Key Factors Affecting Intention to Order Online Food Delivery (OFD)

Abstract

Purpose: This study investigated the impact of key factors such as service quality, perceived benefit and brand familiarity on a consumer's intention to order online food delivery (OFD). In addition, mediating effect of electronic word of mouth (e-WOM) between relationships among selected key variables and OFD purchase intention is also assessed. Research design, data and methodology: This explanatory, quantitative study employed convenience sampling and collected data through online structured questionnaire from 304 respondents who are users of OFD apps based in greater Klang valley region of Malaysia. The data was then subjected to normality and reliability assessment followed by confirmatory factor analysis, validity assessment and structural equation modelling using IBM SPSS AMOS 24.0. Results: Findings revealed that service quality, perceived benefits and brand familiarity affects purchase intention positively and significantly. Perceived benefits demonstrated highest impact on purchase intention followed by brand familiarity and service quality. Findings also suggest that e-WOM fully mediates relationship between brand familiarity and purchase intention, however, the same was not observed for remaining two variables. Conclusions: The study has enriched OFD literature by investigating impact of selected key factors on purchase intention in the context of OFD. Implications, limitations and future research avenues are then discussed.

1. Introduction

Food market is a trillion-dollar industry (Kandasivam, 2017) indicating a big growth opportunity for online food delivery (OFD) business which is currently a small fraction of the sector. According to Statista (2020), the expected market volume of online food delivery sector to be $182, 327 million globally and $410 million for Malaysia by 2024. The percentage of internet usage in Malaysia has increased to 84.2% in 2019 (The Star, 2020) and consequently, the level of trust or confidence for online transactions is continuously increasing among Malaysians. Apart from online shopping and social media, number of consumers adopting to online food shopping as daily life style has also increased (Amir & Rizvi, 2017). Busy working adults and their family members are looking for a convenient way to order food, which saves time, cost and efforts to get their meal (Goh, Ng, Wong & Chong, 2017; Kedah, Ismail, Haque & Ahmed, 2015).

According to an e-consumer survey conducted by Malaysian Communications and Multimedia Commission

(MCMC), majority of the online shoppers (57.2%) prefer home delivery of their purchases. This changing nature of urban consumers also lead to the emergence and growth of the online food delivery services. Broadly, there are two delivery service formats viz. restaurant-to-customer delivery and platform-to-consumer delivery and their revenue was increased from $76m to $145m respectively in Malaysia (Statista, 2019). Users of OFD for both of these types are steadily increasing and expected to be 4.8m and 6.4m by 2023 (Statista, 2019).

Several food delivery companies are offering OFD services in Malaysia such as Food Panda, Food Time, Deliver Eat, Running Man Delivery, Grab Food, Honestbee, Shogun2U and Dahmakan. Food Panda is pioneering company which is also in forefront of the competition and recorded 100% growth in 2017 (Rosli, 2018).

The customer purchase intention in OFD has been discussed time to time (Amir & Rizvi, 2017; Kandasivam, 2017; Novito, 2017). Many obstacles in the online food shopping have been addressed by researchers. Such issues are payment security, receiving low-quality food products, web loading time, transaction problems etc. (Amir & Rizvi, 2017). However, studies related to understanding on combined effect of key factors such service quality, perceived benefit, brand familiarity on purchase intention to OFD are sparse. Additionally, mediating effect of e-WOM in this context is less addressed specially when it comes to Malaysian OFD consumers.

Existing literature consists few studies related to customer intention to order OFD within and outside Malaysia. Time saving and hedonic motivation are two main element that affect consumer attitude and intention to order OFD (Yeo, Goh, & Rezaei, 2017). According to Prabowo and Nugroho (2018) time saving orientation and Hedonic motivation were external factors affecting consumer’s usefulness perception. On the other hand, Quevedo-Silva, Freire, Lima-Filho, Brandao, Isabella and Moreira, (2016) pointed out that attitude and perceived difficulty were significant factors that affect purchase intention. Ketabi, Ranjbarian and Ansari (2014) found that subjective norm, attitude, behavioral control, social influence and credibility impacts purchase intention. Ha, Nguyen, Nguyen and Nguyen (2019) claimed that perceived ease of use, perceived usefulness, trust, subjective norm and attitude exist positive affect on buyer’s online shopping intention. However, such studies emphasis more on factors related to OFD adoption, technical factors related to OFD portals and hedonic factors. When it comes to factors such as brand familiarity, perceived benefits and e-word of mouth, impact of such factors on purchase intention of OFD is less explored in Malaysia.

Exponential demand in OFD in Malaysian market has also fuelled a cut-throat competitions and companies are struggling to offer appropriate value proposition for their consumers. OFD is a frequent phenomenon and retaining customers is of utmost importance which ensures recurring revenue. M-Commerce itself is a highly competitive industry and companies are in race to offer better value to customers and to gain competitive advantage. To remain relevant in the market and stay ahead of the competition, it becomes necessary for OFD businesses to focus on providing value to the customers. It is crucial to have consumer purchase intention as a factor of consideration for OFD businesses as the online food delivery sector is expected to remain bullish. However, consumer purchase intention is always been challenging over the years due to the myriad of competition from other OFD providers. Furthermore, e-WOM has its role in deciding consumer behavior as most of the consumers read other purchasers review before making their decision. OFD is a sensitive market when it comes to WOM as a variety of aspects such as safety, quality, taste, trends etc. drives e-WOM in the context of OFD.

In the face of the importance mentioned above and the changing consumer behavior towards OFD services in Malaysia, studies highlighting the contributing factors towards OFD services among urbanites still remain sparse in the existing literature.

By addressing these gaps, current study intends to provide clearer understanding to marketers and strategists in the industry to comprehend the importance of consumer psychology especially in their behavioral intention. Therefore, aim of the study is to investigate impacts of selected key factors such as service quality, perceived benefit, brand familiarity and e- WOM. on consumers’ intention to order OFD. Research questions of the study include, does service quality impacts intention to order OFD? Does perceived benefits impacts intention to order OFD? Does brand familiarity impacts intention to order OFD? Does e-WOM mediates relationship among these key factors and intention to order OFD?

2. Literature Review

2.1. Critical review of key theories

Theory of Reasoned Action (TRA) depicts the method of forecasting human behavior and illustrates the link between beliefs, attitudes, subjective norm, intentions and behavior of individuals (Fishbein & Ajzen, 1975). Theory of Planned Behavior (TPB) proposed by Ajzen (1987) is an extension TRA by adding element of Perceived Behavior Control (PBC). TPB was then widely used by empirical studies to examine customer behaviour (Dastane, 2020; Dastane, & Fazlin, 2017; Hei & Dastane, 2017).

Davis (1989) formulated Technology Acceptance Model (TAM) which become a widely accepted as a guideline to forecast, interpretation and raise comprehension of various information system adoption for user acceptance. Aziz and Wahid (2018) used “perceived usefulness” as a term in their study which expressed the benefits related to online buying experience like price, convenience, enjoyment and relationship of consumer and supplier. Haba and Dastane (2019) mentioned that TAM describe that consumer’s perceptions of usefulness and ease of use via internet affects consumer behavior of online purchase through internet. In this study, we used the word” perceived benefit” which similar to Haba and Dastane (2018) studied to examine the relationship between perceived benefit with the customer intention to order hailing services online.

The previous studies have focused on TPB and TAM to specially examine the elements subjective norms and attitude toward buyer’s intention on buying via internet (Haba & Dastane, 2018; Hung, Chen & Huang, 2014; Kim, Ferrin & Rao, 2008; Pei, Paswan & Yan, 2014; Santos, 2003; Tham, Dastane, Johari, & Ismail 2019). However, service quality, perceived benefit, brand familiarity and word of mouth (WOM) are not covered in a single model to assess their impact on repurchase intention. Yeo et al. (2017) have focused on behavior intention, attitude, post usage usefulness, convenience motivation, hedonic motivation, prior online purchase experience, price saving and time- saving part for the study in Malaysia, specifically OFD services. Prabowo and Nugroho (2018) reproduced same model of Yeo et al. (2017) by restricting its object.

Figure 1: Proposed Framework

Existing literature has investigated the aspects of service quality on purchase intention mainly for in-store busses and online businesses but when it comes of OFD, service quality construct need a contextualization and literature is sparse in this domain. Similarly, for the construct of perceived benefits, perception of benefits from OFD is different than that of other online services which requires further investigation. In terms of brand familiarity, the literature is vast for conventional market. For online markets, various researchers have used brand familiarity in the context of credibility, low risk, reputation etc. But when it comes to OFD, this aspect is not fully explored or investigated in present literature.

This study seeks to explore and analysis the interactive effects that buyer’s intention to order online food delivery (OFD) of selected four factors of service quality, perceived benefit, brand familiarity and e-WOM by combining aspects from both disciplines of IT and Business.

2.2. Hypothesis Development

Chandon, Wansink and Laurent (2000) interprets E-service quality as comprehensive consumers’ assessment and determination about the quality and electronic service delivery in the fictitious market. Services which consist of delivery services along with customer services related to pre-purchase and post purchase. According to Donni, Dastane, Haba and Selvaraj (2018) there are high expectations of consumer on timely delivery of purchased product. Celik (2016) mentioned that online buyer looks for uncomplicated customer services especially on return policy. Timely delivery of products and product return issues are main concerns of consumers (Dastane, Goi & Rabbanee, 2020). Consumer’s future purchase intention will have negative impact by failure of customer service (Dastane & Fazlin, 2017; Jallow & Dastane, 2016). Therefore, the first hypothesis is constructed to examine the effects of Service Quality on consumer intention to order online food delivery (OFD).

H1: Service Quality positively and significantly impacts OFD purchase intention.

Perceived benefit is defined by Baker, Hutchinson, Moore and Nedungadi (1986) as a self-confidence of customer that she or he will get well off from the online purchase with a particular website. Park and Stoel (2005) interpreted that Perceived benefits are conviction about the good results related to behavior react to a perceived threat. Wee, Ta and Cheok (1995) inspected convenience, price and product type are three elements that high probably affect online customer buying intentions. In addition, they mentioned that product types and convenience affect buyer intention and make buyer overindulge in buying via internet. Laroche, Kim and Zhou (1996) mention that utilitarian value will motivate consumers do the online purchase. Ismagilova, Dwivedi, Slade and Williams (2017) conceptualised rage of perceived benefits for M-commerce which include convenience, economic and information as aspects of utilitarian value. Several studies noticed that price benefit have positively affected buyer’ attitudes and intentions for buying food via internet (Ketabi et al., 2014; Yeo et al., 2017). Therefore, the second hypothesis was formulated to examine the buyer intention in context OFD.

H2: Perceived benefit positively and significantly impacts OFD purchase intention

Konstantopoulou, Rizomyliotis, Konstantoulaki and Badahdah (2018) defined brand familiarity as consumers previous awareness about brand’s information. Both customers and retailers prefer brands with familiarity (Oluwafemi & Dastane, 2016). Hennig-Thurau, Gwinner, Walsh and Gremler (2004) found that positive attitude toward certain brands will affect his/her preferences on certain brands and affect their purchase intention. Lee and Youn (2009) also have same opinion that consumer would have more confidence on the brand with familiarity and effect the purchase intention of the brand. The below hypothesis created to examine the effect of brand familiarity toward OFD purchase intention.

H3: Brand familiarity positively and significantly impacts OFD purchase intention

Rehman, Ilyas, Aslam, and Imran (2016) defined WOM as a verbal correspond and a method to propagate a message. WOM acts as a significant part in marketplace as it assists customers to form their behaviors and perspectives (Cheng & Krumwiede, 2018). In Cheng and Krumwiede (2018) point of view, is it important that e-commerce strategy to grow Small- or medium-sized enterprises (SMEs) especially eWOM which use social media as a pathway, number of social media influencer with millions of followers may bring in potential consumers for business. WOM become more effective than conventional advertisement due to it is cheaper cost with prompt proliferation (Hennig-Thurau et al., 2004). Lee and Youn (2009) stated that it is important to understand customers need of word of mouth (WOM) communication. At the same time, both studies mentioned that buyer’s purchase intention are greatly impacted by Word of mouth (WOM) as buyers mostly depending on others buyer’s review on the product (Rehman et al., 2016). e-WOM on social network effectively affects consumer’s purchase behavior (Ismagilova et al., 2017). Following hypotheses are furnished to examine the effects of e-WOM on consumer intention to order online food delivery (OFD) and concurrently this study evaluate the effects of services quality, perceived benefit and brand familiarity on e-WOM.

H4: E-WOM mediates relationship between service quality and OFD purchase intention

H5: E-WOM mediates relationship between perceived benefits and OFD purchase intention

H6: E-WOM mediates relationship between brand familiarity and OFD purchase intention

3. Research Method

This research employs explanatory approach with quantitative research method based on primary data collection. Data collection is carried out by a self-administered, electronic based questionnaire which was circulated using Google form. Recipients of questionnaire were informed about the objective of this study, purpose of data collected, confidentiality and academic purpose of the data. This participation in survey was on voluntary basis. Convenience sampling technique was used to select sample of 304 respondents based in greater Klang Vally of Malaysia during March 2020 to June 2020. This was a period of COVID-19 pandemic lockdown which resulted in shift of majority of activities online ensuring all respondents have experience of ordering OFD.

The responses were measured on 7-point Likert scale, where scales range from strongly disagree (1) to strongly agree (7). Questionnaire in this study consists of two sections, 32 questions in total. First section (25questions) was intended to provide information measuring the perception of the client of the constructs which are services quality, perceived benefits, brand familiarity, word of mouth on how they affect respondent’s intention order OFD. Second section was designed to collect demographic information about a respondent including gender, age, ethnicity, education and occupation and balance two general question that ask regarding the experience and frequency order OFD.

The collected data was then subjected to analysis. Demographic analysis was conducted initially to assess the attributes of data set followed by normality and reliability assessment. Confirmation factor analysis (CFA), validity assessment and structural equation modelling was then conducted using IBM SPSS AMOS 24.0.

4. Analysis and Results

4.1 Demographic Analysis

The questionnaire was sent to 350 targeted respondents out of which 304 responses were obtained. This sample consists of 92 males (30.3 percent) and 212 females (69.7 percent). The age groups of 46 and above (84 respondents, 27.6%) and 36-40 (83 respondents, 27.3%) are top two highest percentage age group among 304 respondents. Most of the respondents are Chinese, which are 209 (68.8%), followed by Malay 66 (21.7%), Indian 21 (6.9%) and others are 8 persons (2.6%). The greater part of respondents’ own bachelor’s Degree (48.7%), followed by diploma (22.7%), masters (16.8%), others (8.2 percent), doctorate (2%) and matriculation and below (1.6 percent). The majority respondents are employed (76.3%), followed by self-employed (14.2%), unemployed (7.2%) and last 2.3 % are student.

4.2 Normality Assessment

Kim (2013) mentioned that ideal value of skewness and kurtosis should be between -1.0 to + 1.0 as the rule of thumb which mean the sample data show rationally close to normal distribution. The value of skewness between -2.0 to +2.0 and kurtosis’s value within -7.0 to +7.0 still consider acceptable and normal for data set more than 300 (Hair, Black, Babin, Anderson & Tatham, 2006). Normality assessment results of this study show normal distribution because the skewness’s and Kurtosis’s column demonstrates all the statistical values within the range of -2.0 to +2.0 and -7.0 to +7.0 respectively.

4.3 Reliability Assessment

Reliability was assessed by calculating Cronbach’s alpha values. Overall, total 25 items resulted in Cronbach’s Alpha value 0.946 which presents excellent benchmark for the internal consistency. On the other hand, when assessed by individual variables, Service Quality displayed score of 0.808, Perceived Benefits (PB) of 0.783, Brand Familiarity of 0.892, Word of mouth of 0.875 and Purchase Intention of 0.883. All variable’s Cronbach’s Alpha values are within the range of 0.783 and 0.892 demonstrate that the assessment scale is reliable and consistent the rule of thumb which states the alpha value have to more than 0.7 (DeVellis, 1991).

4.4 Measurement Model

The initial run CFA was conducted but the model fit indices did not meet the required level as suggested by Hair et al (2014). The iterative process of deleting items with factor loading less than 0.6 were conducted with deletion of just one item at a time and such item to be deleted first which is having the lowest factor loading. Inspecting the fitness indexes after each deletion, and the process then repeated until achieving the fitness indexes (Hu & Bentler, 1999). Also, modification indices were checked for cross loadings with set threshold of 15 and if above considered for deletion. The final measurement model is presented in figure 2. Overall model fit assessment for this study show that all indicators achieved the acceptance level. CFA result for RMSEA of 0.066, GFI at 0.920 for incremental fit, CFI, IFI, TLI show the outcome 0.964, 0.964 and 0.954 respectively. Finally, the parsimony fit ‘s index Chisq/ Df at 2.301 which is lower than 3.0 is acceptable. The measurement of model fit is then acceptable and shows good model fit for majority of indices, so this study can be continued with validity assessment following recommendations of Kline and Rosenberg (2010).

Figure 2: Measurement Model

4.4. Validity Assessment

Gerbing and Anderson (1992) suggests composite reliability (CR) value of 0.70 or above in order to conclude that measurement model archives convergent validity. CR for the five constructs Service Quality, Perceived Benefit, Brand Familiarity, Purchase Intention and e-WOM in were calculated. The reliability of all constructs in this study gets confirmed with CR values surpassed the threshold of 0.70 as mention above, within range of 0.77 to 0.92. Hair et al. (2014) stated that Average Variance Extracted (AVE) and CR) are tool to assess the convergent validity of the measurement model with the rule of thumb for the values is 0.5 or (Bentler, 1990; Kline & Rosenberg, 2010). The results displayed in table 1 depicts all values above 0.5 supporting that scale measures archives convergent validity. Similar approach was used by several studies in the past (Keong & Dastane, 2019; Lakhwani, Dastane, Satar & Johari, 2020).

Table 1: Convergent and Divergent Validity Assessment

The discriminant validity can be proved when the value of AVE is higher than Maximum Shared Squared variance (MSV) (Gerbing & Anderson, 1992). Referring to table 2, all criteria are thus satisfied and it can be claimed that the model has attained discriminant validity.

Table 2: Discriminant Validity Assessment

4.5 Structural Equation Modelling

As the model has achieved convergent, divergent and discriminant validity, structural model can be formed to test the impact of independent variables on dependent variable.

4.5.1. Direct Effect

Figure 3 shows the structural model with direct effect by eliminating mediator variable. the Chisq/ Df valued at 1.786 achieved the accepted level (<3.00) and at the same time SEM model fit indices values for GFI, CFI, IFI and TLI at 0.950, 0.982, 0.982 and 0.976 show that they are above the threshold (>0.900). Lastly RMSEA in the table state that achieve the level (<0.08) as well, value at 0.051 (Kline & Rosenberg, 2010). This model demonstrates good model fit (Fornell & Larcker, 1981).

Figure 3: Structural model for direct effects

Table 3 shows standardised estimates and corresponding significance level with direct effects of independent variables. It can be observed that direct impacts of all three variables namely service quality, perceived benefits and brand familiarity on purchase intention are significant (p<0.05). Perceived benefits demonstrated highest impact of 28.4% on purchase intention followed by brand familiarity (21.2%) and service quality (15.5%). Thus, hypotheses H1, H2 and H3 are accepted.

Table 3: Standardised estimates and significance level

4.5.2. Mediating Effect

The structural model is then modified by adding e- WOM as a mediator. Figure 3 presents SEM with mediation effect. The modified model achived good model fit as model indices values are well within acceptable range. Corresponding values are Chisq/Df of 2.301, GFI of 0.920, CFI of 0.964, IFI of 0.964, TLI of 0.954 and RMSEA of 0.066. Table 4 shows standardised estimates and corresponding significant level with mediating effect.

Figure 4: Structural model for mediating effect

Table 4: Standardised estimates and significance level

The mediating effects are calculated by multiplying indirect effects of independent variable on dependent through mediating variable. If such effect is higher than direct effect then mediation occurs. Significance level decides whether such effect is partial or full. It can be observed that impact of service quality and brand familiarity on purchase intention is insignificant (P>0.05) after addition of mediation. Effect of service quality through mediating variable (.072 X .503 = 0.03) is less than its direct effect (.16) and so there is no mediating effect. However, for brand familiarity, such indirect effect is more than direct resulting in full mediation. For perceived benefits as well, indirect effect is less than direct effect resulting in no mediation by e-WOM. Thus, hypotheses H₄ and H₆ are rejected but hypothesis H5 is accepted.

4.5.3. Discussion on findings

As current study has evaluated the impact of key variables on purchase intention of OFD, such variables were established in previous literature but for traditional or in-store purchases of goods. The study has extended the theoretical understanding of effects of key variables to the OFD literature.

For the traditional markets service quality has always impacted purchase intention and findings of this study confirms the same for OFD as results are in line with findings of

previous research results that proved service quality have direct impact to consumer willingness to purchases in store as well as on the internet (Chang and Wang, 2012; Cantallops and Salvi, 2014). Findings on perceived benefits is aligned with previous studies which concluded that perceived benefit affects online buying intention (Al-Debei, Akroush and Ashouri, 2015; Aziz & Wahid, 2018; Sheikh, Abbas and Mehmood, 2015). Arora and Aggarwal (2018) examine the parts of perceived benefits such as convenience, price and diversity of goods toward attitude shopping via internet and affect intention buying via internet and the investigation get the similar result.

In terms of brand familiarity, the literature related to impact brand familiarity on purchase intention in the context of OFD is sparse, therefore when compared the results with findings related to traditional market, result was aligned with studies done by Laroche et al. (1996) resulted that intention to purchase the brand was positively affected by brand familiarity.

E-WOM is a proven mediator in existing literature. When it comes to current study findings, its full mediation between brand familiarity of OFD purchase intention is in line with exiting studies (Cheng & Krumwiede, 2018). However, our findings contradict findings of (Ismagilova et al., 2017; Rehman et al., 2016) as in the context of OFD, e- WOM did not show mediating effect between relationship among service quality, perceived benefits and OFD purchase intention.

5. Conclusions

The aim of this study was to investigate impact of selected key factors on purchase intention of OFD. In addition, the study aimed to investigate mediating effect of e-WOM between relationships among independent variables and purchase intention. Objective is thus achieved and it can be concluded that service quality, perceived benefits and brand familiarity affects purchase intention positively and significantly. Perceived benefits demonstrated highest impact on purchase intention followed by brand familiarity and service quality. It can also be concluded that e-WOM plays full mediation between relationship among brand familiarity and purchase intention however mediating effect of e-WOM is not observed between relationship among other two independent variables and purchase intention. The model in this research gives better insights into origin and process of consumer willingness to OFD. Practically, this research dedicates better view for online marketers and food delivery service providers to understand an element that affects customers’ intention to order OFD. Marketers can plan to offer more benefits and enhance brand familiarity at the same time app developers can focus on increasing web service quality. OFD business managers need to focus on generating positive e-WOM by providing better service and customer relationship management. However, it should be noted by practitioners that service quality and perceived benefits are must for ensuring purchase intension irrespective of e-WOM. The study has some limitations related to generalizability as convenience sampling is used. Additionally, majority of the respondents are Malaysian Chinese and so this pose restrictions to generalize the findings for overall population. Future researchers can select variables apart from the ones selected in this study and also can consider to employ stratified or quota sampling.