Sweetening the Deal: The Ingredients that Drive Consumer Adoption of Online Grocery Shopping

ABSTRACT

Over the past decades, consumer adoption of online grocery shopping has increased steadily. Yet, overall market share is still comparatively low and retailers start questioning the prospects of the maturing distribution channel. The existing landscape of online grocery channels has seen little innovation nor diversity in terms of business models, reflecting the prevailing assumption that consumer online grocery shopping behavior is largely homogeneous. The present research challenges this notion by updating the understanding of consumer online grocery shopping behavior in a large-scale, representative study of Danish consumers. The results reveal distinct segments of online grocery adopters, which differ in their importance placed on perceived benefits of online grocery shopping. These segments can be targeted based on differences in preference for price, convenience, and service. The findings imply potential for retailers’ differentiation in the market of online grocery shopping.

KEYWORDS:

• Online grocery shopping
• innovation adoption
• consumer behavior

Introduction

The adoption of online grocery shopping, while increasingly steadily, still lags behind other online retailing domains (Melis, Campo, Lamey, & Breugelmans, 2016). Practitioners, who were overly confident about the growth of online grocery, over the past two decades have faced a reality in which consumers are slower than expected in adopting online as a distribution channel for buying groceries. The grocery retail market is different from others as margins are small, competition and concentration is high, all while product differentiation is low (Bolton & Shankar, 2018; Bolton, Shankar, & Montoya, 2010). More, consumers’ grocery shopping behavior is primarily driven by habits and the satisfaction of utilitarian needs (Hoyer, MacInnis, & Pieters, 2010). As a consequence, grocery retailers have to be available conveniently and at all times, practically forcing them to maintain all sales channels such as online shopping despite its challenges and potential losses.

The typical online grocery retailing market, including the Danish market in which the present study was conducted, is represented by the major offline retail chains as well as a few highly specialized pure online retailers. Discounters, on the other hand, seem to be slower at adopting the online channel, despite the increase of discount formats in the offline market (Kumar, Anand, & Song, 2017). The strategy of discount formats to cater to the more price-sensitive instead of service-oriented consumers in the offline market (Cleeren, Verboven, Dekimpe, & Gielens, 2010) seems not to have been introduced to the online grocery market. This highlights the notion of retailers that consumers show a rather homogenous shopping behavior when it comes to shopping for groceries online. Previous research on consumer online grocery behavior, however, indicates that consumers have different motives, such as utilitarian or hedonic, when shopping online (Cervellon, Sylvie, & Ngobo, 2015; Souitaris & Balabanis, 2007). Yet, these studies were not based on a representative sample of consumers who in fact shop online for groceries and leave room for a more refined investigation of heterogeneous adoption of online grocery shopping as an innovation.

The present study aims to overcome this limitation by identifying predictors of online grocery shopping using a representative sample of online shoppers along with a more nuanced classification of consumers’ shopping behavior in comparison to previous research. The results from our online survey show that about one quarter of the population have previously purchased groceries online and suggest major potential for the better differentiation in the grocery retail landscape online. As such, the data reveal three segments of online grocery adopters that can be individually targeted based on differences in consumer preference for price, convenience and service. Based on these insights, online retailers are advised to focus on (a) the price and assortment, (b) the mere convenience, or (c) the service to ease transition between online and offline purchases at the same chain. In addition, to attract more consumers by their online channel, retailers should focus on the perceived social norm, for example, through social media campaigns and continue to highlight the compatibility of online and offline shopping.

Theoretical background

Innovation adoption

The introduction of online shopping as a means to conveniently interact with retailers from the comfort of your own home has eventually disrupted an entire industry but along its way highlighted the issues in consumers’ adoption of innovations (Hoffman, Novak, & Peralta, 1999). From a consumer’s perspective, the shift from shopping in physical stores to shopping online involves the acceptance of new and unfamiliar technology and requires consumers to overcome functional and psychological barriers (Grabner-Kraeuter, 2002). That is, while functional barriers (i.e. trialability and complexity) are linked with the perception of the magnitude of changes resulting from adopting an innovation, psychological barriers (i.e. norm and risk) refer to consumers inner conflicts in beliefs about an innovation (Joachim, Spieth, & Heidenreich, 2018). The innovation of online shopping has created uncertainty about previously known and often routinized behavior (Grabner-Kraeuter, 2002), and has resulted in increased levels of perceived risk (Miyazaki & Fernandez, 2001), skepticism, and distrust (Grabner-Kräuter & Kaluscha, 2003). Consumer trust serves as a mechanism to reduce the complexity of human conduct situations where consumers cope with uncertainty (Luhmann, 2000) and, as such, facilitates the adoption of technological innovations (Gefen, 2000; Gefen, Karahanna, & Straub, 2003). The construct of trust is closely related to perceived risk regarding the characteristics of the innovation (Kim, Ferrin, & Rao, 2008; Mitchell, 1999). The issues in the adoption of online shopping, however, are by no means static as consumers’ knowledge and familiarity tend to increase and with it factors affecting the adoption of innovations change (Waarts, van Everdingen, & van Hillegersberg, 2002). Accordingly, with more experience in the online domain, issues with consumers’ adoption of online grocery shopping can be expected to have been reduced. More than a decade ago, Hansen (2005) identified risk as a minor discriminating factor, but complexity as one of the major factors in online grocery adoption. In support of this finding, a recent study by Driediger and Bhatiasevi (2019) using the technology acceptance model (TAM; Davis, 1989) shows that perceived risk no longer predicts consumers’ adoption of online shopping in Thailand, giving supporting us in revisiting Hansen’s (2005) study and extending it by segmenting the market of online grocery shopping.

Online grocery shopping behavior

A fair amount of studies investigated consumers’ grocery purchase behavior in the last two decades. Geuens, Brengman, and S’Jegers (2003) found consumers to be nonaccepting of an online channel for their grocery shopping, which is in line with the slow adoption rates of this innovation. With increasing developments in the sector, most research finds that convenience and time saving aspects are the most relevant drivers for adopting the online channel while perceived risk is still the main barrier (Blitstein, Frentz, & Pitts, 2020; Handayani, Nurahmawati, Pinem, & Azzahro, 2020; Harris, Dall’Olmo Riley, Riley, & Hand, 2017; Mortimer, Hasan, Andrews, & Martin, 2016; Robinson, Dall’Olmo Riley, Rettie, & Rolls-Willson, 2007). Most research has been done with UK consumers, as this is the most mature market. Ramus and Asger Nielsen (2005), found virtually no differences between UK and Danish consumers on general perceptions of online grocery shopping, suggesting a homogenous consumption behavior online. With an increasing number of consumers adopting online grocery shopping, however, a more diverse customer base might emerge. Concerning heterogeneity in shopping behavior, early research on the topic identified those with time constraints and those with physical constraints to be the most likely to adopt online grocery shopping, where risk concerns played only a very minor role (Morganosky & Cude, 2000), which was studied in-depth by Elms, de Kervenoael, and Hallsworth (2016). This is in line with Hand et al. (2009) that big life changes trigger the initial adoption of this innovation with a higher demand for convenience or time-saving features. Later quantitative research distinguished between utilitarian and hedonic motives, where convenience was more important for utilitarian shoppers and ambiance more important for hedonic shoppers (Cervellon et al., 2015; Souitaris & Balabanis, 2007). There were some inconsistencies for the role of assortment, value for money and service between those studies. The role of price seems to be less important in the online grocery retailing environment (Benn, Webb, Chang, & Reidy, 2015; Melis, Campo, Breugelmans, & Lamey, 2015), but the results are mixed (Degeratu, Rangaswamy, & Wu, 2000). Chu, Arce-Urriza, Cebollada-Calvo, and Chintagunta (2010), for example, found price sensitivity to be moderated by shopping frequency, where those shopping more frequently online were also more price-sensitive. In line with that, experience with the online channel is consistently found to influence shopping behavior (Hansen, 2008; Mortimer et al., 2016). Research using purchase data inferred that consumers are loyal to their offline chain in the beginning of innovation adoption, but refocus their shopping priorities over time (Dawes & Nenycz-Thiel, 2014; Melis et al., 2015). With increasing experience, the importance of assortment increases (Melis et al., 2015), which is especially beneficial for niche products as can be observed in other markets as well (Richards & Rabinovich, 2018), but also private label brands (Arce-Urriza & Cebollada, 2012; Degeratu et al., 2000).

The overview of previous research shows that despite an increasing body of research in the area of online grocery shopping adoption, there still seem to be inconsistencies regarding consumer characteristics and the actual weights attached to different shopping behaviors. The purpose of this study is twofold. First, the study seeks to update the understanding of the factors that facilitate the innovation adoption of grocery shopping. Second, this study seeks to identify the importance of shopping characteristics and potentially identify consumer segments with a representative sample of shoppers in the Danish market.

Method

Participants

One-thousand-five-hundred-eighty Danish online shoppers (47.0% females; age: M = 32.51 years, SD = 9.18) were recruited through a professional panel and paid about 4 USD for their participation, which took approx 15 minutes. The data were collected as part of a larger, representative online study of Danish consumers using self-administered online questionnaires in 2016. Table 1 summarizes the characteristics for the two main segments of the sample: nonadapters and adopters of online grocery shopping. Adopters were those participants who purchased groceries online at least once per year. In other words, nonadopters were those who never purchased groceries online before.

Table 1. Sample characteristics

Demographic factor	Nonadopters (N = 1163)	%	Adopters (N = 417)	%
Gender^a
Female	622	53.6	179	43.2
Male	539	46.4	235	56.8
Age^a
18–24	151	13.0	70	16.8
25–34	169	14.5	100	24.0
35–44	174	15.0	110	26.4
45–54	225	19.3	44	10.6
55–64	234	20.1	50	12.0
65 or older	210	18.1	43	10.3
Education^a
7^th to 10^th class	116	10.1	55	13.4
Upper secondary school^c	149	13.0	56	13.7
Professional education	198	17.3	56	13.7
Short-term higher education (up to 2 years)	100	8.7	52	12.7
Middle higher education (from 2–4 years)	356	31.1	113	27.6
Long higher education (over 4 years)	215	18.8	69	16.8
PhD	9	0.8	9	2.2
Income^ab
< 0.5 times the average	113	11.2	49	12.9
0.5 times the average to average	194	13.9	97	25.6
Average	245	24.2	124	32.7
Average to 1.5 times the average	248	24.5	55	14.5
> 1.5 times the average	212	20.9	54	14.2
Children^a
0	770	67.8	208	50.7
1	160	14.1	78	19.0
2	152	13.4	92	22.4
3 or more	53	4.7	32	7.9

^aPearson’s chi-square test for differences across groups significant at the .001 level

^bAnnual household income relative to 344 847 DKK ~ 53.400 USD

^c3-year education, needed as entry requirement for university education

Procedure

Participants were informed about the purpose, benefits, and risks and discomforts of the study prior to their voluntary participation. They then proceeded with screening questions in regards to their online grocery shopping experience. The screening was followed by the main questionnaire which included self-reported multi-item measurement scales. Participants, who reported previous usage of online shopping, also answered ten additional questions about their beliefs about online grocery shopping. Last, all participants answered demographic questions, were debriefed and paid for their participation.

Measures

Participants’ adoption of online grocery shopping was measured by a self-report question assessing the frequency of shopping online. Scales for the assessment of consumers attitudes toward adopting online grocery shopping were adapted from (Hansen, 2005; Table A1, Appendix). These included perceived social norm (two-item, 5-point agreement scale, 1 – disagree a lot, 5 – agree a lot; CR = .78) perceived complexity (four-item, 7-point agreement scale; CR = .71), perceived compatibility (3-item, 5-point agreement scale; CR = .66), perceived relative advantage (three-item, 5-point agreement scale; CR = .67) and perceived risk (four-item, 7-point agreement scale; CR = .59). In addition, based on the reviewed literature, ten belief items about relevant online grocery shopping characteristics (time saving, independent opening hours, product assortment, correct delivery, fast delivery, price, personal service, choice of best before date, trust mark of retailer, brand of retail chain) were included and ranked in terms of their importance (rank 1 – very unimportant, rank 10 – very important). Demographic questions included age, gender, relative income (relative to the average annual household income of Danish households before tax = 344.847 DKK), education and number of children.

Analysis

As the original study was administered in Danish, all items were translated to English by two independent researchers prior to analysis of the data and reporting of the findings. Statistical analyses were performed in R 3.6.0 using the default stats package except when stated otherwise. First, scales were tested for reliability and measurement validity using the psych 1.8.12 package (Revelle, 2019) and lavaan 0.6–3 package (Rosseel et al., 2018) respectively. Next, composite scores of the scales were computed and differences between nonadopters (76%) and adopters of online grocery shopping tested using two-sided, independent sample t-tests as well as nonparametric tests of independence (Kruskal-Wallis tests) to account for the non-normal distribution of measurement scales. A staged, binary logistic regression analysis was used to predict the adoption of online grocery shopping by stepwise entering the five independent variables (Model 1) and demographic variables (Model 2). Last, a K-means cluster analysis was used to identify segments of online grocery shopping adopters that differ in terms of the importance of online shopping characteristics using the cluster 2.0.9 package (Maechler et al., 2019). Whenever possible, sample sizes, confidence intervals and effect sizes are reported in favor of p values (Wasserstein, Schirm, & Lazar, 2019). Hypotheses were considered confirmed when effect sizes were larger than .1 and confidence intervals excluded zero for continuous and one for dichotomous dependent variables (Cohen, 1992).

Results

Adoption of online shopping

Table 2 summarizes the independent measures of perceived social norm, complexity, compatibility, relative advantage and risk for the two segments of nonadopters and adopters. The results show large, statistically significant differences between nonadopters and adopters in terms of perceived social norm, t(1207) = 16.74, 95% CI: [1.00, 1.26], d = 1.03, perceived compatibility, t(1453) = 15.19, 95% CI: [.79, 1.02], d = .89, and perceived relative advantage t(1434) = 15.56, 95% CI: [.79, 1.02], d = .91. More, results show a small, statistically significant difference between non-adopters and adopters in perceived complexity, t(819) = −3.31, 95% CI: [−.30, −.08], d = .18. Perceived risk, on the other hand, shows no statistical significant difference between nonadopters and adopters, t(825) = −.59, 95% CI: [−.13, .07], d = .03.

Table 2. Independent measures for adopters and nonadopters

Independent measure	Nonadopters			Adopters
	M	SD	N	M	SD	N
Social norm^a	2.06	1.07	826	3.19	1.13	383
Complexity^a	3.29	1.04	1031	3.11	.93	405
Compatibility^a	2.37	1.03	1046	3.27	1.00	409
Rel. advantage^a	2.29	1.02	1028	3.20	.99	408
Risk	3.30	.95	1037	3.27	.85	409

^aDifference between non-adopters and adopters is statistically significant.

While we report the results of independent samples t-tests, it shall be noted that these are in line with non-parametric tests of independence (Kruskal-Wallis tests) used to account for the non-normal distribution of the measurement scales.

Table 3 shows the binary logistic regression models predicting consumer adoption of online grocery shopping (1 = adoption, 0 = nonadoption) as a function of the aforementioned independent measures (see Model 1), and the independent measures and demographic variables (see Model 2). The results of Model 1 suggest that perceived social norm, compatibility and relative advantage are the strongest predictors of consumers’ adoption of online grocery shopping with positive odd ratios between 1.33 and 1.69. In line with the results of the t-tests, perceived risk and, in addition to risk, perceived complexity are found not to be statistically significant as the estimated 95% confidence interval for the odd ratios includes one.

Table 3. Estimates for consumer adoption of online grocery shopping (N = 1580)

Independent variable	Model 1			Model 2
	B (SE)	OR	95% CI	B (SE)	OR	95% CI
Main effects
Social norm	.52 (.08)	1.69	1.44–1.99	.57 (.09)	1.77	1.48–2.11
Complexity	−.11 (.98)	.90	.74–1.09	−.17 (.11)	.84	.67–1.04
Compatibility	.37 (.10)	1.45	1.20–1.75	.43 (.11)	1.54	1.25–1.90
Rel. advantage	.28 (.10)	1.33	1.09–1.61	.29 (.11)	1.33	1.07–1.65
Risk	.03 (.11)	1.03	.84–1.27	−.02 (.12)	.98	.78–1.24
Control variables
Gender (1 = female, 0 = male)				−.66 (.16)	.52	.38-.70
Rel. Income				−.29 (.06)	.75	.66-.85
Model summary
Intercept	−3.70 (.42)	.02	.01-.06	−2.37 (.51)	.09	.03-.20
DF	5			7
Nagelkerke R^2	.54			.67

Model 2 only retains demographic variables that were significant; age, education, number of children, weight, height were excluded accordingly.

Model 2, which considerably improves the model fit over Model 1 (R²_Model2 > R²_Model1), finds that female consumers are less likely to adopt online grocery shopping, as are households with higher relative income. Notable is that the main effects of perceived complexity, relative advantage and risk are lower in Model 2 as gender and relative income account for their previously explained variance. Perceived social norm remains unchanged.

Online shopping customer segments

The segmentation analysis about online shopping adopters’ beliefs about the importance of characteristics of online grocery shopping resulted in a three segment solution at which the differences between the segments were maximized. The three segments differ in terms of all online grocery shopping characteristics (Table 4). Segment 1 represents “price-oriented” customers as it scores highest on price, product assortment, correct and fast delivery of the ordered goods. Segment 2 represents “time optimizers” and scores highest in time saving and independence of opening hours, scores second highest in price and lowest in choice of best before date. Segment 3 is found to represent “cautious” customers, who stand out in deeming personal service, the trust mark score of the retailer and the retail chain brand as most important.

Table 4. Customer online grocery shopping segments

Factor	Segment 1 “Price-oriented” (N = 123)	Segment 2 “Time optimizers” (N = 99)	Segment 3 “Cautious” (N = 90)
	M (SD)	M (SD)	M (SD)
Time saving	5.34 (2.59)	7.61 (2.19)	4.10 (2.69)
Indep. opening hours	3.45 (2.12)	7.84 (1.84)	4.43 (2.76)
Product assortment	7.01 (2.45)	4.70 (2.57)	5.90 (2.66)
Correct delivery	7.19 (2.20)	5.40 (2.48)	5.90 (2.97)
Fast delivery	6.76 (2.13)	5.97 (2.71)	5.33 (2.90)
Price	8.76 (1.51)	7.08 (2.51)	4.33 (2.34)
Personal service	3.39 (2.30)	4.95 (2.85)	5.48 (2.99)
Choice of bb-date	5.32 (2.39)	3.95 (2.45)	5.53 (2.64)
Trust mark	4.41 (2.46)	3.78 (2.57)	6.86 (2.47)
Retail chain	3.37 (2.29)	3.73 (2.40)	7.13 (2.72)

All values printed in bold are statistically different from the values of the other two segments at the p <.001 level.

The shopping frequency turned out not to differ significantly between segments, but the share of shoppers who indicate weekly online grocery shopping was highest in Segment 1 (27%), followed by Segment 2 (24%) and was lowest in Segment 3 (13%). Those who shop at least once per month online for groceries were represented slightly more in Segments 2 and 3 (31%) compared to Segment 1 (24%). About half of all consumers shop for groceries online less than once per month (S1: 50%, S2: 44%, S3: 56%). In terms of demographics, customers in Segment 1 (M_age = 28.29, SD = 13.35, N = 123) are slightly younger than those in Segments 2 (M_age = 33.33, SD = 15.84, N = 99) and 3 (M_age = 30.78, SD = 16.64, N = 90), F(2, 289) = 3.55, f = .14 [f_1.2 = .14, f_1.3 = .07, f_2.3 = .07]. More, customers in Segment 3 (M_weight = 66.81, SD = 28.82, N = 82) weigh less compared to Segment 1 (M_weight = 77.28, SD = 28.59, N = 116) and Segment 2 (M_weight = 72.15, SD = 24.42, N = 94), F(2, 289) = 3.55, f = .16 [f_1.2 = .08, f_1.3 = .16, f_2.3 = .08]. All other demographic variables including income, gender, education and children show no statistically significant difference between the three segments.

Discussion and conclusion

The findings of this research provide new insights into consumer adoption of online grocery shopping and their heterogeneous preferences across segments. Overall, adoption of online grocery shopping is increasing, which is also reflected in the market with more retailers offering an online channel for their consumers. In our representative Danish sample, we find that 24% have tried online grocery shopping before, compared to 15% in 2002 in the US (Hansen, 2005). We find that perceived social norm, compatibility, and relative advantage are the strongest predictors of consumer adoption of online grocery shopping. This is in line with previous research, which found time-saving aspects and convenience to be the main drivers for adoption (Harris et al., 2017; Mortimer et al., 2016; Roberts, Xu, & Mettos, 2003; Robinson et al., 2007). Perceived complexity as well as perceived risk did not predict online grocery adoption, contrary to what was previously observed (Hansen, 2005). This suggests that consumers are more familiar and more trusting toward online grocery shopping than a decade ago, which could be expected due to a larger share of purchases being conducted online in general. Still, despite this development, we find that women are less likely to adopt online grocery retailing. This finding is supported with early research in this domain, which shows that men tend to be more prone to adopting IT innovations (Chau & Lung Hui, 1998). In line with that Handayani et al. (2020) find that gender moderates consumer adoption of online grocery shopping. We deem our finding relevant in that women are still usually the main household responsible for grocery shopping (Cervellon et al., 2015) and therefore a relevant target group for the online grocery channel. Most previous research did not include demographics or did not find any effect for these measures concerning online grocery adoption (Souitaris & Balabanis, 2007). We can therefore extend the previous literature in that income negatively affected adoption of online grocery shopping. We urge further research to investigate the reasons for this. It might be that households with higher income, prefer grocery stores and markets, which carry products in a higher price tier, which are not represented online.

In terms of customer heterogeneity, we find three clusters, of almost equal size to be the best solution. Contrary to previous research, we find one segment, which ranks price as one of the most important aspects of online grocery shopping. We do not find that there is a relation with shopping frequency as Chu et al. (2010), but we find this segment to be younger than the other two segments. The second segment is clearly interested in the time-saving aspects of online grocery shopping as well as independence of opening hours. This is in line with previous research identifying convenience and time-saving aspects as most important drivers (Blitstein et al., 2020; Harris et al., 2017; Mortimer et al., 2016). The third segment is concerned with the chain of the supermarket and the online shops’ trust rating. There were less of the experienced shoppers present in this segment. This is in line with previous research, which suggests that beginners will most likely start out shopping online at their preferred offline chain and only later on move toward other online outlets (Dawes & Nenycz-Thiel, 2014; Melis et al., 2015).

Managerial implications

Our results have managerial relevance as we find heterogeneity in the sample that justifies more diversity on the online grocery shopping market, than is currently the case. As our segments are similar in size, they are all worth considering from a practitioners perspective. One third of our sample is price-oriented, also in the online shopping environment. This suggests that discount formats also can have an attractive position in the online channel. As these consumers are less concerned with the independence of opening hours, click-and-collect models, which are more cost-efficient for the retailers (Belavina, Girotra, & Kabra, 2016; Hackney, Grant, & Birtwistle, 2006) might be a viable option for these consumers – as long as pick-up is possible soon after ordering and the order is flawless.

The second segment are time-optimerers, who choose online grocery shopping to save time and to be independent of store opening times. For this segment, click-and-collect models will be less attractive as those are dependent on the store being open. Pure online retailers might see this group as their main target. As this segment is less price-concerned, they might be willing to pay a premium for the delivery service as long as shopping does not consume their scarce amount of time available.

The third segment is more cautious and inexperienced. While we did not find perceived risk to play a significant role on the aggregate level, this segment might find online grocery shopping somewhat risky as it is important for them that the online shop has a high trust rating. In addition, consumers in this segment tend to patronize the online chain that they know, so established offline retailers can gain from a well-functioning omni-channel strategy that eases consumers into the adoption of the new online format and with additional personal service, might also be able to retain those customers in the online sphere.

Limitations and further research

Despite the strengths of this study, it has limitations that warrant discussion as well as avenues for future research. First, we deal with about 25% missing data in cluster analysis of adopters online shopping factors, which can be attributed to the online setting of our survey. As this data collection was part of a larger project, some less involved participants might have been tired and opted out of answering questions, where it wasn’t required. However, due to our large sample size, we argue that our clusters are sufficiently large to overcome this issue. Second, this research is based on self-reported data and not actual behavior. Future research should therefore aim to validate the importance of the shown factors based on actual purchase behavior. It would be especially interesting to combine the data with demographic and psychographic measures to enhance the validity and at the same time improve the operationalization of the clusters. Lastly, our results find no statistically significant difference between the two groups of adopters and nonadopters of online grocery shopping in terms of perceived risk. This result seems unexpected as nonadopters of innovative technology typically differ in terms of trust and risk perceptions, which would suggest that our nonadopters could be more technology savvy than those of noninnovation adopters of other populations. While it should be noted that Denmark is one of the leading countries in Europe regarding digitalization, future research should look into this issue by collecting data on less accessible and less tech savvy members of the population.

Appendix

Table A1. Descriptive statistics and reliability of scale items

Item	Description	M	SD	N	CR
SNM1	My family members think it’s a good idea to buy food online	2.40	1.35	1117	.79
SNM2	Most of my friends and acquaintances think it’s a good idea to buy food online	2.53	1.28	1030
CPL1	Buying food online is complex because you can not see and feel the products.	3.81	1.20	1357	.70
CPL2	Online shopping is generally complex.	2.83	1.23	1225
CPL3	It’s very difficult to find the products you need when you buy food online.	2.99	1.21	1019
CPL4	It is difficult to order food online.	2.91	1.28	1102
CPT1	Online shopping is attractive in my daily life.	2.60	1.40	1365	.66
CPT2	Buying food online fits well with the way I usually buy food.	2.15	1.27	1344
CPT3r	Generally speaking, online shopping is problematic.	3.34	1.20	1210
RLD1	Buying food online saves a lot of time.	3.29	1.29	1244	.67
RLD2	I prefer to buy food online because I am independent of opening hours.	2.10	1.29	1335
RLD3	You can save a lot of money by buying food online.	2.48	1.27	1116
RSK1	Return and exchange options are not as good online as in offline supermarkets.	3.64	1.24	1044	.59
RSK2	A risk of buying food online is that you can get wrong products or products of lower quality.	3.65	1.17	1265
RSK3	The security of online payment is not good enough.	2.64	1.26	1316
RSK4	There are too many untrustworthy stores online.	3.37	1.16	1175