisfaction variable on the website was seen as having an effect on consumer attitudes and online purchase intention. Perceived benefits and risks that influence consumer attitudes were also examined; however, the stages of the payment, shipping and after-sales service processes were not covered in this research.

METHODS
The research was conducted on the Sayur Box website (www.sayurbox.com) which sells organic fruits and vegetables online in Indonesia. The research location was limited to the Jabodetabek area as this is adjusted to the location of the product delivery set by the company. The study was conducted from May to September 2017. There were 12 latent variables used in this study, consisting of eight exogenous latent variables and four endogenous latent variables. These latent variables were measured by 46 indicators. The definition of measurement for latent variables can be seen in Table  1.  (Bhukya and Singh, 2015).

PR4
Perceived benefits is a belief about positive results related to behavior (Liu et al. 2012).
I can buy organic vegetables or fruits wherever I want if the purchase can be made online (Wani and Malik, 2013).

PB1
Buying organic vegetables or fruits online can save my energy compared to visiting traditional stores (Wani and Malik, 2013;Sinha and Singh, 2016).

PB2
Buying organic vegetables or fruits online can avoid the hassle of driving and parking (Wani and Malik, 2013;Sinha and Singh, 2016).

PB3
Discounts and prizes are available if I buy online (Sinha and Singh, 2016). PB4 Subjective norm refers to social pressure felt to be involved or not involved in behavior (Sentosa and Mat, 2012).
I think my family and colleagues will agree if I buy organic vegetables or fruit online (Liang, 2014).

SN1
People like me have to buy organic vegetables or fruits online (Dakduk et al. 2017). SN2 Media regularly encourages people to shop online (Park, 2003). SN3 The person I value his opinion encouraged me to shop online (Park, 2003). SN4 Perceived behavioral control is how a person's perception of an action is based on their abilities (Sentosa and Mat, 2012).
I have the resources to buy organic vegetables or fruits online (Park, 2003;Sihombing, 2013).

PBC2
Buying organic vegetables or fruits online is easier for me than other methods (Lee et al. 2014).

PBC3
I am able to decide for myself to buy organic vegetables or fruits online (Park, 2003;Liang, 2014).

PBC4
User satisfaction is the response of consumers to the compatibility between expectations and actual performance that is felt after having the products (Kotler, 2005;Rangkuti, 2003).

US1
The website performance is in accordance with user expectations (Eid, 2011). US2 The website knows its users well enough to offer products that are tailored to their needs (Eid, 2011).

US3
Attitude is a positive or negative evaluation in carrying out an action (Sentosa and Mat, 2012).

A1
I will consume more organic vegetables or fruits if available online (Kumar et al. 2014).

A2
Buying organic vegetables or fruits online is better than buying in traditional stores (Al-Debei et al. 2014;Akroush and Al-Debei, 2015).

A3
Buying organic vegetables or fruits online is a convenient thing to do (Park, 2003  Latent Variables Indicators Symbol Online purchase intention is the intention of consumers (website users) to exchange information and conduct transactions online (Xu, 2017). I intend to buy organic fruits or vegetables online in the near future (Dakduk et al. 2017).

OPI1
I predict that I will buy organic fruits and vegetables online in the future (Ali, 2016).

OPI2
There is a great possibility for me to buy organic fruits and vegetables online (Xu, 2017).

OPI3
If I want to buy organic fruits and vegetables, I will buy them online (Xu, 2017). OPI4 The analytical tool used in the study was PLS-SEM. The data were processed using SmartPLS 3 software.
In analysis using PLS-SEM, the amount of data did not become a problem with a minimum of 30 so that the use of 113 data in the study was acceptable. In PLS-SEM, there are two types of analysis: measurement model and structural model.
The development of flowcharts in the study determines the influence of website quality (ease of use, usefulness, entertainment, complementary relationship), user satisfaction, perceived risks, perceived benefits, attitudes, subjective norms, perceived behavioral control on purchasing organic fruit and vegetable products online based on the previous theoretical models as can be seen in Figure 1. The notations used in the model are as follows: x (Indicators of exogenous latent variables); y (Indicators of endogenous latent variables); ξ (Ksi, exogenous latent variables); η (Eta, endogenous latent variables); λx (Lamda (small), outer loading of exogenous latent variables); λy (Lamda (small), outer loading of endogenous latent variables); β (Beta, the coefficient of influence of endogenous variables on endogenous variables); γ (Gamma, the coefficient of influence of exogenous variables on endogenous variables); ς (Zeta (small), model error); δ (Delta (small), measurement error on exogenous latent variables); ε (Epsilon (small), measurement error in endogenous latent variables).

Measurement Model Analysis
The measurement model is interpreted by looking at several things, including: convergent validity, composite reliability, and discriminant validity (Hair et al. 2011). Convergent validity is measuring the amount of outer loading for each indicator against the construct. Outer loading> 0.70 is highly recommended (Hair et al. 2011), but outer loading> 0.55 can still be tolerated (Pirouz, 2006). Based on Figure 2, the value of outer loading for all indicators used in the study is > 0.55 so that the indicators used can be said to be valid for measuring each variable. Besides being seen from the outer loading value, convergent validity can also be seen from the value of average variance extracted (AVE

Analyses of Structural Model and Hypotheses
Structural model analysis was carried out to ensure that structural models are robust and accurate (Hussein, 2015). Structural model testing can be conducted by looking at the value of R square and the value of the path coefficient (Yamin and Kurniawan, 2009; Hair et al. 2011). R square value shows the magnitude of the construct that is able to be explained by other constructs (Yamin and Kurniawan, 2009   Evaluation of the path coefficient value (β-value) is used to examine the significance of the influence among variables and test the hypotheses (Hair et al. 2011). The path coefficient value can be identified after bootstrapping on the model studied. The level of significance for each hypothesis was measured based on the T-table value for 113 data. If the level of significance is high (99% or p <0.01) then t-value is > = 2.36, if the significance level is moderate (95% or p <0.05) then t-value is > = 1.66 and if the significance level is low (90% or p <0.1) then t-value is > = 1.29. Evaluations on path coefficient value, t-value and significance of hypotheses can be seen in Table 4.
Based on Table 4, ease of use has a positive effect on user satisfaction with a moderate significance level (β = 0.267; p <0.05) so H1 is accepted. Meanwhile, the usefulness variable does not affect user satisfaction, so H2 is rejected. The entertainment variable (β = 0.245; p <0.01) and the complement relationship (β = 0.293; p <0.01) have a positive and very significant effect on user satisfaction so that H3 and H4 are accepted. User satisfaction and subjective norm have insignificant t-value on the online purchase intention so that H5 and H7 are rejected. Attitude variable has a positive effect with a high significance level (β = 0.602; p <0.01) on the online purchase intention, so that H6 is accepted. The perceived behavioral control variable has a positive effect with a moderate significance level (β = 0.157; p <0.05) on the online purchase intention so that H8 is accepted. User satisfaction variable (β = 0.192; p <0.01) and perceived benefits (β = 0.581; p <0.01) have a positive effect with a high significance level on attitude variables so that H9 and H11 are accepted. Perceived risks variable has a negative effect with a high significance level (β = -0.202; p <0.01) on attitude variables so that H10 is accepted. Based on the results of the analysis, the results of the model estimation on the the purchase intention research on the Sayur Box website using SmartPLS can be seen in Figure 3.
In the analysis of purchase intention on the Sayur Box website, not all WebQual instruments have influences on user satisfaction on the website, but entertainment and complementary relationship have significant influences on user satisfaction with a high degree of influence. Ease of use variable has an influence on user satisfaction with a moderate significance level, while usefulness variable has no effect on the user satisfaction of the Sayur Box website. User satisfaction with the website has no influence on the online purchase intention even though it has high quality. However, user satisfaction can have an indirect influence on the online purchase intention through the mediator of consumer attitudes. User satisfaction has a significant influence on consumer attitudes with a high influence according to the research conducted by Alonso-Dos-Santos et al.
Perceived risks and benefits influence consumer attitude significantly with a high level of influence in accordance with the research conducted by Doolin et al. (2005) and Wani and Malik (2013). Perceived risks has a negative influence, while the perceived benefits has a positive influence on consumer attitudes.  Consumer intention in purchasing organic fruits and vegetables online is positively influenced by attitude and perceived behavioral control in accordance with the research conducted by Irianto (2015), Suardika et al. (2014), as well as Teng and Wang (2014). The attitude of consumers has a significant influence on online purchase intention with a high degree of influence. Perceived behavioral control has an influence online purchase intention with a moderate influence. Thus, consumer attitude is a key factor that most influences the online purchase intention on the Sayur Box website, in accordance with the research conducted by Mhlophe (2016).

Managerial Implications
Based on the results of the study, the user satisfaction variable has no effect on the online purchase intention. The intention of consumers in purchasing organic fruits and vegetables online is apparently not directly affected by user satisfaction on the website. The most decisive factor is the attitude and perceived behavioral control. Attitude is influenced by perceived risks and benefits and user satisfaction on the website which are the main marketing tools. In the technological aspect, user satisfaction is influenced by ease of use, entertainment and complementary relationship which are instruments of the quality of the website. Therefore, it is important for the management to always improve the quality of the website.
Dissemination of information on the benefits of purchasing organic fruits and vegetables online, the benefits of consuming organic fruits and vegetables, as well as information on organic fruits and vegetables being traded can be a way for the management to foster a positive attitude towards the community. The more information on the product provided, the more knowledge the community has regarding the products being sold, thereby reducing their risk perception. Information can be provided through social media as the closest source of information to the community today. Marketing trends through influencers such as celebrities, vloggers, or other figures can be used by companies to provide that information.
On the Sayur Box website, entertainment and complementary relationship are website quality instruments that most influence user satisfaction. Highly innovative website design is something that can attract consumers. Companies can project company image through website design. Adding live chat features on the website can make it easier for consumers to interact with company representatives such as online customer service/support.

Conclusions
Consumer intention in purchasing organic fruits and vegetables online on the Sayur Box website is influenced by attitude and perceived behavioral control. The attitude of consumers has a significant influence on the intention in online purchases with a high degree of influence. Perceived behavioral control has an influence on the online purchase intention with a moderate influence. Perceived risks and benefits have significant influences on attitudes. Not all WebQual instruments have influences on user satisfaction towards the website. Entertainment, ease of use, and complementary relationship variables affect user satisfaction, while usefulness variable does not affect the user satisfaction on the Sayur Box website. User satisfaction on the website has no influence on the online purchase intention even though the website has high quality. However, user satisfaction can have an influence on the online purchase intention through the mediator of consumer attitude, and user satisfaction has a significant influence on consumer attitudes.
Improving the quality of websites in terms of appearance and addition of live chat feature needs to be taken in account, considering that this greatly affects the quality of the Sayur Box website. In addition, it is necessary to disseminate information on the benefits and organic fruit and vegetable products to improve the positive attitude of consumers so that intention in online purchases on the Sayur Box website is increasing. One medium for disseminating information is through social media which is the current marketing trend.

Recommendations
This research does not require respondents to make purchases of organic fruit or vegetables on the examined website, so that the quality of the website relating to several stages of the transaction cannot be measured. Therefore, further research can be considered to add quality aspects which include ordering, payment, and after-sales services. In addition, the results of the study show that there are still 33.7% of the influences of other variables that have not been identified; therefore, it is suggested that further research can add other variables that can influence the online purchase intention, such as perceived price variable. Meanwhile, suggestions for companies are marketing through social media, adding live chat features on the website, and improving the quality of the website in terms of appearance or design.