Sustainability in e-commerce food shopping: an empirical study from the perspective of consumers

E-commerce food shopping (EFS) has experienced significant growth over the past decade, with the global epidemic further accelerating this trend, making consumers’ daily lives more convenient and affordable (Wang et al., 2024). The rapid development of EFS raises concerns about its impact on the sustainability of our food system. Previous studies have summarized and discussed the potential EFS impact on the three dimensions of sustainability in our food system: economic sustainability (e.g. employment opportunities, and business profit and efficiency), environmental sustainability (e.g. delivery, emissions, carbon footprint, and food and packaging waste), and social sustainability (e.g. social interaction, food safety and quality, and life convenience) (Li et al., 2020, 2024; Prencipe et al., 2024; Siragusa and Tumino, 2022; Yue et al., 2024).

However, these initial findings regarding sustainability in EFS are primarily based on academic perspectives and conceptual models from researchers themselves or on literature reviews (Li et al., 2020, 2024; Prencipe et al., 2024; Siragusa and Tumino, 2022; Yue et al., 2024). There is a lack of empirical studies that systematically explore and confirm the EFS impact on the sustainable development of our food system, particularly from the perspective of consumers, who are among the groups most affected by and benefitting from the advancement of EFS. To our knowledge, no empirical consumer study has been conducted to explore consumers’ perceptions toward and general image of the EFS impact on sustainability.

To fill this knowledge gap, in this study we undertook qualitative research to elicit consumers’ perceptions of the EFS impact on sustainability, using a large sample (n = 2,039) from New Zealand (NZ) and the United Kingdom (UK). Additionally, through a quantitative data analysis approach, we also explored significant consumer perceptions, socio-demographics, and consumption experiences that influence whether consumers have a positive or negative image of the sustainability impact of EFS.

Over the past decade, the rapid deployment of EFS has evolved into two major modes: Business-to-Consumer (Business-to consumer (B2C)) and Online-to-Offline (O2O) (Wang et al., 2020, 2024). B2C sellers typically do not have physical stores in consumers’ local regions, while O2O sellers are usually food stores and restaurants placed in the consumers’ neighborhoods (Wang et al., 2020). The O2O mode has several sub-modes based on the location consumers choose for their final food consumption. These include Online-to-Offline Food Delivery Service (O2O-FDS), where food is delivered to consumers’ homes; Click & Collect (C&C), where consumers pick up their food themselves from the sellers' locations; and Online-to-Offline In-store (O2O-IS), where consumers eat at the sellers’ physical stores or restaurants (Kilders et al., 2024; Wang et al., 2020, 2024). There are also other emerging EFS modes. “New Retail” integrates the O2O approach into traditional food retailing, combining more technologies and innovations with multiple food consumption services, such as delivery, and in-store shopping and dining (Wang et al., 2020, 2024). Live-streaming e-commerce is also on the rise, where consumers can order food and non-food products recommended by sellers during live-streaming interactions hosted by the sellers (Mao et al., 2024; Tan, 2024).

Previous studies have generally confirmed the significant impact of four key factors on consumer adoption of EFS (Wang et al., 2024). These factors include innovation-adoption characteristics (e.g. consumers’ perceived incentives, ease of use, and risks associated with EFS), e-commerce food choice motives (e.g. mood, convenience, unavailability of certain products offline, and quality concerns), e-service quality attributes (e.g. app menu design and the trustworthiness of online sellers), and socio-demographics (e.g. age, marital status, and educational level) (e.g. Abed, 2024; Cao and Wang, 2024; Fu et al., 2025; Jansen et al., 2025; Khan et al., 2025; Liu and Fan, 2025; Wang, 2020; Wang et al., 2020; Wang et al., 2024). Additionally, previous scholars have applied other theories to explore the influences of relevant factors on consumer adoption of EFS, such as the theory of planned behavior and word-of-mouth (e.g. Ahmad et al., 2025; Akar, 2024; Hamid and Azhar, 2023; Liao et al., 2025).

Previous studies have also indicated that different EFS modes are better suited for selling various food categories. For instance, O2O modes are more appropriate for restaurant meals and fresh food products, as O2O sellers are typically local food stores and restaurants, which can ensure the freshness and safety of food and meals (Wang et al., 2020, 2024). In contrast, the B2C mode is more suitable for packaged and processed food products. This because B2C sellers do not have physical stores close by consumers’ locations, resulting in comparatively longer delivery distances than O2O sellers (Wang et al., 2020, 2024). However, with the continued development of EFS logistics and delivery systems, the B2C mode has gradually advanced in fresh food delivery, as evidenced by the rise of the “meal kit business” mode. In this mode, fresh ingredients are delivered to consumers’ homes over longer distances by sellers who need not have physical stores close by consumers’ locations (Park et al., 2024; Shin et al., 2024; Wang et al., 2024).

To date, only a few studies have explored the field of sustainability in EFS. Li et al. (2024) provided a sustainability evaluation model for O2O-FDS, which includes actions to enhance environmental sustainability by joining sustainable development programs, reducing disposable tableware, using degradable packaging, and minimizing total packaging. To enhance economic sustainability, the model highlights the economic advantages delivered by direct profit, delivery speed, and fast refunds (Li et al., 2024). To enhance social sustainability, it emphasizes providing consumers with safe and tasty food, building a positive business image, and increasing job opportunities for the community (Li et al., 2024).

Li et al. (2020) developed a conceptual model that identifies the key impacts of O2O-FDS on the three dimensions of sustainability. The impact on economic sustainability includes increased employment opportunities, low job satisfaction, uncertain effects on the traditional catering industry, and the emergence of new business models (Li et al., 2020). The impact on social sustainability includes more convenient urban living, challenges to public health and traffic, rising concerns about the safety of delivery employees, and uncertain effects on relationships between food and people, as well as between individuals (Li et al., 2020). The impact on environmental sustainability includes increased plastic waste and carbon footprint, along with uncertain effects on food waste (Li et al., 2020).

All other available studies focus on the EFS impact on specific dimensions of sustainability. The results by Siragusa and Tumino (2022) confirmed that EFS is more environmentally sustainable than traditional offline food shopping, with lower greenhouse gas (GHG) emissions. While Varese et al. (2024) indicated that EFS may cause environmental and food waste issues. Scholars have also indicated that the use of green and sustainable packaging and delivery methods can significantly benefit the EFS business in terms of environmental and economic sustainability (Kumar et al., 2025; Prencipe et al., 2024; Siragusa and Tumino, 2022; Tudisco et al., 2025; Yue et al., 2024). Khatami et al. (2024) found that EFS contributed to business resilience and sustainability in food enterprises by broadening equitable food access.

Despite the academic advancements mentioned above, there is still a lack of understanding regarding consumer behavior concerning sustainability in EFS. No empirical studies have systematically explored consumer behavior towards sustainability in EFS, including consumer perceptions and the general image consumers have of the impact of EFS on sustainability.

Perceptions relate to consumers’ selection, organization, and interpretation of product information, which originate from their impressions of a product or service and are influenced by the information they accumulated about the product or service by sourcing it through direct experience or via external engagements (e.g. via memory, marketing stimuli, and socio-cultural and economic influences) (Guerrero et al., 2010; Solomon, 2020; Verbeke, 2000; Wang et al., 2018). The general image of a product/service refers to consumers’ overall positive or negative beliefs about it (Almli et al., 2011; Wang et al., 2018). This general image shapes consumer expectations, which, in turn, influence their final purchasing decisions for the product/service (Almli et al., 2011; Wang et al., 2018). Consumer perceptions and general image represent the cognitive and affective stages, respectively, in their decision-making process for food choices; these correspond to the sequential stages of consumers initially recognizing, then gradually shaping and building preferences between food sourcing alternatives (Almli et al., 2011; Verbeke, 2000; Wang et al., 2018). Therefore, consumer perceptions significantly influence the shaping of their general image of a product/service (Almli et al., 2011). Additionally, consumers with different socio-demographic profiles and previous consumption experiences of a product/service differ in their affective stage (e.g. general image and attitudes) when selecting the product/service including EFS (Almli et al., 2011; Goode et al., 2010; Guerrero et al., 2010; Verbeke, 2000; Wang et al., 2018).

Based on the theoretical background above, this study formulates the following hypotheses:

Data were collected in 2022 through an online questionnaire developed in English and distributed between November and December. A reputable research agency was hired to distribute the questionnaire to its consumer sample panels in NZ and in the UK. A total of 2,039 valid responses were received, with a sample size of 1,020 from the UK and of 1,019 from NZ. Table 1 shows the socio-demographic distribution of the sample, including gender, marital status, age, educational level, occupation, household size, residential location (rural/urban), and financial situation. Participants self-evaluated their financial situation on a seven-point scale, ranging from 1 = Difficult to 4 = Moderate, and 7 = Well-off (Vanhonacker et al., 2010).

Before being asked the survey questions, participants were provided with the following introduction to the concepts of EFS and sustainability: “E-commerce food shopping involves the purchase or ordering of food products and meals through mobile apps and online platforms. In recent times, it has become increasingly important to know the sustainability impact of EFS due to the rapid increase in its market share, particularly during the COVID-19 epidemic”.

Sustainability has three dimensions: (1) social sustainability – to treat people, communities and cultures fairly and appropriately; (2) economic sustainability – to be economically feasible; and (3) environmental sustainability-to protect the Earth’s environment”.

Table 2 presents the measures used in this study. Participants’ general EFS frequency over the past year was measured using an eleven-point ordered scale, adapted from a previous study that used it to explore consumers’ consumption frequencies across three EFS modes including B2C, O2O-FDS, and C&C (Wang et al., 2024). This variable was later recoded into a nine-point ordered scale by combining categories 9, 10, and 11 into a new category, “3 times each week or more”, due to the low response rate in these three categories (2.5% for Category 9: “3–4 times each week”; 1% for Category 10: “5–6 times each week”; and 0.4% for Category 11: “7 times each week or more”; For the frequencies of other scale categories, please refer to Table 3).

Participants who did not select “Never” for the EFS frequency question were asked to estimate their food shopping percentages using four EFS modes over the past year: B2C, O2O-FDS, C&C, and O2O-IS, which are the common EFS modes in the UK and NZ. This measurement instrument was adapted from a previous study that examined consumers’ food shopping percentages across different locations (Vanhonacker et al., 2010). Participants who did not select “Never” for the EFS frequency question were also asked to indicate the top three food/meal categories they most often purchased using EFS in the past year. The measurement instrument and the list of 26 food categories were adapted from previous studies that explored consumers’ food purchase categories via EFS, as well as the actual food categories available on EFS platforms (Wang et al., 2020, 2024).

Participants’ general image of sustainability in EFS was measured using a seven-point scale, ranging from 1 = very negative to 7 = very positive, which was adapted from previous studies that explored consumers’ general image of various food categories (Almli et al., 2011). The measurement instrument was used to assess participants’ general image of EFS’s impact on sustainability as a general concept, as well as on the three sustainability dimensions: social, economic, and environmental sustainability.

Due to the absence of existing instruments on consumers’ perceptions of sustainability in EFS, an open-ended question was developed to gain a qualitative understanding of participants’ perceptions on the EFS impact on sustainability (Ritchie et al., 2013; Wang et al., 2020). After conducting a content analysis of the transcripts from the responses to this open-ended question, participants’ perceptual dimensions of sustainability in EFS were dummy coded and transformed into twenty-eight binary variables for the subsequent quantitative analysis.

Descriptive analyses were conducted to determine the mean values or percentages of participants’ EFS frequency, mode preferences, categorical preferences, and their general image of sustainability in EFS. Chi-square tests or independent sample t-tests were used to assess significant differences between the NZ and UK sub-samples, using IBM SPSS 29 (Caputo et al., 2018; Van Loo et al., 2015). Additionally, content analysis with NVivo 14 was performed to code transcripts of the open-ended responses regarding participants’ perceptions toward sustainability in EFS (Jaeger et al., 2023; Monticone et al., 2023; Pettigrew et al., 2024). A linear regression model was then estimated using STATA 17 to explore the relationship between consumers’ general image of sustainability in EFS, and their perceptual dimensions of sustainability in EFS, socio-demographics, and EFS frequency, in order to identify the significant factors that influence whether they have a positive or negative image of sustainability in EFS. The linear regression model was developed because the dependent variable, the general image of sustainability in EFS, is a continuous variable (Aertsens et al., 2011).

As shown in Table 3, 57.8% of the pooled sample had used EFS, while 42.2% of the participants had not used the service in the past year. A Chi-square test revealed significant differences in EFS usage frequency between participants from NZ and the UK A higher proportion of UK participants reported using EFS in the past year compared to their counterparts in NZ. This aligns with the findings of Wang et al. (2024), which suggest that NZ consumers have lower food consumption frequencies across B2C, O2O-FDS, and C&C services than consumers in the UK and Denmark. This may reflect the reality that Europe, particularly the UK, is a major player in the EFS market, with higher market size, revenue, growth, and penetration rates compared to Oceania, including NZ (Kingpin Market Research, 2024; Statista, 2024).

Regarding the percentages of EFS mode usage, O2O-FDS and C&C accounted for over 60% of the share among participants who used EFS in the past year, both in the pooled sample (n = 1,179) and in the sub-samples from NZ (n = 538) and the UK (n = 641). Independent sample t-tests revealed that NZ had a significantly higher percentage of participants using C&C and a lower percentage using O2O-FDS compared to the UK This could be attributed to the more developed EFS market and the much higher population density in the UK, which result in lower delivery costs and shorter delivery distances than in NZ (Kingpin Market Research, 2024; Statista, 2024; Wang et al., 2020, 2024; Worlddata, 2024). Wang et al. (2024) indicated that consumers living in highly urbanized regions tend to prefer O2O-FDS, whereas those residing in more rural areas find C&C to be more popular. This may further demonstrate the impact of EFS development levels and population density on consumers’ preferences for EFS modes. People living in regions with lower population density and less developed EFS markets tend to prefer C&C over O2O-FDS to save on delivery costs, unlike those in regions with higher population density and more developed EFS markets.

Figure 1 shows the most frequently purchased food categories by participants who used EFS in the past year. Fresh foods and restaurant meals dominate participants’ food choices with EFS in either the pooled sample or the two country samples, which is consistent with the fact that O2O modes accounted for over 80% of participants’ EFS usage. Sellers on O2O platforms are primarily local restaurants and food stores. As a result, consumers used O2O platforms as substitutes for traditional offline grocery shopping and restaurant meal ordering. Previous studies have also shown that O2O modes are better suited for selling restaurant meals and fresh food products compared to the B2C mode, which lacks physical stores in consumers’ locations and requires longer delivery distances (Tan et al., 2023; Wang et al., 2020, 2024). Additionally, independent sample t-tests revealed significant differences in EFS category preferences. NZ participants find restaurant meals, fresh fruits, meal kits, snacks, and fresh seafood more popular than UK participants do. Instead, UK participants find fresh meat, fresh vegetables, dairy products and meat substitutes to be more popular than their counterparts in NZ. This is the first consumer study to identify significant differences in EFS category preferences between countries.

Table 3 also presents participants’ general image of the EFS impact on sustainability as a general concept, as well as its impact on the three dimensions of sustainability. Overall, participants have a positive image of sustainability in EFS, as the mean values of the four general perception variables on the answer scales are all above 4, indicating an average response that is positive, in both the pooled and separate samples. There are no significant differences in this positive perception of sustainability in EFS between the NZ and the UK sub-samples.

Regarding the participants’ general image of the EFS impact on the three dimensions of sustainability, we find that economic sustainability has the most positive image, followed by the impact on environmental sustainability. This aligns with the findings of Li et al. (2024), which indicated that in O2O-FDS businesses economic sustainability is considered to be a more important criterion than social and environmental sustainability.

No previous study has explored consumers’ perceptions toward the EFS impact on sustainability sufficient for developing a codebook. So, the qualitative data from the open-ended question was processed using content analysis with an inductive coding approach, arising directly from the data itself (Jaeger et al., 2023). Furthermore, given the exploratory nature of the study, which aimed to produce new knowledge on the topic, the coding process was completed by a single coder, the first author, while all other project members peer-reviewed the coding structure and results to achieve a shared consensus (Monticone et al., 2023; Pettigrew et al., 2024). Additionally, through a two-round coding process, the transcripts were first coded into text-classes, which were then grouped into text-dimensions, indicating consumers’ perceptual dimensions regarding the EFS impact on sustainability (Guerrero et al., 2010). Finally, the codes for perceptual dimensions were grouped into the three sustainability dimensions based on previous literature regarding sustainability in EFS (Li et al., 2020, 2024; Prencipe et al., 2024; Siragusa and Tumino, 2022; Yue et al., 2024).

A data saturation test was conducted to determine if the sample size was sufficient or further sampling would be needed. We found that the likelihood of new information regarding consumers’ perceptual dimensions toward the EFS impact on sustainability from additional sampled units was very low (Wang et al., 2016). The test examined both time and location data saturation. For time data saturation, the pooled sample was classified into four sub-sample groups based on the order in which participants completed the survey: the first group (n = 500), the second group (n = 500), the third group (n = 500), and the last group (n = 539). The study analyzed the number of perceptual dimensions that appeared in these groups. All perceptual dimensions regarding the EFS impact on sustainability appeared in responses provided by participants in the first group, the initial 500 survey completions. Therefore, no new perceptual dimensions were found in subsequently sampled units from the three later groups, where participants completed the survey after those in the first group. According to location data saturation, the study examined the number of perceptual dimensions appearing in the sub-samples from the two countries: NZ (n = 1,019) and the UK (n = 1,020). All perceptual dimensions appeared in both country groups, indicating no new perceptual dimensions emerged across the two countries. Based on these findings, the study sample size is adequate to achieve data saturation, implying no new information would be obtained from additional sampled units for consumers’ perceptual dimensions toward the EFS impact on sustainability.

Figure 2 illustrates the codes of perceptual dimensions regarding the EFS impact on sustainability. Approximately 12% of the total participants (n = 2,039) provided invalid responses that were judged to be unrelated to the open-ended question. These participants were excluded from the subsequent analysis. Furthermore, around 22% of the total participants indicated that they were unaware or unsure about the topic. About 7.7% of the participants expressed a general opinion on the EFS impact on sustainability (positive, negative, significant, or no influence) but did not provide detailed opinions.

Around 34%, 26% and 15.9% of participants respectively shared specific opinions on the EFS impact on environmental, social and economic sustainability. Their specific opinions contributed nine perceptual dimensions each to the three sustainability dimensions. The perceptual dimensions for economic sustainability include “approaches to influence the economy”, “costs”, “food chain”, “local economy”, “physical stores”, “price”, “quality”, “small business”, and “sustainable consumption”. The perceptual dimensions for environmental sustainability encompass “carbon footprint”, “delivery and transport”, “food miles”, “GHG emissions”, “packaging”, “pollution”, “resource usage”, “waste”, and “influence environment generally”. The perceptual dimensions for social sustainability include “convenience”, “employment”, “food security”, “food variety”, “health”, “social interaction”, “specific groups”, “trustworthiness”, and “influence society generally”. These findings align with previous non-consumer studies, which indicate that the impacts of EFS on sustainability include increased employment opportunities but low job satisfaction, uncertain effects on the traditional catering industry, social relationships, and food waste, the emergence of new business models, more convenient urban living, higher business efficiency, lower GHG emissions, rising challenges and opportunities for public health and transportation, promoting the use of green and sustainable packaging and delivery methods, and an increased carbon footprint (Li et al., 2020, 2024; Prencipe et al., 2024; Siragusa and Tumino, 2022; Yue et al., 2024).

In terms of the economic sustainability, specifically in the perceptual dimension of “approaches to influence the economy”, more participants believed that EFS supported business and the economy (n = 68) than believed it harmed them (n = 12). Supporters indicated that EFS could benefit business and the economy by accelerating the circulation of goods, supporting specific sectors (e.g. delivery and packaging businesses, farmers, and large shops), boosting demand and business efficiency, encouraging sustainable production, expanding the market (e.g. to national and international markets), and saving storage space. On the other hand, other respondents argued that EFS could harm business and the economy by leading to monopolies and product shortages, reducing business profits and control, increasing unsustainable competition, and being inefficient for consumers compared to offline shopping.

In the perceptual dimension of “price”, some participants (n = 47) expressed that EFS has an impact on price increases (e.g. it being expensive, wasting money, and causing price rises), while others (n = 31) indicated a positive influence of EFS on prices (e.g. saving money and decreasing prices). Unfortunately, they did not provide more detailed opinions about the specific impact of EFS on food product or service prices. However, some clues can be found in the sub-codes of the perceptual dimension “cost”, where participants mentioned the cost of delivery or transportation (n = 30) and transaction costs (n = 6) (e.g. reduced retail service, overhead expenses, and paper usage during the transaction process). Overall, it seems that some participants were concerned that EFS could increase food prices due to the additional delivery costs, while others thought EFS would decrease food prices due to benefits, such as lowering transaction costs.

Regarding the EFS impact on economic sustainability, participants also had differing opinions on how EFS can influence “sustainable consumption”, “local economy”, “physical stores”, “small business”, and product “quality”. More participants (n = 40) believed it can boost sustainable consumption by encouraging precise purchasing and avoiding impulse buying, compared to those who felt it increases unsustainable consumption (n = 23) due to resulting in imprecise purchases and overconsumption. Furthermore, participants mainly considered EFS to be harmful to the “local economy” (n = 25), “physical stores” (n = 16), and “small businesses” (n = 13). Only a few participants felt that EFS had a positive influence, with some suggesting it could reshape or support small businesses (n = 2), and others believing it could support or gauge regional businesses and needs (n = 3). In addition, more participants (n = 12) mentioned product quality issues with EFS compared to a few (n = 4) who indicated the good quality of products purchased via EFS.

Several participants contributed to the perceptual dimension of the “food chain”. They mentioned that EFS impacts the food chain by enhancing direct-to-consumer (n = 5) and sustainable (n = 2) chain models. However, some participants also noted that it can make the supply chain longer (n = 1) and more complicated (n = 1), negatively affecting intermediate chain members (n = 1).

In terms of environmental sustainability, particularly regarding the most frequently mentioned perceptual dimension, “delivery and transport” (n = 359), many participants (n = 157) believed that EFS helps reduce the use of personal transportation. Additionally, a significant number of participants highlighted the impact of delivery methods used by EFS on sustainability. More participants (n = 109) identified unsustainable delivery methods (e.g. overuse of vehicles for delivery) compared to those (n = 46) who mentioned sustainable methods (e.g. using bikes or electric vehicles). A small number of participants (n = 7) were unsure whether EFS delivery methods were sustainable. Furthermore, some participants (n = 59) noted the general environmental impact of EFS delivery but did not provide further details.

In the perceptual dimension of “packaging”, more participants (n = 86) expressed concerns about packaging issues related to EFS, including increased packaging usage and the use of plastic and other unrecyclable packaging methods. A smaller number of participants (n = 40) considered EFS to have a positive impact on sustainability through packaging, such as reducing packaging usage, decreasing plastic packaging, and promoting the use of sustainable and recyclable materials (e.g. recyclable paper bags and reusable shipping cartons).

Regarding the EFS impact on environmental sustainability, participants expressed diverse opinions on how EFS can influence “waste”, “resource usage”, “carbon footprint”, “greenhouse gas (GHG) emissions”, “pollution”, and “food mile”. A number of participants indicated the negative impact of EFS on these factors, including an increase in waste (n = 57) (e.g. caused by over-ordering and increased packaging), increased resource usage (n = 25) (e.g. fossil fuels and energy), and the rise in carbon footprint (n = 18), pollution (n = 24), GHG emissions (n = 24) and food miles (e.g. mainly due to heavy delivery services). By contrast, others noted the positive impact of EFS on these factors, including a reduction in waste (n = 48) (e.g. through precise purchasing and reduced packaging), reduced resource usage (n = 30), and the decreased carbon footprint (n = 29), GHG emissions (n = 17), pollution (n = 13) and food miles (n = 1) (e.g. primarily due to the decreased use of personal transportation). Additionally, many participants acknowledged the overall impact of EFS on environment, either positive (n = 52) or negative (n = 29), but did not provide further details. Some also mentioned terms related to general environmental impact (n = 15), fuel usage (n = 1), energy usage (n = 3), carbon footprint (n = 6), and GHG emissions (n = 9) without providing further details.

In terms of social sustainability, participants unanimously agreed that EFS benefits our society in terms of “food security” (n = 42, e.g. enhancing food accessibility and availability) and “convenience” (n = 233, e.g. making people’s daily lives more convenient, easier, and time-saving). At the same time, they also unanimously believed that it poses a threat to our society by causing “trustworthiness” issues (n = 66, e.g. no physical product to inspect when ordering, untrustworthy production processes, freshness concerns, fraud, and expired food), reducing “social interaction” (n = 71), and being unfriendly to “specific groups”, particularly the elderly (n = 19).

Regarding the EFS impact on social sustainability, participants expressed differing opinions on how EFS can influence employment and people’s health. Several participants also acknowledged the overall impact of EFS on society (n = 17), either positive or negative, but did not provide further details. In terms of the dimension “employment”, some participants (n = 31) indicated that EFS provides more job opportunities, particularly in packaging and delivery. However, other participants (n = 25) suggested that EFS reduces employment opportunities in the traditional physical retail industry and due to increased automation in the EFS business model. Some participants (n = 12) also raised concerns about employment exploitation in the EFS industry. Regarding the dimension “health”, some participants (n = 23) expressed concerns about health issues caused by EFS, such as laziness and mental problems. On the other hand, some participants (n = 11) believed that EFS could help people lead healthier lives but did not provide further details on this topic.

Differing opinions were also expressed by respondents with regard to the perceptual dimension of “food variety”. Some subjects (n = 16) noted a limited variety of food with EFS, while others (n = 22) believed that EFS offers a wide variety. Others (n = 23) mentioned using EFS platforms to purchase specific categories of food (e.g. meat, vegetables, fresh food, organic food, green food, animal food and ingredients, functional food, and eco-food). Figure 1 shows that, when using EFS, respondents prefer fresh food over processed food, which is perceived as having a negative impact on health and the environment. In this case, EFS has the potential to become an effective tool for recommending sustainable food products, such as plant-based options.

Regarding the perceptual dimension of “influence society generally”, some participants (n = 9) pointed to the negative impact of EFS, while others (n = 5) noted its positive effects on society. A few participants (n = 3) mentioned that EFS affects society. However, all these participants did not provide further detailed discussions on these perspectives.

The study analyzed the qualitative data using a quantitative approach (Guerrero et al., 2010). A linear regression model was estimated to explain the variation of the participants’ general image of the EFS impact on sustainability by using their perceptual dimensions toward the EFS impact on sustainability, socio-demographics and EFS experiences. This exploratory approach aimed to identify significant factors influencing whether participants have a positive or negative image of sustainability in EFS. A linear regression model was developed in accordance with the continuous nature of the dependent variable, general image of the EFS impact on sustainability, as shown in Table 4 (Joshi et al., 2015; Laerd Statistics, 2024). Table 4 also outlines the independent variables, including participants’ EFS frequency, socio-demographic characteristics, and their perceptual dimensions of the EFS impact on sustainability. Categorical socio-demographic variables, including occupation and marital status, were dummy coded for the regression model (Wang et al., 2024). Additionally, twenty-eight binary variables representing perceptual dimensions were created by dummy coding participants’ frequencies of mentioning those perceptual dimensions in their responses to the open-ended question (1 = Yes, mentioned; 0 = No, did not mention, see section 3.3). Participants who provided invalid responses to the open-ended question were excluded from this regression analysis (n = 244). Consequently, the total sample size used in the linear regression analysis was 1,795. The linear regression model is estimated as follows:

where $yi$ is $participanti$’s general image of the EFS impact on sustainability (See Tables 2–4 for the measurement design and descriptive statistics of the dependent variable). $xik$ is a vector of k independent variables explaining the general image by $participanti$⁠, including perceptual dimensions toward the EFS impact on sustainability, socio-demographics and EFS experiences (See Tables 1–4 for the measurement design and descriptive statistics of the independent variables). $βk$ is a vector of coefficients for the k independent variables, $β0$ is a constant, and $εi$ represents the unobserved error term for $participanti$ (Grebitus and Van Loo, 2022; My et al., 2018; Wang et al., 2024).

Table 5 presents the findings from the linear regression analysis regarding the factors influencing the general image of the EFS impact on sustainability. The analysis shows that EFS experience has a significantly positive impact on this general image. In other words, participants with more experience using EFS are more likely to hold a positive image of its impact on sustainability. As such, H3 is supported. While there are no previous empirical studies specifically examining consumers’ general image of sustainability in relation to EFS, these findings align with earlier research on consumers’ general attitudes toward EFS, which have been shown to have a significantly positive relationship with their purchase behaviors related to EFS (Wang et al., 2020, 2024). Although the general image of sustainability in EFS and the attitudes toward EFS are distinct concepts, both relate to consumers’ affective stage in their decision-making process when choosing EFS (Almli et al., 2011; Verbeke, 2000). Consumers with more positive attitudes toward EFS are likely to have a more favorable image of its sustainability, and in turn, may be more inclined to use EFS.

In terms of perceptual dimensions, “sustainable consumption” and “convenience” have a significantly positive impact on the general image of the EFS impact on sustainability. The significant influence of “convenience” aligns with the views of Li et al. (2020), who noted that EFS contributed to building a more convenient urban living environment. However, the limited studies in the area of sustainability in EFS are predominantly non-consumer-based industry studies, not focusing primarily on the consumption side. To date, no findings from these studies have indicated the EFS impact on sustainable consumption. As such, this study is the first to contribute empirical knowledge showing that “sustainable consumption” drives consumers to form a positive image of sustainability in EFS by encouraging precise purchasing and helping to avoid impulse buying.

By contrast, the perceptual dimensions of “packaging”, “pollution”, “trustworthiness”, “specific group”, “social interaction”, and “local economy” have a significantly negative impact on the general image of the EFS impact on sustainability. In more specific terms, consumers have a negative image of sustainability in EFS because they are concerned that EFS may harm the local economy, increase packaging waste and promote unsustainable packaging methods, contribute to pollution, raise online trustworthiness issues, reduce social interaction, and be unfriendly to specific groups. These sustainability concerns in EFS are partially addressed in previous studies, particularly in relation to social relationships and sustainable packaging (Li et al., 2020, 2024; Prencipe et al., 2024; Siragusa and Tumino, 2022; Yue et al., 2024). The current study is the first to provide empirical evidence on the significantly negative impacts of pollution, trustworthiness, specific groups, and the local economy, as perceived by consumers in relation to sustainability in EFS. In general, H1 is partially supported.

Regarding socio-demographics, two occupational categories “full-time student” and “unemployed, retired, housewife/houseman, or on leave” have a significantly negative impact on the general image of the EFS impact on sustainability. In other words, consumers who belong to these occupational categories are more likely to have a negative image of sustainability in EFS. This aligns with the current study’s findings on the significantly negative effect of the perceptual dimension “specific group” on the general image of sustainability in EFS. Participants indicated that EFS is perceived as unfriendly to specific groups, particularly the elderly. The “unemployed, retired, housewife/houseman, or on leave” category includes the elderly, which may explain the negative impact of this occupational group on the general image. Additionally, Wang et al. (2020) notes that university students are infrequent users of EFS, as they primarily eat in canteens and rarely purchase food or ingredients for cooking for themselves or their families. As such, all these findings suggest that consumers in lower occupational categories—such as students, the unemployed, retirees, housewives/housemen, or those on leave—are specific vulnerable groups for EFS, leading them to have a negative image of its impact on sustainability.

In contrast, both “married” and “single” marital statuses have a significantly positive impact on the general image of the EFS impact on sustainability. It is relatively easy to understand why married individuals have a positive image of sustainability in EFS, as they are shown to be more frequent users of EFS, often buying food and ingredients for family meals, as indicated by previous studies (Wang et al., 2020, 2024). This aligns with the earlier finding that EFS experiences have a significantly positive impact on consumers’ general image of sustainability in EFS. However, this study is the first to recognize that single individuals also have a positive image of sustainability in EFS, despite being less frequent users, as confirmed by prior research (Wang et al., 2020, 2024). Further studies are needed to explore the reasons behind this finding.

Additionally, previous studies show that several socio-demographic factors, including country (e.g. NZ versus the UK), gender, age, financial situation, household size, residential location (rural/urban), and educational level, have significant influences on consumer adoption of EFS (Wang et al., 2020, 2024). However, the current study found that none of these socio-demographic factors have a significant impact on consumers’ general perception of sustainability in EFS. As such, H2 is partially supported.

This study has some limitations. First, as the first to present several novel findings (e.g. the positive perception of sustainability in EFS among single individuals), it is difficult to further discuss or compare these results due to the lack of existing relevant research. Future studies are recommended to explore and explain the underlying rationale. Second, the coding process was conducted by a single coder, which resulted in a limitation due to the absence of coding results from a second coder to calculate inter-coder reliability (e.g. Cohen’s Kappa). However, in this study, reliability was enhanced by involving other three experienced researchers with strong backgrounds in qualitative research and relevant subject areas, who reviewed and confirmed the coding results. Third, another limitation related to subjectivity stems from participants’ self-reported responses to the open-ended question. However, the large sample size and the use of a data saturation test help mitigate this limitation and support a comprehensive and generalizable understanding of consumers’ perceptions of sustainability in EFS based on the qualitative findings of this study.

The results of this study have academic implications as it is the first empirical research to systematically explore consumers’ perceptions and image of sustainability in EFS. Utilizing a large sample size, the study provides reliable and generalizable findings that can inform the design of future related studies. Additionally, the study employed an innovative approach, analyzing qualitative data quantitatively to identify the significant perceptions that influence consumers’ positive or negative image of sustainability in EFS. It contributes to a good example in food consumer research to identify information from the qualitative data and explain it with regression. It also serves as a valuable applied example for future research to adopt the same approach when exploring novel topics with limited prior material to guide research design.

The findings of this study also offer significant policy and managerial implications. EFS is rapidly growing worldwide, and local governments and food marketers can hardly afford to ignore this trend, as it will profoundly impact or reshape local food systems and businesses. The significant factors identified in the quantitative regression model, along with their relevant qualitative insights from the content analysis, can serve as valuable references when formulating policies and marketing strategies to promote or regulate the development of EFS. First, the positive impact of “sustainable consumption” and “convenience” on consumers’ general image of sustainability in EFS highlights an opportunity for marketers to emphasize these aspects. By encouraging mindful purchasing, reducing impulse buying, and fostering a more convenient urban living environment, EFS can be positioned as a tool for promoting sustainable consumption and lifestyles, aligning with broader sustainability goals. Additionally, the findings serve as a warning for local governments and food marketers in developing policies and marketing strategies related to EFS. The negative associations with EFS, including “packaging”, “pollution”, “trustworthiness”, “specific groups”, “social interaction”, and “local economy”, highlight critical areas requiring targeted interventions. Policymakers and marketers should prioritize developing eco-friendly packaging standards, supporting the e-commercialization of local food products to reduce EFS food miles, and providing specific EFS training programs for vulnerable groups, such as the elderly. They should also prioritize the implementation of robust food traceability systems to build trust in EFS, while addressing concerns about social isolation (e.g. promoting live-streaming e-commerce to enhance interaction between online buyers and sellers) and environmental challenges, such as pollution (e.g. recommending sustainable transportation options, such as using electric vehicles and bicycles for delivery). All relevant policies, marketing strategies and actions will be essential in ensuring that the growth of EFS maintains a sustainable pace.

This study was funded by the University of Waikato Strategic Research Fund and the Waikato Management School Contestable Research Fund. The manuscript was submitted to the journal for review and publication, and subsequently revised, while the first author was visiting the University of Tokyo during his sabbatical leave as a Japan Society for the Promotion of Science (JSPS) Invitational Fellow.