Ho Chi Minh City, 2022 STATE BANK OF VIETNAM THE MINISTRY OF EDUCATION AND TRAINING HOCHIMINH UNIVERSITY OF BANKING VO THI KIM NGAN STUDENT CODE 030805170155 FACTORS AFFECTING CUSTOMER INTENTION T.
INTRODUCTION AND THESIS SUMMARY
The necessity of the topic
Online food delivery (OFD) apps have become familiar to customers, especially the young, allowing them to order meals online in minutes via mobile apps or websites OFD services connect customers with partner restaurants through internet-based ordering and delivery platforms, delivering meals fast and without travel or lines As a significant segment of retail e-commerce, the OFD market has seen rapid growth worldwide in recent years In Vietnam, the OFD sector is projected to reach about 449 million USD by 2024, up from around 207 million USD today (Euromonitor International, 2020).
Compared with prior years, global eating-out activity at restaurants declined by 58% due to government lockdowns The COVID-19 outbreak has seriously affected the restaurant industry, driving customers to adapt their lifestyles and shift shopping habits from bricks to clicks As a result, the pandemic appears to have a situational effect that positively enhances consumer intention toward online food delivery (OFD) services.
Ho Chi Minh City issued social distancing measures, including stay-in-place and stay-at-home orders that forced food service operations to close or operate with restrictions Consequently, shopping and dining shifted online, turning many activities into digital processes, including food purchasing Online Food Delivery (OFD) services grew rapidly, delivering meals directly to customers’ homes Restaurants undertook digital transformation by boosting social media marketing and partnering with OFD brands to retain customers, stabilize revenue, and quickly adapt to the evolving epidemic situation.
Despite notable milestones in the online food delivery (OFD) market, usage surged during the pandemic A recent survey by Sach Trang shows OFD adoption rising from 68% in 2020 to 88% in 2021 Because COVID-19 is highly contagious and strict contact restrictions plus government distancing directives limited movement, people increasingly ordered food and drinks from OFD brands instead of leaving home Consumers now prefer OFD over offline shopping for convenience, competitive pricing, a broader range of options, fast delivery, and greater access to information.
To meet the rising demand among customers, OFD brands have significantly expanded supplies and improved quality by upgrading app features and service capabilities Yet purchase intention is shaped by many other factors beyond product quality and convenience To stay competitive and grow in today’s market, OFD brands must monitor external influences on their business strategy while also considering diverse customer factors, so they can develop an effective, data-driven competitive strategy.
These study findings offer valuable insights for both restaurants and OFD brands in identifying the factors that most significantly influence customer intent to use online food delivery (OFD) services By clarifying which determinants drive usage, the results can help restaurants choose the right distribution channels for their operations and enable OFD brands to optimize website features and services that customers value most Additionally, while OFD usage intention has been extensively studied in developed countries, it remains under-researched in developing markets such as Vietnam, presenting a meaningful gap this study begins to address.
With the above considerable changes and the increasingly popular status of purchases through apps, the study "Factors Affecting Customer Intention To Use
Online Food Delivery (OFD) services in Ho Chi Minh City are analyzed to identify the factors that influence customers’ purchase intention and to propose practical strategies for OFD brands The study highlights key drivers such as price sensitivity, delivery speed and reliability, app usability, restaurant quality, brand trust, promotions, and perceived safety, all of which shape consumer decisions in Ho Chi Minh City By understanding these factors, OFD providers can tailor marketing efforts, optimize operations, and enhance user experience to improve conversion rates and foster customer loyalty The recommended actions include transparent pricing, faster and more reliable delivery, stronger restaurant partnerships, advanced app features, personalized promotions, and rigorous hygiene standards to meet local expectations These insights offer a roadmap for OFD brands to strengthen competitive positioning, attract new users, and increase repeat purchases in Ho Chi Minh City’s dynamic online food delivery market.
Research objectives
This study aims to identify the factors influencing customers’ purchase intention when using online food delivery apps in Ho Chi Minh City, providing actionable insights for optimizing OFD sales The findings will propose strategies to enhance the sales efficiency of online food delivery (OFD) businesses operating in Ho Chi Minh City.
Objective 1: Investigating and analyzing the factors affecting customers purchase intention through the OFD applications in Ho Chi Minh City
Objective 2: Evaluating the effect of factors from which to examine and appraise the strengths and weaknesses in sales activities on OFD applications to customers in Ho Chi Minh City
Objective 3: Proposing practical implications and strategies to develop sales activities of OFD businesses to customers in Ho Chi Minh City.
Research questions
The study will clarify the following research questions:
(1) Which are the factors that affect customer intention to use Online Food Delivery services in Ho Chi Minh City?
(2) To what extent do these factors affect on customer intention to use Online Food Delivery services in Ho Chi Minh City?
Subject and scope of the research
Research Subjects: Factors affecting customer intention to use OFD services in HCMC
Survey Subject: Customers in HCMC who are intending to use OFD services
Research Scope: o Location: Ho Chi Minh City o Duration: from 01/03/2022 to 01/05/2022
Research method and data
Primary data: Based on the research questionnaire, survey forms were sent directly to respondents or via social media / emails to collect data from responses of customers in HCMC
Secondary data were collected from business documents and reports of OFD enterprises to establish the empirical basis for the study The theoretical framework and related research on purchase intention were gathered from books and peer‑reviewed scientific journals available in the Banking University archives This combined data approach provides a comprehensive literature review and supports rigorous analysis of consumer purchase intention within the OFD sector.
This study adopts a mixed-methods research design, integrating quantitative and qualitative approaches with data processed and analyzed in SPSS 20.0 A comprehensive review of domestic and international literature informs the development and adjustment of the measurement scale and the construction of an appropriate research model Scale reliability is assessed using Cronbach's Alpha and exploratory factor analysis (EFA), while regression analysis is employed to evaluate the impact of the components on the outcomes The study concludes by summarizing the findings and outlining practical implications and recommendations.
Practical significance of the research
This study investigates the factors influencing customer purchase intention through on-demand food delivery (OFD) apps in Vietnam By presenting up-to-date data and insights, the findings help business managers in the Vietnam OFD market to develop effective strategies and optimize decision-making Additionally, the results lay a solid foundation for future research, guiding scholars to analyze customer intention and behavior in OFD platforms.
Thesis structure
This thesis is structured with 5 chapters as below:
Chapter 1: Introduction and Thesis Summary
This chapter establishes the necessity of the study with a clear introduction and outlines its contribution to the Online Food Delivery (OFD) industry It provides an overview of customer intention toward ordering food online and identifies the factors driving OFD usage The chapter also restates the research objectives, research questions, subject, and scope, ensuring a coherent framework and alignment with the study’s goals.
This chapter outlines the essential concepts and theoretical foundations, presents the reference models guiding this study, and provides a synthesis of international and domestic research on OFD usage intention to identify gaps in the literature and to inform the construction of the proposed research model.
Chapter 3: Research method and data
This chapter outlines the research process and the development of measurement scales, which will be refined through additional adjustments to enhance the model's stability and fit with the collected data Chapter 3 provides a detailed discussion of the research process, scale development, and the statistical data analysis techniques used to interpret the findings.
Chapter 4: Research result and discussion
This chapter reports descriptive statistics, reliability analysis, and a calibrated research model using an appropriate scale, setting the stage to assess how different factors influence customer intention The results identify which factors have the most significant impact on customer intention.
Based on the findings from Chapter 4, this chapter concludes the identified problems facing OFD brands and presents actionable solutions and strategies to improve their quality and competitive standing It offers concrete recommendations for enhancing brand performance, customer experience, and service reliability, while outlining the anticipated impact of these improvements on OFD brand equity In addition, the chapter acknowledges research limitations and outlines directions for future studies to guide ongoing enhancement efforts.
Chapter 1 analyzes the current social context and the rationale for selecting the research topic, explaining why this study matters It provides a general overview and outlines the implications for the Online Food Delivery (OFD) business, highlighting the study’s significance to industry stakeholders The chapter also includes a basic review of customer intentions toward OFD services.
This section provides a concise overview of the research objectives, research questions, subject and scope, and the research methods and data, highlighting how these preliminary elements establish the foundation for the study to be developed in the next chapters By detailing the aims, questions, scope, and data collection approach upfront, the thesis creates a clear roadmap for the research process and informs the design of subsequent chapters The content structure section confirms that the thesis comprises five chapters, with a precise presentation of each part outlined below to guide readers through the research journey.
LITERATURE REVIEW
Concepts
2.1.1 Definition of Customer Buying Behavior
Customer buying behavior encompasses all actions customers reveal during the exchange process, including researching options, making a purchase, using and evaluating products, and spending on goods that satisfy their needs It can also be viewed as how customers decide to allocate their assets—money, time, and effort—to purchase and use goods or services to achieve satisfaction.
Buying behavior, as defined by Kotler & Keller (2011), refers to how individuals or groups purchase and dispose of products to satisfy their needs and desires Similarly, Solomon et al (1995) describe customer buying behavior as the process of selecting, acquiring, utilizing, and disposing of products to fulfill personal satisfaction Enis (1974) adds that it is a process in which inputs, through thought processes and actions, lead to the satisfaction of needs and wants Together, these perspectives frame purchasing decisions as a dynamic sequence of choices, product use, and disposition aimed at achieving personal or collective satisfaction.
Despite numerous definitions, customer buying behavior ultimately refers to the process of selecting, acquiring, and disposing of products and services based on customers' needs and desires Experts and academics agree this process is not static; it evolves over time as buying characteristics vary with changing physical and psychological demands Understanding these dynamics helps businesses anticipate shifts in consumer demand and tailor their strategies accordingly.
Purchase intention is the motivation that drives a consumer to be willing to perform a buying behavior Ajzen (1991) defines intention as the motivating factor behind taking action, while Morinez et al (2007) describe it as a state where a buyer is inclined to acquire a specific product under particular conditions The formation of purchase intention is complex because it relates to consumer behavior, perceptions, and attitudes Ghosh (1990) notes that purchase intention can be an effective predictor of the buying process During the purchasing journey, customers may be influenced by internal and external factors.
Purchase intentions are a key indicator for evaluating the effectiveness of a new distribution channel and for determining which geographic markets and customer segments should be targeted (Morwitz et al., 2007) To analyze consumer behavior, it is crucial to understand the attitudes, evaluations, and internal elements that lead to purchase intention (Fishbein and Ajzen, 1977).
2.1.3 Definition of App Purchase Intention
As technology advances, consumer purchasing behavior increasingly centers on app purchase intent—the likelihood that a purchase will be completed within a mobile application—mirroring traditional purchase intent but executed through the app network Delafrooz et al (2011) define app purchase intention as the certainty that a customer will conduct the purchase through the application, while online shopping intention represents the strength of a consumer’s intent to buy via the Internet This evolution highlights that in-app purchases and online shopping share the same core drivers of purchase intent, with the distinction lying in the channel through which the transaction occurs.
(2013) described online purchase intention as willingness to buy via the internet
Likewise, online purchase intention has been defined as the degree to which a buyer is willing to acquire a product from an online application (Pavlou, 2013) George
Online shopping, as defined in 2004, refers to the frequency with which customers purchase goods or services through social networks Compared with websites and traditional sales channels, online purchases conducted via smartphone applications are more advanced and distinctive in form (Christian Fuentes et al., 2017).
This study demonstrates that the intention to purchase through the mobile application reflects consumers' willingness to engage in actual purchasing behavior on online food delivery (OFD) platforms Put differently, higher purchase intention signals a greater likelihood of buying products via OFD apps, highlighting the link between purchase intent and consumer behavior in online food delivery Understanding this relationship can inform strategies to optimize app design, messaging, and incentives to boost sales on OFD applications.
2.1.4 Online Food Delivery (OFD) Services
Online food delivery (OFD) is the process by which food ordered online is prepared and delivered to the consumer In Vietnam, the OFD market has expanded with brands such as Grab Food, GoFood, and Baemin OFD involves ordering directly from local restaurants via mobile applications to have meals delivered to a specified location When a customer places an order through an OFD app and selects a payment method, the restaurant prepares the meal and a delivery driver delivers it to the customer Users can monitor the order status and contact drivers through the app.
Online Food Delivery (OFD) services let customers choose their preferred restaurants and favorite meals and have them delivered to their doorstep, creating a convenient and personalized dining experience The COVID-19 era has further boosted OFD adoption due to social distancing measures and safe operating practices OFD platforms offer clear advantages such as no need to travel, time savings, and fewer order mistakes.
Online food delivery (OFD) services offer convenient access to meals from a diverse selection of restaurants at flexible times and locations They provide current information on restaurant details, promotions, menu options, and customer reviews, and often include order tracking and real-time delivery-driver location Thanks to advances in OFD technology, the food and beverage (F&B) industry can expand market reach, strengthen customer relationships, improve productivity, and minimize order mistakes.
Theoretical perspectives
Fig 2.1 The Buyer Decision Process
According to Phillip Kotler (2011), there are 5 stages in the Buying Process:
Need recognition: customers recognize their own needs and want to satisfy them Those needs arise in daily life, probably from personal desire or commercials, friends of friends, etc
Information search: when customers have interests in a product, they will seek the product information through personal sources, advertisements, friends, the Internet and so on
Evaluation of alternatives: customers will use the available information, based on the wanted product characteristics to evaluate between brands and make the final decision
Purchase decision: Purchase intention could not turn into Actual purchase if one of these two factors occurs: Attitudes of others; and Unexpected situations
Post-purchase behavior describes how customers evaluate a product after use They consider various factors—price, quality, performance, durability, and overall value—to decide whether they will continue using the product in the future This assessment informs their intention to repurchase or discontinue use, which in turn affects customer retention and brand loyalty.
2.2.2 Theory of Planned Behavior – TPB
Numerous studies have applied several theories to investigate customer intention, with the Theory of Planned Behavior (TPB) by Ajzen (1991) serving as the central framework; this theory is an updated and expanded version of the Theory of Reasoned Action (TRA) developed by Fishbein and Ajzen (1975).
Fig 2.2 Theory of Planned Behavior
According to the Theory of Reasoned Action (TRA), two core factors—Attitude (ATT) and Subjective Norms (SN)—shape both intention and behavior, while the Theory of Planned Behavior (TPB) extends TRA by adding Perceived Behavioral Control (PBC) TPB posits that consumers base their decisions on a logical assessment of options, guided by motivation and factual information, which leads to a Behavioral Intention (BI) that ultimately guides actual behavior In TPB, the most accurate predictions of behavior come from BI, which is influenced by ATT, SN, and PBC, making these three factors the key determinants of BI and, consequently, of consumer behavior.
Attitudes: refers to an individual's extent of favorable or unfavorable assessment of a behavior, a person's evaluation of the outcomes obtained from making a specific behavior
Subjective Norms: A person's perception is influenced by the opinions and judgments of important people (parents, spouses, friends, etc.) that the behavior should or should not be performed
Perceived behavioral control: relates to an individual's perception of the ease or difficulty of carrying out a behavior; it depends on the resources availability and the situation to perform the behavior.
2.2.3 Theory of Perceived Risk – TPR
Perceived Risk (opposite of Perceived Safety) is identified as the level of risk that consumers believe in terms of purchasing on e-commerce platform and it negatively affect customer intention
Fig 2.3 Theory of Perceived Risk
Additionally, it also the main factor prevents the desire stage convert to actual purchase during the buying process The use of technology frequently entails risks, which include two aspects:
Product Quality Risk: the possibility that products received by customer are different from the image on the web/app, the product quality is low, etc
Online Transaction Risk: possible risks when customers make e- commerce transactions on electronic devices (password disclosure, data theft, financial fraud, insecurity in the payment system, etc.)
The TPR model serves as the theoretical basis for research on purchase intention and consumer behavior and has recently been applied to online food delivery (OFD) applications Recent studies in OFD contexts show that perceived risk negatively impacts online purchase intention.
Davis (2000) developed the Technology Acceptance Model (TAM) to explain and predict technology adoption and actual use Grounded in the Theory of Reasoned Action (TRA), TAM analyzes the relationships and influences of the factors that shape how and why people accept and use technology.
Fig 2.4 Technology Acceptance Model (TAM)
Perceived usefulness and perceived ease of use positively influence behavioral intention, which ultimately leads to actual system use; moreover, these factors are themselves shaped by external variables, underscoring the contextual nature of technology adoption.
Perceived usefulness, defined by Davis (1989) as the degree to which a person believes that using a particular information system would enhance his or her job performance, is supported by four key quality dimensions: communication, which links all subjects within the information system; system quality, where a high‑quality system enables easier online task performance and more effective information extraction; information quality, ensuring that the system outputs are trustworthy, accurate, and timely; and service quality, which requires a responsive and secure user experience to make the usage process convenient Together, these factors determine how useful users perceive the system to be and influence its adoption, performance impact, and overall user satisfaction.
Perceived ease of use, defined by Davis (1989), is the degree to which a person believes that using a particular system would require little effort This perception is shaped by external variables such as the user's ability to operate electronic devices, prior experience with technology, level of knowledge, and training.
2.2.5 Consumer Behavior In E-Commerce Environment
Fig 2.5 Online Customer Attitude And Behavior Model
Li and Zhang (2002) proposed a ten-factor model to map the relationships among variables in online shopping, highlighting how Attitude toward Online Shopping is directly influenced by five factors: External Environment, Demographics, Personal Characteristics, Service/Product Characteristics, and Web/App Quality Among these, Service/Product Characteristics and Web/App Quality have a direct impact on Consumer Satisfaction, underscoring their role in digital commerce success The model’s diagram presents a clear process flow with stages—Antecedents, Attitude, and Intention—demonstrating how initial conditions affect attitudes and ultimately shape purchase intention in online retail.
Decision making and online purchasing are closely linked in the digital marketplace, with consumer satisfaction present at every stage and dependent on active customer participation in the online buying process This two-way relationship creates ongoing feedback that can influence both current decisions and future purchases, highlighting the importance of engagement, usability, and perceived value in shaping online shopper outcomes.
Empirical studies
Chanmi et al (2021) analyzed six factors—Perceived Usefulness, Perceived Ease of Use, Price Saving Benefit, Time Saving Benefit, Safety Perception, and Trust—that shape Customer Intention to Use Online Food Delivery (CIU) during the COVID-19 pandemic Using data from 1,045 U.S consumers aged 18 and older, the study found that all six factors positively influence CIU, with Perceived Usefulness emerging as the strongest predictor and Trust as the second strongest Consumers are more willing to use OFD if they perceive it offers real benefits, including food quality and order accuracy comparable to restaurant dining, and if safety is assured The analysis also shows that younger generations (Gen Y and Gen Z) are more inclined to use OFD than Baby Boomers Furthermore, both Price Saving Benefit and Time Saving Benefit significantly enhance CIU, indicating that customers perceive OFD as offering more benefits than risks during the pandemic.
Saqib et al (2021) examined four factors shaping OFD usage intention (online food delivery) during the COVID-19 pandemic, finding that Innovativeness and Optimism significantly and positively affect customers’ intention to use OFD services In developing countries, consumers are more inclined to try new innovative technology, viewing it as delivering efficiency and flexibility through time and cost savings and improved pricing Conversely, Discomfort and Insecurity dampen usage intention, as some individuals—especially older adults—feel anxious about using OFD apps and lack experience or control with smartphones Consequently, many customers remain reluctant to order food online because the internet is perceived as a risky platform that lacks the tangible reassurance of physical interaction.
According to Sangeeta et al (2020), Indian consumers who ordered food through online food delivery (OFD) apps during the epidemic exhibited distinct characteristics from those who did not The study collected data from 462 respondents and examined six factors to identify key differences between the two OFD customer groups It found that higher purchase frequency, stronger perceived benefits, and greater product involvement are associated with more frequent OFD use, while higher perceived threat reduces the intention to order via OFD Additionally, the OFD purchasing rate is 56% higher among the younger generation compared with the older group.
Vincent et al (2017) examined seven variables influencing online food delivery (OFD) services through the Technology Acceptance Model (TAM) The findings show that higher perceived usefulness after use and stronger convenience motivation significantly boost customers’ attitudes and their OFD usage intention Time-saving emerges as a key factor, since customers can order hot meals anytime and anywhere without traveling Perceived safety also exerts a noticeable influence on intention, while enjoyment and entertainment provided by OFD apps help foster a positive attitude Based on these insights, firms can attract more customers through sales promotions, offering competitive prices and discounts.
Chetan et al (2019) investigated the factors shaping consumer behavior toward online food delivery (OFD) services, examining how consumer attitudes mediate these effects Based on data from 170 customers in India's food-retail sector, the study found that convenience, control, and ease of information positively influence customer satisfaction and emerge as the most significant drivers of OFD adoption The results show that users can order food with just a few clicks anytime and anywhere, while also gaining control over restaurant choice, food selection, and payment methods Additionally, OFD apps provide visual food information and real-time order tracking, enhancing the overall user experience However, technology anxiety negatively affects some users, particularly older adults, who struggle to perform tasks on smartphones when ordering food.
Annaraud and Berezina (2020) examined consumer intention to use online food delivery (OFD) services by evaluating food quality, fulfillment, customer service, control, and convenience, based on data from 303 frequent OFD users across the United States who interact with platforms such as Uber Eats, Seamless, and GrubHub The study offers practical implications to enhance OFD customer satisfaction: restaurants should use heat-preserving devices and sturdy packaging to maintain food quality; staff should prepare meals precisely to ensure order accuracy; OFD brands should simplify the ordering process and present information clearly to boost credibility; and apps should integrate support features like a chatbot or a call button to enable prompt problem reporting.
Yen (2015) published "Research On The Factors Affecting The Online Shopping Intention of Consumers" with the aim of discovering and evaluating the factors that have significant impact on online shopping intention through internet platforms, providing findings for future researchers The study identifies four major factors affecting the Consumer OFD Purchase Intent: (1) Perceived Consumer Benefits; (2) Usability; (3) Perceived Safety; (4) Subjective norm The author conducted the research through regression analysis using data from 244 observed samples in the Vietnam OFD market.
Khoa (2019) identified several factors influencing online purchase decisions on smartphone applications in Ho Chi Minh City The study employs the Technology Acceptance Model (Davis, Bagozzi & Warshaw, 1989) and uses exploratory factor analysis (EFA) and Cronbach's Alpha with a sample of 300 customers who frequently shop online The results indicate that perceived usefulness, perceived ease of use, infrastructure conditions, and return policy positively impact the decision to purchase online via smartphone apps, whereas perceived risk negatively affects online purchase decisions.
In the context of Covid-19, Lien & Trang (2021) conducted research on the topic
A study conducted in Ho Chi Minh City during the Covid-19 period examines how the pandemic has altered online shopping intentions, revealing that 200 respondents were analyzed with analytical methods and that several factors substantially influence purchase intention, notably Perceived Usefulness, Reference Group, Safety & Security, and Reputation; these findings highlight the importance of practical utility, social influence, trust and safety, and seller reputation in shaping consumers’ online shopping decisions amid the pandemic.
On the contrary, (5) Perceived Risk is considered as the element has negative impact on customer intention
Trang (2021) investigates the determinants of consumers' intention to use the Baemin online food-ordering app in Ho Chi Minh City The study collected 178 responses and analyzed the data with SPSS, employing exploratory factor analysis, reliability testing, difference testing, and regression modeling By integrating the TAM framework, the findings show that perceived usefulness has the strongest impact on intention to use Baemin, followed by social influence and perceived ease of use.
(4) Perceived Reliability and (5) Perceived Price is the final factor
Table 2.1 Summary of recent empirical studies
Factors affecting customer intention to use online food delivery services before and during the
Perceived Ease of Use; Price Saving Benefit; Time Saving Benefit;
All of the factors positively influence on OFD Usage Intention
(OFDO) services in Pakistan: the impact of Covid-
Innovativeness; Optimism have positive effects, meanwhile, Discomfort; Insecurity have negative effects on OFD Usage Intention
Customers response to online food delivery services during COVID-
Affective and Instrumental Beliefs; Perceived
19 outbreak using binary logistic regression
Involvement impact positively but Perceived Threat impact negatively on OFD Usage Intention
Consumer experiences, attitude and behavioral intention toward
Convenience Motivation; Post- Usage Usefulness;
Price and Time Saving; Online Purchase Experience;
With a higher perception of Post-usage Usefulness; Convenience Motivation; Perceived Safety, Customers Attitude will increase significantly
Understanding consumer behavior towards utilization of online food delivery platforms
Convenience; Control; Ease of Information have positive effects on Customer Satisfaction and OFD Usage Intention
Predicting satisfaction and intentions to use online food delivery: What really makes a difference?
All of the variables impact strongly on both Customer
Research on the factors affecting the online shopping intention of
Factors that strongly influence OFD Purchase Intent are Perceived Consumer Benefits and Subjective norm
Research on factors affecting online purchase decisions on smartphone applications in
Usefulness; Ease of Use; Perceived Risk;
Online Purchase Decision on Smartphone
All of the factors (except
Perceived Risk has negative result) have positive impact (+) on Online Purchase Decision on Smartphone
All of the variables affect intention of
HCMC consumers in the period of Covid-
Safety and Security & Online Shopping
Intention positively, whereas, Perceived Risk impact negatively on Online Shopping Intention
Factors affecting customer intent to use Baemin application to purchase online food in HCMC
Perceived Ease of Use; Perceived Risk; Perceived Reliability;
All of the factors (except
Perceived Risk) have significant effects on Baemin Usage Intention
Research gap
After reviewing the studies with related topics to this thesis, discussed below are several research gaps that are noted:
International studies: although OFD service intention and behavior has been thoroughly researched in developed countries (US, UK, India, etc.), there is a lack of study on OFD services in Vietnam
Research conducted in Vietnam has largely focused on overall online shopping intention, with OFD usage intention remaining underexplored Moreover, there are no studies addressing OFD usage during the COVID-19 pandemic in Ho Chi Minh City (HCMC) As a result, this study may reveal differences when compared with existing research from other regions or time periods, highlighting the unique impact of the pandemic on OFD adoption in Vietnam.
This thesis explores the factors that influence customers’ intention to use Online Food Delivery (OFD) services in Ho Chi Minh City (HCMC) By addressing existing research gaps, the study aims to contribute to the development of Vietnam’s OFD industry and offer actionable insights for OFD platforms to boost adoption and growth across the Vietnamese market.
Research model and hypotheses development
Among reference models, the Technology Acceptance Model (TAM) remains the most widely used framework in research on technology, e-commerce, and online shopping Consequently, this study adopts Davis, Bagozzi, and Warshaw’s (1989) TAM as the foundation for developing the research model, emphasizing the core constructs of Perceived Usefulness and Perceived Ease of Use Building on prior investigations into OFD usage intention, the model also incorporates additional determinants—Subjective Norm, Perceived Safety, and Perceived Price—to provide a more comprehensive understanding of consumer adoption behavior in online food delivery contexts.
Rooted in prior research, this study identifies five key determinants of customers’ intention to use online food delivery (OFD) services in Ho Chi Minh City: perceived usefulness, perceived ease of use, subjective norm, perceived safety, and perceived price These factors illuminate how users evaluate OFD platforms and form their adoption decisions, providing guidance for strategies to enhance user uptake in Ho Chi Minh City.
Finally, the proposed research model is as follows:
Fig 2.6 The Proposed Research Model
2.5.2 Hypotheses development a) Perceived Usefulness – PU
Perceived Usefulness (PU) is the subjective belief that using a specific application will enhance one’s performance in an organizational context, and in the context of online food delivery (OFD), PU reflects how useful customers find ordering food online Prior studies indicate that PU has a significant positive impact on customers’ intention to use OFD services, with younger shoppers often preferring online channels to save time and avoid opportunity costs When consumers perceive that OFD apps save time, reduce effort, and offer flexibility, their intention to use OFD services rises Therefore, PU is defined here as the extent to which customers believe that using OFD services is useful for ordering food online, and the study proposes hypothesis H1: PU positively influences the intention to use OFD.
H1: Perceived Usefulness has a positive effect (+) on Customer Intention to Use OFD b) Perceived Ease of Use – PEU
Perceived Ease of Use (PEU) is defined as the extent to which customers believe a system can be used without effort (Davis, 1986) When the web or mobile app interface is straightforward to access and easy to operate, customers’ intention to purchase online increases significantly (Ramayah & Ignatius, 2005) Prior research consistently shows that PEU has a positive and significant impact on willingness to use online food delivery (OFD) services, with higher PEU leading to stronger OFD Usage Intention (Roh & Park, 2019) An easy-to-use application motivates customers to shop and attracts potential users, underpinning hypothesis H2.
H2: Perceived Ease of Use has a positive effect (+) on Customer Intention to Use OFD c) Subjective Norm – SN
Ajzen (1991) defined Subjective Norm (SN), also known as Social Influence, as the perceived social pressure to engage in or abstain from a behavior SN arises from the opinions and judgments of reference groups—such as parents, spouses, and friends—about whether the behavior should be performed, and it can either increase or decrease a customer’s buying intention There is a close link between SN and buying intention (Chang, 1998) For example, studies by Yen (2015) and Trang (2021) show that SN significantly affects the intention to use online food delivery (OFD) services Therefore, Hypothesis H3 is proposed.
H3: Subjective Norm has a positive effect (+) on Customer Intention to Use OFD d) Perceived Safety – PS
Perceived Safety (PS) refers to an individual's perception of trust and certainty when engaging in a particular behavior, and it is the opposite of Perceived Risk Pavlou (2003) classifies online risks into four categories: financial risk, product risk (low quality), privacy risk (personal information illegally disclosed), and security risk (credit card information stolen) Bhatnagar et al (2000) imply that if these risks can be minimized, PS will increase, and so will the intention to shop online.
Large and familiar online food delivery (OFD) brands tend to make customers feel safe and secure when using their services This suggests that PS has a positive influence on customer intention Consequently, the study proposes hypothesis H4.
H4: Perceived Safety has a positive effect (+) on Customer Intention to Use OFD e) Perceived Price – PP
Perceived Price (PP) is not a real price of a product, it is the price assessed by the customer Depends on their personal evaluation, the product will be “cheap” or
Perceived price significantly shapes online consumer judgments, with some buyers viewing products as “expensive” (Kashyap & Bojanic, 2000) Since online purchases prevent direct viewing of the product, consumers often infer quality from price signals (Jiang & Rosenbloom, 2005) In online food delivery (OFD) services, customers view ordering meals via apps as money-saving and conducive to price comparison due to numerous promotions and a diverse selection of restaurants (Hasslinger et al., 2007) Therefore, Hypothesis H5 is proposed.
H5: Perceived Price has a positive effect (+) on Customer Intention to Use OFD
This chapter defines the key concepts and theoretical foundations of customer intention toward OFD services and establishes a solid theoretical framework for analysis It outlines core reference models, including the Technology Acceptance Model (TAM) and the Theory of Reasoned Action (TRA), to explain how perceived usefulness, ease of use, attitudes, and behavioral intention influence OFD adoption It also reviews relevant foreign and domestic research conducted from 2019 to 2021 to situate the study within current scholarship Finally, it identifies research gaps to clarify the unique contribution of this thesis to the literature on OFD customer intention.
Based on the abovementioned foundation, the hypotheses for independent variables and the research model are presented at the end of Chapter 2.
RESEARCH METHODOLOGY
Research process
This thesis is conducted through 6 phases and each step ensures the objectivity and generality of the topic The research process is presented in figure 3.1 as follows:
This study defines its objective, research questions, scope, and subjects, then presents a literature review that covers key concepts, theoretical perspectives, and empirical studies Building on this foundation, the research model and hypotheses are introduced Qualitative research identifies the components and refines the measurement scales, while quantitative research assesses the factors influencing OFD usage intention Data are collected and analyzed using SPSS 20.0, and the process concludes with a thorough revision to finalize the study.
Scale development
In order to conduct the scale development, the author has inherited based on the TAM Model and referred to the available scales from previous studies
The study is measured using 5-Point Likert Scale in which respondents express their degree of agreement (1 - Strongly disagree; 2 - Disagree; 3 - Neutral; 4 - Agree; 5
Table 3.1 Preliminary and modified scale
Code Preliminary scale Modified scale Source
I order food on mobile apps to save more time and efforts
I think that using OFD services saves more time and efforts
Be able to compare among different options
I can easily compare among different options
(restaurants, meals, etc.) on OFD apps
I can order food on mobile apps anytime and anywhere
I think that using OFD services helps me to order food at anytime and anywhere
I receive plenty of information (promotions, new menus, etc.) through OFD apps
Perceived Ease of Use – PEU
PEU1 I find it easy to understand how mobile apps work
I find it easy to understand how OFD apps work
PEU2 The interfaces of mobile apps are easy to use
The interfaces of OFD apps are easy to use
PEU3 The payment systems are easy to use
The payment systems of OFD apps are easy to use
I can easily order and receive food through OFD apps
My family, friends and colleagues support me using OFD services
I intend to use OFD services because my family, friends and colleagues are using OFD services
SN3 The current society thinks
I intend to use OFD services because the current society (people, media, internet, etc.) thinks OFD is a good option
SN4 Ordering food online is I think that using OFD suitable for the current situation services is suitable for the current trend and situation
PS1 OFD services are safe to use Tu Thi Hai
Professional app interface makes me feel safe when using OFD services
I feel secure when making payments
I feel secure when making payments through familiar OFD apps
The more detailed and clear the restaurant/dishes information, the safer I feel when using OFD services
PP1 Prices are equivalent to the food quality
Prices of OFD services are equivalent to the food quality
OFD services have attractive promotions to me
PP3 Reasonable prices Prices of OFD services are reasonable to me
PP4 Prices are clearly presented
Prices are clearly presented on OFD apps
Customer Intention to Use OFD – CIU
CIU1 I will use OFD apps to order food
I will use OFD services in the future
CIU2 I am planning to use OFD services in the future
CIU3 I will often use OFD services
I will recommend others to order food on the mobile app(s) that I use
I will recommend other people to use OFD services to order food
Several modifications are made to make the scale more adequate and straightforward, which will be discussed in detail in qualitative research section.
Qualitative research
Qualitative research aims to explore, supplement, and refine the scales of observed variables for use in subsequent quantitative analysis The study unfolds in two stages: (1) developing a research model and a preliminary scale grounded in theoretical foundations and a review of related literature; (2) conducting one-on-one interviews with five online food delivery (OFD) customers to adjust the preliminary scale The participants are experienced OFD users who have more than two years of online food ordering experience through OFD applications.
In order to make the scale more adequate and straightforward, one-to-one discussions will be performed to specify needed modifications for the preliminary scale:
Participants will complete a qualitative research questionnaire consisting of two parts (Appendix 1: Discussion Outline) In Part I, they will identify the factors that affect customers’ intention to use OFD services in Ho Chi Minh City (HCMC) based on their own experiences and viewpoints In Part II, participants will offer suggested adjustments and express personal opinions about the scale statements, indicating whether they are clear and easy to understand.
After completing the interview, the author will modify the questionnaire based on the collected data
Discuss again with the participants using modified scales When the discussion questions give the same results repeatedly without any new changes, the qualitative research process will end
Participants evaluated the scale statements for clarity and ease of understanding and assessed whether any variables should be added, removed, or modified The majority agreed with the developed scales for factors affecting customers’ intention to use online food delivery (OFD) services in Ho Chi Minh City (HCMC) Some respondents suggested adding and refining a few words to improve coherence, detail, and readability, ensuring the questions are not confusing or open to misinterpretation.
To complete the qualitative research phase, the researcher synthesized the collected data and subsequently refined the measurement scale, culminating in a quantitative survey comprising 24 observed variables (Appendix 2: Survey Questionnaire) The instrument features the Perceived Usefulness (PU) scale as a central component of the survey design.
PU represents the degree to which customers believe that online food delivery (OFD) services provide meaningful benefits to them The preliminary measurement scale comprises four observed variables, and aside from minor wording adjustments, no changes were made after the qualitative research phase.
PU modified scale is presented as follows:
PU1 I think that using OFD services saves more time and efforts
PU2 I can easily compare among different options (restaurants, meals, etc.) on OFD apps
PU3 I think that using OFD services helps me to order food at anytime and anywhere
PU4 I receive plenty of information (promotions, new menus, etc.) through OFD apps b) Perceived Ease of Use (PEU) scale
Perceived Ease of Use (PEU) is defined as the degree to which customers believe that no effort is required to use a particular system To enhance accuracy, several corrections were made to the statements, and these adjustments did not alter the underlying construct The preliminary PEU scale consists of four observed variables, and there were no changes after the qualitative research process The PEU modified scale is presented below.
Table 3.3 Perceived Ease of Use scale
PEU1 I find it easy to understand how OFD apps work
PEU2 The interfaces of OFD apps are easy to use
PEU3 The payment systems of OFD apps are easy to use
PEU4 I can easily order and receive food through OFD apps c) Subjective Norm (SN) scale
Social norms (SN) are understood as the perceived social pressure to embark on or abstain from a given behavior The preliminary SN scale comprises four observed variables, and aside from wording adjustments, there were no changes after the qualitative research The modified SN scale is presented here.
SN1 My family, friends and colleagues support me using OFD services
SN2 I intend to use OFD services because my family, friends and colleagues are using OFD services
SN3 I intend to use OFD services because the current society
(people, media, internet, etc.) thinks OFD is a good option
SN4 I think that using OFD services is suitable for the current trend and situation d) Perceived Safety (PS) scale
PS refers to an individual's perception of uncertainty and the consequences of engaging in a particular behavior After the qualitative research process, four observed variables remained unchanged, and only one statement required a single word addition to provide more detail The PS modified scale is presented below, offering a concise framework for measuring perceived risk and the potential outcomes associated with the behavior in question.
PS1 OFD services are safe to use
PS2 Professional app interface makes me feel safe when using OFD services
PS3 I feel secure when making payments through familiar OFD apps
PS4 The more detailed and clear the restaurant/dishes information, the safer I feel when using OFD services e) Perceived Price (PP) scale
PP is the price assessed by the customer, and the product is perceived as “cheap” or “expensive” depending on their personal evaluation There are no statement additions or deletions, so the number of observed variables remains four The PP modified scale is presented as follows:
PP1 Prices of OFD services are equivalent to the food quality
PP2 OFD services have attractive promotions to me
PP3 Prices of OFD services are reasonable to me
PP4 Prices are clearly presented on OFD apps f) Customer Intention to Use OFD (CIU) scale
CIU is defined as the customer’s intention to continue using or to adopt OFD services The measurement scale remains unchanged, consisting of four observed variables after qualitative research The variable “I am considering using OFD apps to order food” has been rewritten as “I am considering ordering food through OFD apps” to improve clarity and SEO relevance while preserving the concept of a customer’s intention to use online food delivery services.
“I will use OFD apps to order food” The CIU modified scale is presented as follows:
Table 3.7 Customer Intention to Use OFD scale
CIU1 I will use OFD services in the future
CIU2 I am planning to use OFD services in the future
CIU3 I will often use OFD services
CIU4 I will recommend other people to use OFD services to order food
Summary of qualitative research results
Scale modifications in the qualitative research are summarized as below:
Rewrite the statement of one observed variable in CIU scale
Several slight adjustments were made to help respondents understand properly
In conclusion, the model “Factors affecting customer intention to use OFD services in HCM” uses 7 components and 24 observed variables in total.
Quantitative research
Quantitative research evaluates the influence of factors through a systematic process that includes data collection, data analysis, and descriptive statistics The data gathered from survey questionnaires are then analyzed to assess scale reliability and validity and to test the model fit, ensuring robust measurements and credible conclusions.
Quota sampling is a non-probability sampling method used to select the study sample According to Hair et al (2014), the minimum sample size for conducting Exploratory Factor Analysis (EFA) is 50, with a preference for 100 or more The recommended ratio of observations per analytic variable ranges from 5:1 to 10:1, though some researchers suggest a 20:1 ratio for more robust results In this context, the term “number of observations” refers to the required number of valid questionnaires, while a “measurement variable” denotes a survey question that serves as a measurement item.
By selecting targeted survey respondents who are familiar with and intend to purchase through OFD applications, this study collects relevant research data The survey comprises 24 questions using a 5-point Likert scale, representing 24 observed variables across different factors Applying Hair et al (2014) with a 5:1 ratio, the study determines a sample size of 200 to ensure the minimum reliability needed for data analysis This methodological approach provides robust data to understand factors driving purchases via OFD apps and yields reliable insights for optimization and decision-making.
Data for this study were collected using a questionnaire designed by the author based on the proposed research model to survey respondents The survey targeted two groups: customers who have used OFD services and those intending to use OFD services in Ho Chi Minh City (HCMC) Conducted during the COVID-19 outbreak, the data collection emphasized safety, with questionnaires distributed online or delivered directly to respondents After finalizing the instrument, Google Forms were distributed via social networks and email, and the author compiled the completed responses.
Research location: Ho Chi Minh City
Data analysis is performed through the following steps:
Step 1: Collect the responses, remove invalid ones (same point for all variables or blank answers); then data is encrypted, entered, cleaned and analyzed using SPSS 20.0 software
Step 3: Evaluate scale reliability by Cronbach Alpha analysis
Step 4: Exploratory Factor Analysis (EFA)
Step 5: Multivariable regression analysis and test the hypotheses
Several analytical methods will be used in this study, including:
The scale reliability is assessed by the internal consistency method through the
Cronbach's alpha coefficient measures internal consistency reliability, with higher values indicating greater reliability of a measurement scale As Tho & Trang (2007) state, Cronbach's alpha analysis is essential to identify and remove inappropriate variables, because such variables can generate dummy variables and distort the findings.
Cronbach's alpha is a measure of internal consistency reliability, but it does not indicate which items should be removed or retained; it merely shows whether the measured items form a related set To refine item selection, Cronbach's alpha should be used together with the corrected item-total correlation (CITC) to identify items that contribute little to the underlying construct and may be excluded, as noted by Trong & Ngoc (2005).
The criteria that need to be satisfied include:
Cronbach Alpha Reliability Coefficient: higher than 0.8 is a good scale; from
0.7 to 0.8 is usable; lower than 0.6 can be used in case the research concept is new or new in the research context (Nunnally, 1978; Peterson, 1994; Slater, 1995; cited by Trong & Ngoc, 2005) The author selects scales with Cronbach's Alpha coefficient of at least ≥ 0.6 in this research
Corrected Item-Total Correlation is a key measure of internal consistency, indicating how well each observed variable (item) correlates with the remaining scale items; the higher this index, the stronger the association with the rest of the items and the more reliable the item To improve reliability, items with a corrected item-total correlation below 0.3 should be removed from the dataset.
Exploratory Factor Analysis (EFA) is a statistical method used to reduce a large set of interrelated measurable variables into a smaller set of meaningful factors By extracting these latent factors, EFA preserves the essential information content of the original data while simplifying interpretation and enabling more robust analysis, as described by Hair et al (2009).
Method: For multidimensional scale, use Principal components analysis with
Varimax rotation and breakpoint when extracting factors with Eigenvalues ≥
1 For unidimensional scale, the Principal Components factor extraction method is used When the total variance extracted is ≥ 50%, then the scale is acceptable (Tho & Trang, 2007)
Standard: Factor loadings must be ≥ 0.5 to ensure the practical meanings of
Exploratory Factor Analysis (EFA) relies on factor loadings to gauge the relationship between observed variables and latent constructs In practice, the value levels of factor loadings are classified as follows: loadings above 0.3 are the minimum acceptable level, above 0.4 indicate important significance, and above 0.5 reflect practical significance When selecting the threshold for factor loadings, the sample size matters: if the sample size is at least 350, it is acceptable to choose loadings greater than 0.3; if the sample size is about 100, the loadings should be greater than 0.55; and if the sample size is about 50, the criterion is not specified in the provided content.
Building on the theoretical foundation, this study models “Factors Affecting Customer Intention To Use OFD Services In HCMC” using 24 observed variables analyzed through exploratory factor analysis The analysis employs Principal Components Analysis with Varimax rotation and retains factors with eigenvalues greater than 1 (the Kaiser criterion) Consequently, the study will test related requirements, comprising assessments of reliability and validity and the criteria for factor extraction and retention.
Bartlett’s test is statistically significant (Sig < 0.05)
The value of KMO must > 0.5 then factor analysis is suitable for the data
Observed variables with factor loading coefficient ≤ 0.5 will be removed
Keep only factors with Eigenvalue ≥ 1 and Percentage of variance is ≥ 50%
3.4.3.3 Multivariate regression analysis a) Pearson's correlation analysis
To test the hypotheses, the data on satisfactory scales are subjected to Pearson correlation analysis and regression analysis Pearson correlation analysis assesses the linear relationship between the dependent variable and the independent variables When the absolute value of the Pearson correlation approaches 1, the strength of the linear relationship increases It is also essential to examine correlations among the independent variables, since such correlations can significantly affect the regression results by introducing multicollinearity (Trong & Ngoc).
Enter method of multiple regression will be applied to measure the factors affecting customer intention towards OFD services
The hypothesis testing process includes the following steps:
and adjusted are used to assess the multivariable regression model fit
Testing hypotheses about the model fit and each regression coefficients
Test for multicollinearity using Variance Inflation Factor (VIF < 10)
Specify the degree of factors influence (the higher the beta coefficient, the stronger the impact of that factor)
To ensure model reliability, the study checks for assumption violations, including the independence of residuals measured by the Durbin-Watson statistic (ranging from 0 to 4) Following the regression analysis, an independent samples t-test was performed for gender, while a one-way ANOVA was conducted for occupation and income to examine differences in customer satisfaction among those intending to purchase through the OFD application across gender, occupation, and income groups.
Chapter 3 presents preliminary scale development and detailed research methods In the qualitative research, using the one-on-one discussion technique, several adjustments were made for the scale refinement In the quantitative research, through a survey by the questionnaire, the author collected data with a sample size of 200 OFD customers The scale of factors affecting customer intention to use OFD services includes 5 components based on 24 observed variables By using SPSS 20.0, data is encrypted, entered, cleaned and analyzed
In addition, this chapter also indicates related sections such as sample size calculation and data analysis methods (reliability assessment, EFA & regression analysis).
RESEARCH RESULTS
Characteristics of the survey sample
The sample was collected through the survey questionnaire and was conducted during the period from April 19th to May 9th, 2022 A total of 300 questionnaires were distributed, of which 100 were delivered directly to the respondents and 200 Google forms were sent via social networks and emails The survey results obtained
A total of 236 responses were collected (offline: 83 and online: 153) 36 responses were discarded as unsatisfactory due to issues such as duplicated answers and blank or incomplete sheets, leaving 200 valid responses for the research analysis Table 4.1 presents the statistical analysis of the collected sample.
Table 4.1 Characteristics of the survey sample Characteristics (n = 200) Frequency Percent (%)
Source: Results of data analysis
Table 4.1 summarizes the survey results from 200 participants The gender statistics show that 51% of men use online food delivery (OFD) services, compared with 49% of women, indicating only a 2 percentage-point difference between the two gender groups in online food ordering.
Age statistics show that the 18–25 and 25–30 age groups together form the majority of OFD users, accounting for 23% and 69.5% respectively Participants in these two groups typically have Internet knowledge and experience with Online Food Delivery (OFD) services By contrast, the 30+ age group represents the smallest share at 7.5% The primary reason appears to be lower technology adoption among older adults.
Occupation statistics show that office staff account for the largest share at 68.5%, followed by civil servants at 15.5% Records indicate that employees working for companies tend to order food online together through OFD apps to exploit promotions and preferential shipping prices The remaining three groups—freelancers at 9.5%, housewives at 2.5%, and others at 4%—represent the smaller shares.
Through the Income Statistics, it is clear that the 10 – 15 million group has the most OFD service users, which constituted the highest proportion with 78% and the 5 –
Among the income groups surveyed, the 10 million group ranks second, accounting for 13.5% of participants, and people in these income groups are mostly office staff—the primary OFD customers The under-5 million group represents the lowest category with zero participants The remaining two groups—the 15–20 million bracket (6.5%) and the above-20 million bracket (2%)—comprise the rest of respondents.
Table 4.2 Descriptive statistics of the survey sample
Code N Minimum Maximum Mean Standard
Source: Results of data analysis
In the Perceived Usefulness (PU) scale, agreement scores for observed variables in the PU group predominantly range from 3.7 to 4.0 on the Likert scale, placing responses in the '4' mid-range This indicates that OFD customers in Ho Chi Minh City generally agree with the viewpoints measured by the Perceived Usefulness construct.
Across the observed variables in the Perceived Ease of Use (PEU) group, agreement scores predominantly range from 3.9 to 4.0 on the Likert scale, placing them in the mid-to-high '4' category This pattern indicates that OFD customers in Ho Chi Minh City (HCMC) strongly align with the viewpoints reflected in the Perceived Ease of Use construct.
Within the SN group, observed variables predominantly fall between 3.4 and 3.8 on the Likert scale, placing them in the '4' mid-range This pattern indicates that OFD customers in Ho Chi Minh City generally agree with the viewpoints captured on the Subjective Norm scale.
The point of agreement with observed variables in the PS group mostly at 3.0, in the
“4” intermediate range of the Likert scale Therefore, OFD customers in HCMC agree with the viewpoints in Perceived Safety scale
Within the PP group, the observed variables predominantly range from 3.3 to 3.6 on the '4' mid-point of the Likert scale, showing that OFD customers in Ho Chi Minh City agree with the viewpoints reflected in the Perceived Price scale.
In the CIU group, observed-variable agreement scores predominantly range from 3.4 to 3.7 on the Likert scale, placing them in the mid-range around the 4 mark This pattern indicates that OFD customers in Ho Chi Minh City generally agree with the viewpoints captured by the Customer Intention to Use OFD scale.
Testing the scale reliability
Before conducting exploratory factor analysis (EFA), the first step is to assess scale reliability Cronbach's Alpha measures internal consistency by examining the degree of correlation among observed variables within a single factor, indicating whether the items reliably reflect the same underlying construct A satisfactory Cronbach's Alpha suggests the scale is reliable, supporting the suitability of the data for EFA and the credibility of the resulting factor structure.
Analysis identifies which observed variables participate in measuring the factor's concept and which do not If Cronbach's Alpha value is ≥ 0.6, the scale will be acceptable Observed variables which have Corrected Item-Total Correlation value ≤ 0.3 must be removed (Nunnally & Burnstein, 1994) Using the statistical data processing SPSS 20.0 software, each observed variable is put into the test and gives the following results.
Table 4.3 Cronbach’s Alpha analysis results
Scale Mean if Item Deleted
Scale Variance if Item Deleted
Cronbach's Alpha if Item Deleted Perceived Usefulness (PU): Cronbach’s Alpha = 0.800
Perceived Ease of Use (PEU): Cronbach’s Alpha = 0.848
Subjective Norm (SN): Cronbach’s Alpha = 0.879
Perceived Safety (PS): Cronbach’s Alpha = 0.916
Perceived Price (PP): Cronbach’s Alpha = 0.822
Customer Intention to Use OFD (CIU): Cronbach’s Alpha = 0.908
Source: Results of data analysis
Based on Table 4.3, all component scales demonstrate a Cronbach's Alpha above 0.6, indicating strong internal consistency and inter-item correlation within each factor In addition, the Corrected Item-Total Correlation values are all above 0.3, ranging from 0.333 to 0.859, which confirms the adequacy of the observed variables and supports retaining all items.
Analysis revealed that the PP3 item on the Perceived Price scale reduced internal consistency, as removing it increased Cronbach's alpha from 0.822 to 0.922 Consequently, the authors retained the remaining variables and excluded the observed variable PP3 from further analysis.
Table 4.4 Perceived Price scale after removing the observed variable “PP3”
Scale Mean if Item Deleted
Scale Variance if Item Deleted
Cronbach's Alpha if Item Deleted Perceived Price (PP): Cronbach’s Alpha = 0.922
PP3 Being removed to increase Cronbach's alpha coefficient
Source: Results of data analysis
Exploratory Factor Analysis (EFA)
In the EFA analysis, Principal components analysis will be used with Varimax rotation and breakpoint when extracting factors with Eigenvalues > 1 Several relevant requirements will be assessed, as follows:
Bartlett’s test is statistically significant (Sig < 0.05)
If the value of KMO > 0.5 then factor analysis is suitable for the data
Remove observed variables with factor loading coefficient ≤ 0.5
Keep only factors with Eigenvalue ≥ 1 and Percentage of variance is ≥ 50%
4.3.1 EFA analysis of independent variables
Following the results of scale reliability testing, an exploratory factor analysis (EFA) was performed on the observed variables PU1, PU2, PU3, PU4, PEU1, PEU2, PEU3, PEU4, SN1, SN2, SN3, SN4, PR1, PR2, PR3, PR4, PP1, PP2, and PP4 The output results are displayed in Table 4.5, Table 4.6 and Table 4.7.
Table 4.5 KMO & Bartlett's test result of independent variables
Source: Results of data analysis
Table 4.5 reports a KMO value of 0.753, above the 0.5 threshold, and a Bartlett’s test of sphericity (Sig = 0.000) below 0.05; together, these results confirm that the data are suitable for Exploratory Factor Analysis (EFA) and that the observed variables are sufficiently correlated to justify using EFA for this dataset.
Kaiser-Meyer-Olkin Measure of
Table 4.6 Total Variance Explained of independent variables
Initial Eigenvalues Extraction Sums of Squared
Source: Results of data analysis
The Total Variance Explained is 74.701% > 50%, it means that 5 exploratory factors explained 74.701% of the dataset variance As a result, these 5 factors can be used in further data analysis
Table 4.7 Rotated Component Matrix result of independent variables
Source: Results of data analysis
After the components being rotated, results in Table 4.7 indicate 5 factors as follows:
Factor 1 (includes PS1; PS2; PS4; PS3) measures the components of
“Perceived Safety”, thus, keep the name unchanged “Perceived Safety”
Factor 2 (includes SN1; SN4; SN2; SN3) measures the components of
“Subjective Norm”, thus, keep the name unchanged “Subjective Norm”
Factor 3 (includes PEU1; PEU4; PEU2; PEU3) measures the components of “Perceived Ease of Use”, thus, keep the name unchanged “Perceived Ease of Use”
Factor 4 (includes PP4; PP2; PP1) measures the components of “Perceived Price”, thus, keep the name unchanged “Perceived Price”
Factor 5 (includes PU3; PU1; PU2; PU4) measures the components of
“Perceived Usefulness”, thus, keep the name unchanged “Perceived Usefulness”
From the above results, there is no need to adjust the original theoretical model In addition, the author will test the research model's hypotheses as the next step
4.3.2 EFA analysis of dependent variables
After evaluating scale reliability, the author performed Exploratory Factor Analysis (EFA) on the dependent variables CIU1, CIU2, CIU3, and CIU4 The EFA results are shown in Tables 4.8, 4.9, and 4.10, outlining the factor structure and the relationships among the CIU indicators to validate the measurement model.
Table 4.8 KMO & Bartlett's test result of dependent variables
Source: Results of data analysis
Table 4.8 shows a Kaiser-Meyer-Olkin (KMO) value of 0.831, which exceeds the 0.5 threshold, indicating sampling adequacy for factor analysis; Bartlett's test of sphericity is highly significant (p = 0.000), suggesting the correlation matrix is not an identity matrix Together, these results confirm that exploratory factor analysis (EFA) is appropriate for this data set and that the observed variables are closely correlated, justifying factor extraction.
Kaiser-Meyer-Olkin Measure of
Table 4.9 Total Variance Explained of dependent variables
Initial Eigenvalues Extraction Sums of Squared
Source: Results of data analysis
The Total Variance Explained is 78.649% > 50%, it means that exploratory factor explained 78.649% of the dataset variance As a result, this factor can be used in further data analysis
Table 4.10 Rotated Component Matrix result of dependent variables
Source: Results of data analysis
After rotating the components, Table 4.10 shows that the rotated components CIU4, CIU1, CIU3, and CIU2 load on the factors for “Customer Intention to Use OFD,” and therefore the construct name remains “Customer Intention to Use OFD.”
Pearson's correlation analysis and Multivariate regression analysis
Pearson’s correlation coefficient is used to quantify the strength of the linear relationship between two quantitative variables Before performing linear regression analysis, it is essential to examine the correlation between the dependent variable and each independent variable, as well as the correlations among the independent variables themselves.
According to Trong and Ngoc (2008), the Pearson correlation coefficient quantifies the strength and direction of the linear relationship between quantitative variables It measures the association between a dependent variable and its independent variables, with the absolute value of r indicating how tightly the variables are linearly related Values of r near 1 (in absolute value) reflect a very strong linear relationship, while values near 0 suggest a weak linear association A positive correlation (r > 0) means the variables tend to increase together, whereas a negative correlation (r < 0) indicates that as one variable rises, the other tends to fall.
On the contrary, if the correlation coefficient is negative, then it is a reverse relationship
A zero or near-zero r value indicating no linear relationship between two variables does not mean that the variables are completely unrelated; non-linear associations may still exist Consequently, the linear correlation coefficient should be used only to measure the strength of the linear relationship, as Trong (2008) emphasizes When two variables are correlated, the Pearson correlation coefficient suggests a linear association with |r| values greater than 0.1 However, testing for correlation between two independent variables does not effectively detect multicollinearity in regression analysis, since multicollinearity requires additional diagnostics beyond simple pairwise correlation tests.
Multicollinearity occurs when independent variables are highly correlated, causing the model to receive redundant information and making it difficult to isolate the unique effect of each predictor on the dependent variable This tight correlation inflates the standard errors of the regression coefficients and reduces the statistical significance of their tests, so the coefficients become less meaningful even when R-squared remains high In practice, multicollinearity can be diagnosed in SPSS during multiple regression analysis using the Collinearity Diagnostic option.
To determine, quantify, and evaluate the influence of five factor groups derived from exploratory factor analysis (EFA), the author conducted a multivariable linear regression analysis using SPSS version 20.0, with the variables encrypted.
Table 4.11 Encrypt the factor groups
Factor group Observed variables Encrypted variables
Perceived Usefulness PU1; PU2; PU3; PU4 F_PU
Perceived Ease of Use PEU1; PEU2; PEU3; PEU4 F_PEU
Subjective Norm SN1; SN2; SN3; SN4 F_SN
Perceived Safety PS1; PS2; PS3; PS4 F_PS
Perceived Price PP1; PP2; PP4 F_PP
Customer Intention to Use OFD CIU1; CIU2; CIU3; CIU4 F_CIU
The next step is to analyze the correlation coefficient, the results are as follows:
Table 4.12 Pearson correlation coefficient F_PU F_PEU F_SN F_PS F_PP F_CIU
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
Source: Results of data analysis
All correlations between the factors F_PU, F_PEU, F_SN, F_PS, F_PP and the dependent variable F_CIU are statistically significant (p < 0.01), indicating a strong relationship between the independent variables and F_CIU.
The correlation between the factors F_PU, F_PEU, F_SN, F_PS, F_PP has Sig > 0.01, showing that the independent variables are not correlated with each other
Analysis shows that Customer Intention to Use OFD is closely related to the factors of Perceived Usefulness, Perceived Ease of Use, Subjective Norm, Perceived Safety, and Perceived Price It also indicates that a multiple regression model can be used to measure the influence of these factors on OFD Usage Intention among customers in Ho Chi Minh City (HCMC).
The multiple regression equation of this study is presented as follows:
F_CIU = C + β1*F_PU + β2* F_PEU + β3* F_SN + β4* F_PS + β5* F_PP + ε
F_PEU: Perceived Ease of Use
F_CIU: Customer Intention to Use OFD
Regression results are presented in the following tables after performing the regression on SPSS 20.0 software, using the “Enter” method
Model R R 2 Adjusted R 2 Std Error of the Estimate Durbin-Watson
Source: Results of data analysis
With the primary consideration of this topic is to find the relationship and explanation level of the factors affecting “Customer Intention to Use OFD ”, then
R 2 = 0.832 and adjusted R 2 = 0.827 The multiple regression model is suitable in measuring the degree and direction of the independent variables impact in the model on the dependent variable
Table 4.13 presents the regression analysis results and uses the Durbin-Watson (d) statistic to test for autocorrelation The d value ranges from 0 to 4, and a value near 2 indicates that the residuals do not exhibit first-order serial correlation Therefore, the absence of first-order autocorrelation in the residuals is suggested when d is approximately 2, as noted by Trong & Ngoc (2008).
The Durbin-Watson statistic for the model is d = 2.164, a value close to 2, indicating that the residuals are independent and exhibit no detectable correlation, and thus the autocorrelation assumption is not violated.
In the next step, the author will check whether the built regression model is suitable by ANOVA analysis
Table 4.14 ANOVA analysis result Model Sum of Squares df Mean Square F Sig
Source: Results of data analysis
According to the ANOVA results in Table 4.14, the regression model is statistically significant (F = 191.494, p = 0.000), indicating a good fit with the dataset at the 5% significance level The predictors—Perceived Usefulness, Perceived Ease of Use, Subjective Norm, Perceived Safety, and Perceived Price—significantly influence Customer Intention to Use OFD (Online Food Delivery), with p-values below 0.05, reflecting a less than 5% chance of a Type I error for these effects.
Likewise, the author will examine and assess the Beta coefficient of each independent variable, the regression analysis results are presented below:
B Std Error Beta Tolerance VIF
Source: Results of data analysis
The results indicate that the significance level (Sig.) for the independent variables is satisfactory, with Sig < 0.05, meaning these variables significantly influence the dependent variable Consequently, the author concludes that hypotheses H1, H2, H3, H4, and H5 are accepted.
Table 4.15 shows that the VIF coefficients are below 10, indicating that the model does not suffer from multicollinearity According to Trong & Ngoc (2008), multicollinearity occurs only when the VIF exceeds 10.
The regression equation according to the standardized Beta coefficient of this study is rewritten as follows:
F_CIU = 0.223*F_PU + 0.686* F_PEU + 0.249* F_SN + 0.172* F_PS + 0.468* F_PP
The meanings of the above regression coefficients are provided below:
Holding all other factors constant, a one standard deviation increase in perceived usefulness corresponds to a 0.223 standard deviation increase in customer intention to use OFD This positive relationship highlights the key role of perceived usefulness in shaping consumers’ willingness to adopt online food delivery services.
Likewise, in the condition that other factors remain constant, when the factor
"Perceived Ease of Use" increases by 1 standard deviation unit, then the factor
"Customer Intention to Use OFD" will increase to 0.686 standard deviation units
Next, in the condition that other factors remain constant, when the factor
"Subjective Norm" increases by 1 standard deviation unit, then the factor
"Customer Intention to Use OFD" will increase to 0.249 standard deviation units
Similarly, in the condition that other factors remain constant, when the factor
"Perceived Safety" increases by 1 standard deviation unit, then the factor
"Customer Intention to Use OFD" will increase to 0.172 standard deviation units
Finally, in the condition that other factors remain constant, when the factor
"Perceived Price" increases by 1 standard deviation unit, then the factor
"Customer Intention to Use OFD" will increase to 0.468 standard deviation units.
Hypothesis testing
Table 4.16 Summary of hypothesis testing
H 1 : Perceived Usefulness has a positive effect (+) on Customer
H 2 : Perceived Ease of Use has a positive effect (+) on Customer
H 3 : Subjective Norm has a positive effect (+) on Customer
H 4 : Perceived Safety has a positive effect (+) on Customer
H 5 : Perceived Price has a positive effect (+) on Customer Intention to Use OFD
Based on Table 4.16, all five research hypotheses are accepted, specifically: a) Perceived Usefulness
Hypothesis H1: Perceived Usefulness has a positive effect (+) on Customer Intention to Use OFD
Beta coefficient = 0.223, Sig = 0.000 < 5% => Hypothesis H1 is accepted
Comment: this factor has a positive influence (+) on “Customer Intention to Use
Online food delivery (OFD) apps offer the convenience of ordering meals anytime and anywhere, helping users save time and effort The more customers recognize the usefulness of OFD services, the higher their intention to use them In addition, perceived ease of use—how user-friendly the app is—significantly influences adoption, as intuitive interfaces and smooth navigation improve willingness to order through OFD platforms Together, perceived usefulness and perceived ease of use drive engagement and repeat usage of OFD services, reinforcing a positive user experience and long-term adoption.
Hypothesis H2: Perceived Ease of Use has a positive effect (+) on Customer Intention to Use OFD
Beta coefficient = 0.686, Sig = 0.000 < 5% => Hypothesis H2 is accepted
This factor positively influences the customer intention to use online food delivery (OFD) services The customer intention to use OFD rises significantly when the interface is easy to access and straightforward to operate When the app is simple to use, consumers feel comfortable and are more likely to rely on OFD apps for their food orders.
Hypothesis H3: Subjective Norm has a positive effect (+) on Customer Intention to Use OFD
Beta coefficient = 0.249, Sig = 0.000 < 5% => Hypothesis H3 is accepted
Comment: this factor has a positive influence (+) on “Customer Intention to Use
Online food delivery (OFD) customers are highly influenced by social proof from people they know When parents, friends, or other trusted individuals endorse or support ordering through OFD apps, their willingness to place online food orders increases substantially In this context, perceived safety emerges as a key factor; trust signals such as secure payment, reliable delivery times, accurate order tracking, and credible reviews from familiar sources can boost consumer confidence and drive greater adoption of OFD platforms.
Hypothesis H4: Perceived Safety has a positive effect (+) on Customer Intention to Use OFD
Beta coefficient = 0.172, Sig = 0.000 < 5% => Hypothesis H4 is accepted
Comment: this factor has a positive influence (+) on “Customer Intention to Use
Online food delivery (OFD) users are more likely to use OFD services when they perceive strong safety in online purchasing Big, familiar OFD brands maintain credibility and brand loyalty by frequently implementing safety policies and programs, especially at the payment stage, which helps users overcome fears of financial risk In addition, perceived price influences adoption, as transparent pricing and perceived value reinforce trust and willingness to order.
Hypothesis H5: Perceived Price has a positive effect (+) on Customer Intention to Use OFD
Beta coefficient = 0.468, Sig = 0.000 < 5% => Hypothesis H5 is accepted
Comment: this factor has a positive influence (+) on “Customer Intention to Use
Online food delivery (OFD) customers often believe that ordering meals through apps saves money and enables easy price comparison The abundance of promotions and a diverse range of price options encourage consumers to choose OFD services over traditional ordering methods, driven by savings and convenience As a result, users are stimulated to prefer OFD platforms for meal purchases, leveraging discounts and the ability to compare prices in real time.
Normal Distribution
Residuals may fail to follow a normal distribution due to model misspecification, non-constant variance (heteroscedasticity), or an insufficient sample of residuals for reliable analysis Therefore, multiple residual diagnostics are necessary to detect such violations This study assesses the residual distribution by constructing histograms and a P-P plot (probability–probability plot) to evaluate normality and identify departures from the assumed model The findings guide model refinement and improve the robustness of statistical inferences.
(Source: Results of data analysis)
Figure 4.1's frequency distribution histogram shows that the residuals are normally distributed, with a mean effectively zero (Mean = 6.25E-17) and a standard deviation close to one (Std Dev = 0.987) This near-zero mean and unit-variance pattern indicate the residuals closely follow a normal distribution, so the normal distribution hypothesis is not violated.
(Source: Results of data analysis)
In Figure 4.2, the P-P Plot histogram shows that the observation points do not scatter too far from the expected line (diagonal)
4.7 Testing for differences among demographic groups
4.7.1 Customer Intention to Use OFD between Gender groups
When the variable "Customer Intention to Use OFD" was put into the Independent Sample T-Test analysis with the control group "Gender", SPSS software has given the following results:
Table 4.17 Equality of variances test between Gender groups
Levene's Test for Equality of Variances t-test for Equality of Means t df Sig
95% Confidence Interval of the Difference
Source: Results of data analysis
The T-Test’s Sig value = 0.387 > 0.05, it means that there is no statistically significant difference in “Customer Intention to Use OFD” between Gender groups
4.7.2 Customer Intention to Use OFD among different Age groups
When the variable "Customer Intention to Use OFD" was put into the One-way ANOVA analysis with the control group "Age", SPSS software has given the following results:
Table 4.18 Homogeneity of variances test among Age groups
Levene Statistic df1 df2 Sig
Sum of Squares df Mean
Source: Results of data analysis
In the equality of variances test, the Sig value is 0.329, which exceeds 0.05 This indicates that the variances of “Customer Intention to Use OFD” are equal across age groups, satisfying the homogeneity of variances assumption Therefore, the one-way ANOVA is appropriate for analyzing differences in customer intention to use OFD among the different age groups.
The results of ANOVA analysis have Sig = 0.827 > 0.05, thus, the hypothesis
"equal mean" is accepted With the observational data, it is insufficient to confirm that there are differences in “Customer Intention to Use OFD” between Age groups
4.7.3 Customer Intention to Use OFD among different Occupation groups
When the variable "Customer Intention to Use OFD" was put into the One-way ANOVA analysis with the control group "Occupation", SPSS software has given the following results:
Table 4.19 Homogeneity of variances test among Occupation groups
Levene Statistic df1 df2 Sig
Sum of Squares df Mean
Source: Results of data analysis
Levene's test for equality of variances yielded a significance value of 0.511, which is above the 0.05 threshold This indicates that the variances of "Customer Intention to Use OFD" are equal across Occupation groups, satisfying the assumption of homogeneity of variances Consequently, the results of the one-way ANOVA can be used to compare customer intention across occupations.
The results of ANOVA analysis have Sig = 0.674 > 0.05, thus, the hypothesis
"equal mean" is accepted Therefore, it is insufficient to confirm that there are differences in “Customer Intention to Use OFD” between Occupation groups
4.7.4 Customer Intention to Use OFD among different Income groups
When the variable "Customer Intention to Use OFD" was put into the One-way ANOVA analysis with the control group "Income", SPSS software has given the following results:
Table 4.20 Homogeneity of variances test among Income groups
Levene Statistic df1 df2 Sig
Sum of Squares df Mean
Source: Results of data analysis
Levene's test for equality of variances yielded a Sig value of 0.439, which exceeds the 0.05 threshold This indicates that the variances of 'Customer Intention to Use OFD' are equal across income groups, permitting the use of a one-way ANOVA to compare mean levels of customer intention to use OFD among different income levels.
The results of ANOVA analysis have Sig = 0.389 > 0.05, thus, the hypothesis
"equal mean" is accepted With the observational data, it is insufficient to confirm that there are differences in “Customer Intention to Use OFD” between Income groups
In this chapter, the analytical results of the study are presented, beginning with a description of the survey sample characteristics and an assessment of scale reliability using Cronbach's Alpha All scales were found to be reliable and qualified for further analyses To enhance the reliability of the Perceived Price scale, the item PP3 was removed, resulting in an increased reliability coefficient.
Next, the author performed Exploratory Factor Analysis (EFA) on both the independent and dependent variables The Rotated Component Matrix for the independent variables revealed five convergent factors, with no modifications and the original variable names retained The same was true for the dependent variables.
Regression analysis indicates that five factors significantly influence Customer Intention to Use OFD, with all research hypotheses supported (p < 0.05) The frequency distribution histogram and the P-P plot are appropriate and show no violations of assumptions Furthermore, there are no differences in Customer Intention to Use OFD across demographic groups defined by gender, age, occupation, and income.