Rural Consumers’ Willingness to Pay for Online Grocery Delivery Services during COVID-19 Pandemic: Preliminary Evidence from South Africa

Abstract

This study examines the willingness of rural consumers to pay (WTP) for online grocery delivery services during COVID-19, as well as the factors affecting their WTP. To achieve this, the contingent valuation method was used to examine WTP, as it measures consumers’ WTP for services without having directly used or paid for the service. Given the ordinal ranking of the WTP variable, the ordered probit model was used to examine factors affecting WTP. The descriptive findings showed that most of the rural consumers were aware of online grocery delivery services and WTP for online grocery delivery services during COVID-19. However, there were mismatches between what rural consumers prefer and the services that are offered by the food retailers. The empirical results showed younger consumers, and those with higher levels of education, higher levels of income, and larger household size are more likely to be WTP amounts higher than the premium amount. This implies that the food retailers should target high-income earning young consumers with higher levels of education and larger household sizes. Overall, the findings provide information on consumers’ awareness, preferences, and WTP, which would assist the retailers in developing marketing, pricing, payment, and delivery strategies that are appropriate for the rural consumers.

1. Introduction

The lockdown restrictions, which were implemented in various forms across the globe to contain the spread of COVID-19, resulted in a surge in e-service innovation because of the high demand for online grocery services. Some of the notable restrictions that were imposed in the food retail sector include social distancing and a limit on the number of customers inside stores and at checkout points [1, 2]. These restrictions, together with consumers’ concerns about their health, triggered consumers to shift away from in-store shopping to online grocery shopping, which resulted in an explosion in demand for online grocery delivery services [2, 3, 4].

This propelled food retailers to develop their own online shopping services [5] or to expand their existing services to meet the demand. Thus, COVID-19 and associated restrictions have led to an upswing in online grocery shopping across the world [2, 6]. In South Africa, the major food retail stores have restructured their business model by introducing online grocery delivery services, which they had not previously offered [7], leading to a surge in online sale of grocery. Although the COVID-19 pandemic has led to an upswing in online grocery shopping, online grocery services have not been expanded to rural consumers due to long distances between despatch and delivery points, lack of technical skills, and limited access to internet and digital payment solutions [3, 8, 9, 10]. Rural consumers are defined herein, in line with Sanchez-Diaz et al. [9], as being marginalized consumers—consumers who were not considered by food retailers when designing online grocery delivery services and who do not have access to such services because of one or more of these factors.

However, this does not negate the fact that there are several benefits that could accrued to both the rural consumers and food retailers, if online grocery delivery services were to be extended to rural areas. From a consumer’s perspective, online grocery delivery services would be beneficial if the consumer could order a variety of food items at a lower price and for a lower delivery cost [11, 12, 13, 14, 15]. Thus, rural consumers would prefer online grocery shopping over in-store shopping if they could purchase the food items they want online, provided that the costs of doing so (delivery costs) are lower than the costs involved in purchasing in-store (traveling costs) are.

From the perspective of the retailers, e-service innovation is driven by profit maximization. In other words, retailers strive to increase sales by offering online grocery delivery services to maximize profit. This implies that, if retailers cannot deliver groceries or find a third-party (logistics) service provider that could deliver the groceries at profitable cost, they will not expand their services to rural consumers [9]. It is on this basis that the online grocery delivery services, as well as studies on online grocery delivery services, have been confined to urban areas rather than rural areas [4, 8, 16, 17].

This study departs from the aforementioned studies in that it focuses on whether rural consumers are willing to pay for online grocery delivery services and factors influencing their willingness to pay (WTP). Since online grocery delivery services are confined to urban areas, a contingent valuation method (CVM) is used to examine WTP, as it measures consumers’ WTP for services, without having directly used or paid for the services. As such, it is assumed that, while there are no available online grocery delivery services in rural areas, the rural consumers would prefer online grocery shopping, if the costs of doing so (delivery costs) are lower than the costs of traveling to physical stores are. Thus, there is a need to examine rural consumers’ WTP for online grocery delivery services.

After determining consumers’ WTP, the factors influencing their WTP are examined. These include socioeconomic and other factors (awareness, preference, and COVID-19-related factors). By so doing, this study provides food retailers with a better understanding of which class of rural consumers to target, should they expand their services to rural areas. Moreover, by focusing on awareness, this study informs retailers on whether rural consumers are aware of their online grocery delivery services and thus assists them to raise awareness about their services in rural areas. By focusing on preference, this study informs retailers on what rural consumers like, and thus, assists them to adjust, remodel, and expand their services to accommodate rural consumers. Lastly, by focusing on COVID-19, this study informs service providers on whether rural consumers’ WTP would remain after the subsiding of COVID-19—when lockdown restrictions are lifted, vaccines have been rolled out, and consumers are free from fears of COVID-19.

2. Aim and objectives

The aim of this study is to examine rural consumers’ willingness to pay for online grocery delivery services. The objectives are as follows:

1. to determine the willingness of rural consumers to pay for online grocery delivery services during COVID-19; and

2. to examine the factors influencing rural consumers to pay for online grocery delivery services during COVID-19.

The remainder of the sections in this study are arranged as follows. Section 3 provides an overview of the South African food retail stores and their associated grocery delivery services, together with a review of related literature. Section 4 focuses on the case study, which includes an overview of the study area, data collection, and sampling procedures. Section 5 discusses the findings, as derived from both the descriptive and empirical analyses. Section 6 provides key findings from this study and the implications thereof, as well as descriptions of the limitations of the study and of recommendations for future research.

3. Literature review

As this study encompasses online grocery delivery services, the review of literature focuses on an overview of grocery delivery services provided in South Africa. Thereafter, a review of previous studies on online grocery delivery services is conducted. Finally, a key summary of the review is presented in order to highlight a gap in the literature that this study is intending to fill.

3.1 An overview of the South African food retail stores and their associated grocery delivery services

South African e-commerce is booming because of a growing trend for using online shopping. For instance, e-commerce in South Africa grew by 66% in 2020 (during the start of the COVID-19 pandemic), relative to 2019 (pre-COVID-19). This growth is attributed to high mobile cellphone penetration, secure payment options, and a shift from shopping in brick and mortar stores to shopping online during the COVID-19 pandemic. However, prior to COVID-19, much of the growth was driven by online shopping for nonfood items such as electronics, clothing, and apparel. Moreover, while there is a notable growth, e-commerce in South Africa is still in the infancy stage, as compared with European markets [18].

According to a Deloitte’s study, more than 70% of South Africans are shopping online at least once a month [19]. The top five online stores in South Africa, in terms of net sales, are Takealot.com (US$ 602 m), Superbalist.com (US$ 85 m), Woolworths.co.za (US$ 57 m), mrp.com (US$ 33 m), and amazon.com (US$ 27 m) [20]. It is noted that these online stores, except Woolworths.com, which specializes in food, homeware, fashion, and beauty, are not food retail stores. This suggests that while South Africa has experienced an upswing in grocery shopping during COVID-19, groceries are not among the most-purchased items by online shoppers. This point is in line with Deloitte Africa’s 2020 Digital Commerce Survey, which ascertained that the most popular items among online shoppers in South Africa, in line with global online shopping preferences, were household appliances, electronics, health products, clothing, and footwear [21]. The report further revealed that the major reasons for online shopping among the respondents were convenience, COVID-19 concerns, and saving time [19]. The key point derived from this survey, in the context of this study, is that lockdown restrictions and fears over COVID-19 have triggered South Africans into shopping online, resulting in growth in online shopping.

A general growth was experienced in the food sector, overall, as a result of the growth in consumers’ demand for online grocery delivery services from online suppliers. In particular, some businesses have restructured their models to allow for the delivery of goods and services, which they had not previously offered. As a case in point, a gifting and floral distributor called NetFlorist restructured its business model during the early stages of COVID-19 to include the delivery of essential goods such as food items. Another case is Quench (an alcohol delivery app), which remodeled its business strategy by integrating the delivery of food items uplifted from Woolworths stores to customers. In the same way, Bottles, an alcoholic delivery app, has partnered with Pick n Pay to deliver groceries to Pick n Pay online shoppers [7]. Other prominent grocery delivery apps developed in the early stages of the COVID-19 pandemic in South Africa are as follows [7]:

• Sixty60, Checkers’ exclusive, 1-hour grocery delivery app;

• Bottles for the delivery of groceries to Pick n Pay online shoppers;

• Quench for Woolworths online shoppers;

• Zulzi for the delivery of groceries from various retail stores including Pick n Pay and Woolworths;

• OneCart for the delivery of groceries from Specials, Pick n Pay, Food Lover’s Market, and Woolworths;

• Mr D Food for the delivery of fast foods from various restaurants to customers; and

• Bolt Food (formerly Taxify) for the delivery of food from those retailers that do not have delivery partners.

It is noted that, as the pandemic advanced, some food retail stores restructured their business models, rebranded, and partnered with third-party delivery companies for the delivery of goods and services. For instance, Pick n Pay rebranded by relaunching its “Bottles” service, its on-demand grocery delivery app, as Pick n Pay Asap! in 2021. The retailer also partnered with Mr D, an online delivery platform, for delivery of groceries in 2022. Another case is Makro, which partnered with OneCart for the same-day delivery of groceries and liquor (when permitted) to customers.

These e-service innovations have resulted in a surge in online food sales and the growth of the grocery retail market in South Africa. The top three retailers that are at the forefront of this surge are Checkers, Pick n Pay, and Woolworths. The online grocery delivery services (APPs) offered by these retailers, as well as delivery details, are presented in Table 1.

	Checkers (Sixty60 App)	Woolworths (Woolies Dash App)	Pick n Pay (Pick n Pay Asap!)
Number of stores offering service	158 in all provinces	31 stores in 3 provinces	392 stores in 7 Provinces
Delivery fee	R35	R35	R35
Delivery distance	Subject to location	Within 5 km of store	Subject to location
Delivery time	Same day, 60 minutes	Subject to time slot	Same day, 60 minutes
Delivery slot	As selected on checkout, depending on store	Every day, 10:00–20:00	Mon to Fri: 09:00–19:00 Saturday: 09:00–17:00
Minimum items (value)	R100	None	R50
Item limit	35	30	35
Order tracking	Live tracking, WhatsApp updates	Personal shopper	Real-time tracking
Refund and substitution	Yes	Yes	Yes
Delivery times	1-hour slot during operating hours	1-hour slot during operating hours	1-hour slot during operating hours
Customer rewards integration	Yes	No	Yes

Table 1.

Selected online grocery delivery services by retailers.

Source: Adapted from Staff Writer [22].

Of these stores, Checkers’ Sixty60 is the largest online grocery delivery service, with 75% of online sales in the last quarter of 2021, followed by Pick n Pay Asap! at 13%, and Woolworth’s Woolies Dash at 12%. Moreover, Checkers’ Sixty60 has the highest presence, with 158 stores across the country that offer online grocery delivery services. Thus, Checkers has pioneered the online grocery delivery space in South Africa, having offered services prior to the advent of COVID-19, thus gaining the largest market share of the online grocery retail market.

Other prominent players in the online grocery market are Makro (wholesaler) and Takealot (e-commerce). Makro offers the same-day delivery of groceries and liquor, in partnership with OneCart. Through OneCart, customers can order more than 13,000 grocery and liquor products at a delivery fee of less than R100. On the other hand, Takealot, the largest e-commerce business in South Africa, charges R60 for delivery and offers free delivery for orders over R450, and there is no standard delivery time. Depending on the item and vendor, some items can take 3 days or more to source and then deliver.

In summation, the food retailers in South Africa have restructured their business models, rebranded, and partnered with third parties for the delivery of groceries to customers. Of the retailers, Checkers has the largest market share of the online grocery retail market in South Africa because of its strong presence across South Africa. However, as in other countries, online grocery shopping in South Africa is confined to urban consumers. As such, the next section focuses on a review of related studies, with a focus on urban consumers.

3.2 Review of related literature

Previous studies on online grocery shopping can be summarized into four groups of studies, as follows. The first group of studies encompasses studies on grocery shopping behavior [2, 4, 8, 23]. The main findings from these studies are that, with the advent of COVID-19, there was a shift in consumer shopping behavior (panic buying) [4], a strong upswing in online grocery shopping [8], a surge in online sales [2, 23], and a shift from in-store shopping to online shopping [24].

The second group of studies are those that used theories of adoption to identify factors that influenced the adoption of online grocery shopping. These studies identified perceived ease of use, time pressure, attitudes, social norms, relative advantage, and perceived risks as potential factors that influenced the adoption and usage of online grocery shopping in various countries [17, 18, 25, 26, 27, 28, 29].

The third group of studies include studies on factors that influenced the adoption of online grocery delivery services, specifically during COVID-19 [4, 8, 30]. These studies found that fear of COVID-19 and concern over ones’ health, fear of being infected, age, income, and gender were more likely to influence the adoption of online grocery delivery services.

The fourth group of studies involve those that predicted the future of online grocery delivery services. In these studies, it is predicted that online grocery shopping will continue after COVID-19 [4], as grocery shopping is a recurring and habitual process, which cannot be changed easily [31]. Moreover, a hybrid-shopping channel, involving both online and in-store shopping, could emerge post-COVID-19, rather than a split in choice between online shopping or in-store shopping [4].

The fifth group of studies are studies on consumers’ WTP for online grocery delivery services [32, 33]. This includes a study, which found that “perceived inconvenience of shopping groceries in stores” had a positive influence on consumers’ WTP for home delivery and use of home delivery of groceries in Europe [32]. Another study conducted before COVID-19 found that French consumers’ WTP for e-grocery was low, though they would adopt e-grocery if home delivery services were to be provided [33]. The results further showed that majority of the consumers reported that they will not pay more for the delivery of grocery items. This is because the negative factors associated with in-store shopping did not increase consumers’ acceptance of delivery fees (i.e. driving distance to stores, the time it takes to shop in-stores, and bad feeling associated with in-store shopping). Overall, the findings from these studies suggest that the willingness of consumers to pay for online grocery delivery services was low (before COVID-19 pandemic).

In contrast, studies conducted during COVID-19 showed that consumers were WTP for online grocery delivery services. This includes a study which examined consumers’ WTP for e-grocery in Milan and Rome during COVID-19 [34]. The results showed that those adopters of e-grocery (i.e. those who have already purchased grocery online more than once) in Milan were WTP for e-grocery more than others. In terms of knowledge about e-grocery, the study found that there was no difference in WTP between consumers with and without knowledge about e-grocery in both Rome and Milan. In terms of frequency of buying, it was found that the lowest frequent buyers in Rome would pay more for e-grocery than the others would.

Concerning gender, the study found that females in Rome were more willing to pay for both the enlargement of the product range and the reduction of the lead time, while those in Milan were WTP less. It was further found that, in both Milan and Rome, full-time workers were WTP two times more than the unemployed for a wider product range, while high-income earners would spend three times more on a wider product range than the low-income earners. The study further found that in both cities, middle-aged consumers were WTP two times more than the over sixties for the preparation of orders [34].

Another study which was conducted during COVID-19 found that consumers in China had higher WTP for vegetables and meat [35]. Various factors, such as anticipated duration of COVID-19, direct exposure to those infected with COVID-19, online shopping shares, income, and gender, had a positive influence on their WTP higher prices for food. The results further attributed the higher WTP to panic about the future and concerns over the increased cost of the food supply. Moreover, higher-income consumers were WTP more than their counterparts.

Overall, studies conducted before COVID-19 suggest that the willingness of consumers to pay for online grocery delivery services was low. In contrast, those conducted during COVID-19 showed that consumers were WTP for food and online grocery delivery services. The main gap in literature, derived from the reviewed studies, is that there are no empirical studies on rural consumers’ WTP for online grocery delivery services and factors influencing their WTP, hence, this study. In particular, based on the reviewed studies, there are no studies that used a contingent valuation method to examine rural consumers’ WTP for online grocery services, without having directly used or paid for the services.

However, while some studies focused on WTP for online grocery delivery services, they were conducted before COVID-19. As such, the findings of these studies cannot provide an insight into the influence of COVID-19-related factors on consumer’s WTP for online grocery delivery services. In addition, while a study by Yue et al. [35] used CVM to examine WTP during COVID-19, it focused on WTP for higher food prices rather than WTP for online grocery delivery services. Accordingly, the findings of their study cannot provide information on rural consumers’ WTP for online grocery delivery services and factors affecting their WTP, hence, this study.

4. Research methods

4.1 Study area

A preliminary study on rural consumers’ willingness to pay was conducted in Ga-Mothapo village for several reasons, as follows. Ga-Mothapo village is located in the jurisdiction of Polokwane Local Municipality, 35, 7 km east of Polokwane (the capital city of Limpopo Province). The stores that are offering online grocery delivery services in Polokwane (i.e. the Mall of the North), at the time of conducting this study, are Checkers, Woolworths and Pick n Pay. These services are confined to consumers in the urban area of Polokwane, meaning they are not delivered to rural consumers in Ga-Mothapo village. This study focuses on the WTP for online grocery delivery services of those consumers who have not previously directly used or paid for such services; hence, Ga-Mothapo village was chosen as the case study area.

4.2 Data collection

The study was based on the collection of primary data, which were collected through face-to-face interviews in 2021, that is, during the COVID-19 pandemic. Face-to-face interviews were conducted to maximize the quality of data collected and to minimize the nonresponse rate. A simple random sampling procedure was used to select a sample size from the overall household population. Household numbers were written on slips of paper that were placed inside a box, from which 120 households were randomly selected and surveyed. The respondents were individuals responsible for grocery-purchasing decisions, as well as the cooking, preparing, and serving of food.

The questionnaire had three separate sections. The first section covered the socioeconomic characteristics of the respondents, while the second section covered online grocery shopping and delivery services. Details of other information related to awareness and preferences were embedded in the second section. The third section covered willingness to pay. The questionnaire was pretested, according to the guidelines published by GAO [36], to improve the reliability and validity of the data collected. In line with ethical standards, the respondents were asked for consent and given assurance that the information collected will be used only for the purpose of the research and will be treated with confidentiality.

5. Descriptive and empirical results

This section provides the results for the socioeconomic characteristics of rural consumers and other factors such as awareness, preference, and COVID-19-related factors. The descriptive results for the willingness of rural consumers to pay for online grocery delivery services are also embedded in this section. This is followed by the empirical findings for factors influencing rural consumers to pay for online grocery delivery services.

5.1 Descriptive results

5.1.1 Socioeconomic characteristics of rural consumers

Descriptive analyses were conducted to describe the socioeconomic and other characteristics of rural consumers. Accordingly, the first set of questions covered their socioeconomic characteristics. The descriptive results for the socioeconomic characteristics of rural consumers are presented in Table 2.

Variable	Outcome		Percentage (100%)
Categorical variables
Gender	Male		47.5
	Female		52.5
Age	Below 21 years		5
	Between 22 and 30 years		65
	Between 31 and 40 years		30
	Above 40 years		0
Educational level	No formal education		0
	Primary education		10
	Secondary education		50
	Tertiary education		40
Employment status	Employed		17.5
	Unemployed		57.5
	Self-employed		25
Income level	Less than R5 000		52.2
	Between R5 000 and R9 999		37.5
	Between R10 000 and R14 999		7.5
	Between R15 000 and R19 999		2.8
	Above R20 000		0
Source of income	Wages/salaries		23.7
	Self-employment		31.3
	Investment		8.8
	Grants		36.2
	Pension		0
Continuous variables
	Min	Max	Mean	Standard deviation
Household size	1	10	2	0.952
No. of employed people in a household	1	3	1	0.844

Table 2.

Socioeconomic characteristics of rural consumers.

The findings for gender showed that 52.5% of the respondents were female, while 47.5% were male, which signifies that most of the rural households in the case study were headed by females. In terms of age, the majority of the respondents were between the ages of 22 and 30 years (65%), while none of the respondents was above the age of 40 years. This signifies that most of the households in the case study were headed by young people. Concerning education levels, the majority of the respondents had attained the secondary education level (50%), while fewer respondents had attained only the primary education level (10%).

The findings regarding employment status indicated that most of the respondents were unemployed (57.5%), while fewer were employed (17.5%). The results are in line with the country’s unemployment rate of 34.9% experienced in the third quarter of 2021 [37]. In terms of the levels of income, most household heads in the case study earned an income of less than R5 000 (52.2%), while none of them earned an income of above R20 000.

With regard to sources of income, grants comprised the main source of income for most of the household heads (36.2%), while investment was the least in terms of source of income (8.8%). This finding is substantiated by the fact that social grants comprise the main source of income for most of South African consumers. The statistics for household size show that the average household size was 2, with a minimum of 1 person in a household and a maximum of 10 people. The average number of employed people in a household was 1, with a maximum of 3 and a minimum of 1.

5.1.2 Rural consumers’ awareness and preference

The second set of questions, following those on socioeconomic characteristics, was on online grocery shopping and delivery services. The aim was to find out about the awareness of the rural consumers regarding online grocery delivery services, as well as about their preferences. Given that the rural consumers in the case study had not used or paid for such services, their preferences were measured hypothetically—what would they prefer, if they were able to use and pay for the services. The descriptive results for awareness and preferences are described in Table 3.

Variable	Outcome	Percentage (100%)
Aware	Yes	70
	No	30
Source of information	TV	35
	Radio	7.5
	Newspaper	0
	Friends/family	15
	Social media	42.5
Method of ordering	Web browser	35
	Phone app	60
	Phone call	5
	Other	0
Frequency of ordering	Not often	15
	Once a month	60
	Two or more than two times a month	25
	Once a week	0
	Two or more than two times a week	0
	When I have money	0
Food items to order	Staple food	7.5
	Fruits and vegetables	17.5
	Canned foods	0
	Frozen foods	37.5
	Dairy	27.5
	Pasta	2.5
	Boxed foods	5
	Other	2.5
Time period to order	Beginning of the month	17.5
	Only 15^th	15
	Only 22^nd	27.5
	Only 25^th	27.5
	Month-end	12.5
Method of payment	Cash	60
	Debit card	10
	Credit card	30
	Electronic funds transfer (EFT)	0
	Mobile payment	0
	Direct deposit	0
	Others	0
Main reason for ordering	COVID-19	30
	Affordability	12.5
	Busy schedule	17.5
	Time-saving	22.5
	Distance	10.5
	Long queues	2.5
	Other reasons	5.0
Concern over COVID-19 and associated regulations	“Afraid to get infected”	46
	“Tired of long queues”	21
	“Tired of wearing masks”	21
	“Afraid to die due to COVID-19”	12

Table 3.

Descriptive statistics for awareness and preference.

Since this study focuses on consumers’ willingness to pay for services, in a case where they have never used or paid for those services, the respondents were first asked if they were aware of online grocery delivery services. The findings are that the majority of the rural consumers were aware of online grocery delivery services, while fewer were not aware. This finding is attributable to the fact that most of the respondents were young people (65%), with access to online information. Those who stated that they were aware were asked to state their source of information about online grocery delivery services. Most of them stated that social media was their source of information about online grocery delivery services (42.5%). These results are attributed to the strong media presence by stores that offer online grocery delivery services (i.e. Checkers, Woolworths and Pick n Pay).

After determining awareness, respondents were asked about their preferences, without having used or paid for the services. In terms of preferred method of ordering, the majority of the rural consumers stated that they would prefer to order through phone apps (60%), suggesting that phone apps are the most-preferred method of ordering among rural consumers. This finding is contradictory to the observation that online grocery services have not expanded to rural consumers because of their limited access to internet, among other things [8]. Concerning the frequency of ordering, most of them stated that they would order once a month, suggesting that the frequency of ordering by rural consumers would be low.

Consumers were further asked about the food items that they would order. Most of them stated that they would order frozen foods (37.5%), confirming the preference of most rural consumers for ordering frozen foods. In terms of the period of ordering (based on the popular paydays), most of the consumers indicated that they would prefer to order on the 22^nd (27.5%) or 25^th (27.5%). This signifies that most of the rural consumers would order once in a month, and on the 22^nd or 25^th of every month. In terms of the method of payment, most of the respondents stated that they would prefer to use cash to pay for online grocery delivery services (i.e. 60%), although this method is not used by stores for the payment of online grocery delivery services. This signifies that rural consumers in the study area have not yet embraced digital payment solutions. In other words, there appears to be a mismatch between what rural consumers prefer regarding the method of payment and the services that are offered by the retail stores. However, as of the date of writing this study, cash payment is only offered by Takealot, an e-commerce business, and not by the food retail stores (Checkers, Pick n Pay and Woolworths) or by Makro, a wholesaler.

Consumers were also asked about the main reason why they would order groceries online for delivery. Most of them stated “COVID-19” as being the main reason they would order groceries online for delivery, signifying that most rural households would order groceries online due to concerns over or fears of COVID-19. These findings are in line with those of previous studies [4, 23]. Respondents were further asked regarding what it is about COVID-19 that would compel them to order groceries online. The results entail that most would order groceries online for fear of being infected with COVID-19 (46.6%). This finding raises a question of whether rural consumers would still order groceries online after COVID-19 has subsided—when lockdown restrictions are lifted and consumers are vaccinated and free from fear of COVID-19.

5.1.3 Rural consumers’ preferred grocery stores and online delivery services

The third set of questions involved questions on the preferred grocery stores and delivery services to underline the store that rural consumers would like to order from and the delivery service they would like to use. The results are described in Table 4.

Delivery service	Outcome	Percentage (100%)
Preferred delivery service	Takealot	57.1
	Bolt Food	35.7
	Bottles, NKA Pick n Pay Asap!	0
	Zulzi	0
	Woolies Dash	0
	Sixty60	7.2
	Others	0
Preferred store	Shoprite	39.3
	Woolworths	22.4
	Pick n Pay	21.4
	Makro	10.6
	Spar	6.3
	Others	0
Preferred store’s online grocery delivery service	Bottles, NKA Pick n Pay Asap!	26.8
	Woolworths’ Woolies Dash	23.2
	Checker’s Sixty60	50

Table 4.

Preferred grocery delivery services.

Six delivery services were placed before respondents with an option to specify their preferred delivery service. This entailed prominent services by the food retail stores, wholesalers, distributors, and e-commerce. It is important to note that Checkers’ Sixty60 is the only online grocery delivery service by a food retail store that the consumers prefer (i.e. only 7.2% of the respondents). This is because, while the services by food retail stores are confined to urban areas rather than rural areas, Checkers’ Sixty60 is the most prominent and more established service among the stores’ online grocery delivery services. Hence, Checkers has the largest market share of online grocery retail market in South Africa [24]. The striking results are that majority of the rural consumers have identified Takealot, an e-commerce, as their most preferred delivery service (57.1%). This is alluded to the fact that Takealot is the most popular e-commerce that is offering delivery services across all provinces of South Africa, including rural areas.

Given this, food retail stores, those offering and not offering online delivery services (excluding Checkers), were placed before respondents for them to choose a store they would prefer to order from, if that store were to provide online grocery delivery services. The store that most of the rural consumers would order from is Shoprite (39.3%), a store that is not offering grocery delivery services, as of this writing. This finding is attributable to the fact that Shoprite is more accessible to rural households, as it has more stores around the country [38].

In light of this, the respondents were further asked, which of the food retail stores offering online grocery delivery services they would order from. Most of the consumers stated that they would order from Checkers (50%), followed by Pick n Pay (26.8%) and Woolworths (23.2%). This finding is in line with the earlier observation that, out of the services offered by the food retail stores, Checker’s Sixty60 is the online grocery delivery service that rural consumers prefer.

5.1.4 Rural consumers’ WTP for online grocery delivery services

After establishing preference, consumers were asked about their WTP for online grocery delivery services. Thus, the fourth set of questions covered rural consumers’ willingness to pay for online grocery delivery services. The results are presented in Table 5.

Variable	Outcome		Percentage
WTP	Yes		95
	No		5
	Min	Max	Mean	Standard deviation
WTP amount	10	50	24.5	10.11

Table 5.

Willingness to pay for online grocery delivery services.

The respondents were asked about their theoretical WTP for online grocery delivery services, although they had never used or paid for those services. The results are that the majority of the rural consumers (95%) are WTP for online grocery delivery services. However, the average amount that they are WTP (R24.50) is lower than the delivery fees charged by stores for the delivery of groceries (R35.00). This suggests that there is a mismatch between the amount that rural consumers are WTP for online grocery delivery services and the amount charged by stores.

The contingent valuation method (CMV) was used to determine the willingness of rural consumers to pay for online grocery delivery services (Objective 1). This method was used because it measures consumers’ WTP for services, without them having directly used or paid for the service in the past. In applying CMV, a payment card was used to present possible levels of WTP. To determine the levels of willingness, respondents were asked, “suppose your favorite grocery store has a delivery service that has a price premium, will you pay more for the delivery service?” The levels of WTP vary from those of previous studies in that they account for consumers who are not WTP for online grocery delivery services. Moreover, they account for whether consumers would be WTP amounts higher than the premium amount. The rationale is to validate whether online shopping induce consumers to pay more for the cost of food delivery services during COVID-19, as observed in China by Yue et al. [35]. The amount that retailers are charging for delivery was used as a premium amount. The results for the levels of WTP are set out in Table 6.

Question	Outcome	Percentage
Suppose your favorite store has a delivery service, how much would you be willing to pay?	Not willing to pay	0
	Willing to pay exactly R35	55
	Willing to pay extra 9% of R35	17.5
	Willing to pay extra 10–14% of R35	17.5
	Willing to pay extra 15–20% of R35	10
	Willing to pay more than 20% of R35	0

Table 6.

Classes of WTP.

In terms of the classes of WTP, most of the rural consumers (55%) are WTP exactly the same amount that the stores are charging in the urban areas for the delivery services (R35). Further to this, none of the consumers is WTP more than 20% of the amount charged by stores for delivery services. These results are in line with those of Nicholson and Snyder [39], who found that consumers are willing to pay lower prices. However, the results are contradictory to those of Yue et al. [35], who found that online shopping induced Chinese consumers to pay more for the cost of food delivery services during COVID-19.

5.2 Empirical results

Given the four classes of WTP and their ordinal ranking, the ordered probit model was used to examine factors influencing rural consumers to pay for online grocery delivery services (Objective 2). However, three classes of WTP were used in the analysis, as the fourth class of WTP had very low observations. The ordered probit model was chosen because it has been widely used to evaluate WTP [40, 41, 42]. The model is set up around a latent regression that begins with the following equation [42]:

$$Y\ast =X^{’}\beta +\epsilon$$E1

where Y* = WTP, X’ is a vector of predictor variables, β represents a vector of coefficients, and ε is the error term. Y* is unobserved, and what can be observed is as follows:

$$\begin{array}{l}y=0\mathit{\text{if}}\mathrm{y}\ast \le {\mathrm{\mu }}_1;y=1\mathit{\text{if}}0\le \mathrm{y}\ast <{\mathrm{\mu }}_2;y=2\mathit{\text{if}}\mathrm{\mu }1\le \mathrm{y}\ast \le {\mathrm{\mu }}_3;\dots ;y=j,\\ \mathit{\text{if}}{\mathrm{\mu }}_{\mathrm{j}-1}<\mathrm{y}\ast <{\mathrm{\mu }}_{\mathrm{j}},\text{and}{\mathrm{\mu }}_0=-\infty ,\\ {\mathrm{\mu }}_{\mathrm{m}}=\infty \end{array}$$E2

In Eq (2), the μ’s are unknown parameters to be calculated with β. The following probabilities are derived, after normalizing the mean and variance to 0 and 1.

$$\mathit{\text{Prob}}\left(y=0|X\right)=F\left(X^{\prime }\beta \right)$$

$$\mathit{\text{Prob}}\left(y=1|X\right)=F\left({\mu }_3–X^{\prime }\beta \right)-f\left(X^{’}\beta \right)$$

$$\mathit{\text{Prob}}\left(y=2|X\right)=F\left({\mu }_2–X^{\prime }\beta \right)-f\left({\mu }_3–X^{\prime }\beta \right)$$

$$\dots \mathit{\text{Prob}}\left(y=j|x\right)=1-F\left({\mu }_{j-1}–X^{\prime }\beta \right)$$E3

All the probabilities must be positive, and then the following condition will be established:

$$0<{\mu }_3<{\mu }_2<{\mathrm{\mu }}_1<..\dots <{\mu }_{j-1}$$

This model estimation aims to identify the relevant factors that explain consumers’ WTP for online grocery delivery services. The final model is specified as follows:

$${\mathrm{WTP}}^{\ast }={\mathrm{\beta }}_0+{\mathrm{\beta }}_1{X}_1+{\mathrm{\beta }}_2{X}_2+{\mathrm{\beta }}_3{X}_3+{\mathrm{\beta }}_4{X}_4+..\dots +{\mathrm{\beta }}_{20}{X}_{20}+\epsilon$$E4

where: WTP* = WTP for online grocery delivery services; β₀, β₁, β₂, β₃, β_3, … , β₂₀ = parameters to be estimated; X₁, X₂, X_3, …, X₂₀ = predictor variables (i.e. socio-economic variables, awareness and preference variables); and ε = Disturbance term.

5.2.1 Model robustness

The empirical analyses were conducted using the Statistical Packaging for the Social Sciences (SPSS). The (-2) log-likelihood of the estimated model is 223.101, which implies that the model can be relied upon to predict WTP. A Nagelkerke pseudo-R² of 0.67 was obtained, which signifies that the predictor variables account for 67% of the variation in WTP. Ten variables of the 20 predictor variables were found to be significant. The results are presented in Table 7.

Variable	Coefficient	Standard error	T-statistics
Socio-economic characteristics
X1	0.554	1.315	0.421
X2	−0.662	0.495	1.783*
X3	0.930	0.481	2.035**
X4	0.928	0.354	2.621***
X5	−0.100	0.324	0.308
X6	−0.547	0.638	0.857
X7	−0.394	0.340	1.158
X8	0.576	0.2430	2.370**
Awareness variables
X9	0.719	0.367	1.959*
X10	0.063	0.205	0.307
Preference variables
X11	−1.870	0.484	3.863***
X12	0.813	0.299	2.719***
X13	0.186	0.2498	0.746
X14	−0.251	0.598	0.419
X15	0.286	0.534	0.535
Preferred store and grocery delivery services
X16	0.520	0.238	2.184**
X17	0.777	0.404	1.923*
X18	−0.835	0.589	1.417
Reasons for ordering
X19	0.189	0.410	0.460
X20	0.385	0.318	1.210
Model summary
(−2) Log-likelihood		223.101
Pseudo R-square
Cox and Snell R-square		0.64
Nagelkerke R-square		0.67

Table 7.

Ordered probit model results.

^*

p < 0.1.

^**

p < 0.05.

^***

p < 0.01.

Source: Research Data (2021).

5.2.2 Interpretation and discussion of the results

The ordered probit model was also used to derive the predicted probabilities and marginal effects for each level of WTP. The results for the estimated coefficients and marginal effects are discussed concurrently. The results for the predicted probabilities, as well as the marginal effects, are presented in Table 8. As per default, the values for the predicted probabilities for the levels of WTP sum to 1, while the values for the marginal effects for the three levels of WTP are equal to 0.

WTP levels	WTP = 1	WTP = 2	WTP = 3
Predicted probabilities	0.40	0.30	0.30
Variables	Marginal effects
Age of the consumer (X2)	0.472	−0.264	−0.208
Level of education (X3)	−0.091	0.047	0.044
Level of income of the consumers (X4)	−0.076	0.042	0.030
Household size (X8)	−0.054	0.032	0.022
Awareness (X9)	−0.018	0.008	0.010
Method of ordering (X11)	0.082	−0.044	−0.038
Frequency of ordering (X12)	−0.087	0.046	0.041
Method of payment (X15)	−0.0013	0.009	0.004
Preferred delivery service (X16)	−0.001	−0.002	0.003
Preferred store variable (X17)	−0.001	−0.002	0.003

Table 8.

Results for predicted probabilities and marginal effects.

The age of the consumer (X₂) had a negative influence on the WTP for online grocery delivery services. This suggests that older consumers are less likely to be willing to pay for online grocery delivery services, relative to younger consumers. The marginal effects for age were positive for the first level of WTP (WTP to pay exactly R35), but negative for the rest of the levels (i.e. willing to pay extra 9% of R35 and willing to pay extra 10–14% of R35). The implication is that older consumers are less likely to be willing to pay an amount greater than the premium amount, compared with younger consumers. These results are similar to those of Eriksson and Stenius [4], who found that older consumers are less likely to be adopters of online grocery shopping.

The level of education of the consumer (X₃) had a positive influence over the WTP for online grocery delivery service, meaning that the higher the level of education achieved is, the more likely the consumer would be WTP for online grocery delivery services. The marginal effects for the level of education were negative for the first level of WTP, but positive for the other two levels. This implies that rural consumers with higher levels of education are more likely to be willing to pay amounts higher the premium amount, relative to those with lower level of education. These findings are in line with those of a previous study, which found that higher education level increases the likelihood of doing online grocery shopping [5].

The level of income of the consumer (X₄) had a positive influence over the WTP for online grocery delivery services. This suggests that rural consumers who earn higher incomes are more likely to be willing to pay for online grocery delivery services than those earning lower incomes would be. The marginal effects for the level of income were negative for the first level of WTP, with positive effects for the remaining two levels. This signifies that rural consumers with higher levels of income are more likely to be willing to pay amounts higher than the premium amount, relative to those with lower levels of income. These results are in line with those of Eriksson and Stenius [4], who found that the adoption of online grocery shopping was influenced by income.

Household size (X₈) had a positive influence over the willingness of rural consumers to pay for online grocery delivery services. This signifies that the larger the size of the household is, the more likely it would be that a consumer in that household would be willing to pay for online grocery delivery services. The marginal effects for household size were negative for the first level of WTP, but positive for the other two levels. The implication is that rural consumers with larger households are more likely to be willing to pay amounts higher than the premium amount, relative to those with smaller households. These finding are in line with those of Eriksson and Stenius [4], who found that the higher the household size is, the more likely the adoption of online grocery shopping would be.

The variable, awareness about online delivery services (X₉), had a positive influence on rural consumers’ WTP for online grocery delivery services. This denotes that rural consumers who are aware about online grocery delivery services are more likely to be willing to pay for online grocery delivery services. The marginal effects for awareness were negative for the first level of WTP and positive for the remaining levels. This implies that rural consumers who are aware of online grocery delivery services are more likely to be willing to pay amounts higher than the premium amount, relative to those who are unaware. This is contradictory to findings of a previous study, which showed that there was no difference in WTP between consumers who were aware and unaware of e-grocery in Rome and Milan [34].

The variable for the method of ordering (X₁₁) that a consumer would prefer to use had a negative influence on rural consumers’ WTP for online grocery delivery services. This means that rural consumers who prefer to order groceries via phone apps are less likely to be willing to pay for online grocery delivery services. The marginal effects for method of ordering were positive for the first level of WTP and negative for the last two levels. The implication is that rural consumers who prefer to order via phone apps are less likely to be willing to pay amounts higher than the premium amount, relative to those who prefer other methods (phone calls and the web). This finding asserts the observation that rural consumers have not yet embraced digital innovation [3, 8, 9, 10]. Hence, those preferring to order grocery via phone apps are less likely to be willing to pay for online grocery delivery services.

The variable for frequency of ordering (X₁₂) had a positive influence on a rural consumer’s WTP for online grocery delivery services, suggesting that the higher the frequency of ordering is, the more likely it is that the consumer would be WTP for online grocery delivery services. The marginal effects for the frequency of ordering were negative for the first level, but positive for the other two levels of WTP. This implies that rural consumers who would order grocery more frequently are more likely to be willing to pay amounts higher than the premium amount, relative to those who would order less frequently. This is contradictory to the finding that the lowest frequent buyers in Rome would pay more for e-grocery than the others would [34].

The variable for method of payment (X₁₅) had a positive influence on rural consumers’ WTP for online grocery delivery services. This means that rural consumers who would pay for online grocery delivery services in cash are more likely to be willing to pay for online grocery delivery services. The marginal effects for the method of payment were negative for the first level of WTP and positive for the other two levels. The implication is that rural consumers who would pay using cash are more likely to be willing to pay amounts higher than the premium amount, relative to those who would use other payment methods (debit and credit cards). These findings are attributed to the earlier observation that rural households have not yet embraced digital payment solutions.

The variable for preferred delivery service (X₁₆) had a positive influence on the willingness of rural consumers to pay for online grocery delivery services. In other words, rural consumers who prefer to use Takealot are more likely to be willing to pay for online grocery delivery services, as compared with those who prefer to use other delivery services. The marginal effects for the preferred delivery service were negative for the first level, but positive for the other levels of WTP. Thus, rural consumers who prefer Takealot are more likely to be willing to pay amounts higher than the premium amount, relative to those who prefer other online grocery delivery services (Bolt Food, Sixty60, Woolies Dash, Pick n Pay Asap! and others). These results are in line with the observation that many consumers in the United States (US) purchased groceries through an online platforms such as Walmart, the second largest e-commerce retailer in the US, during COVID-19 pandemic [43].

The preferred store variable (X₁₇) had a positive influence on rural consumers’ WTP for online grocery delivery services. This suggests that rural consumers who would like to order from Shoprite, a store that does not offer online grocery delivery services, are more likely to be willing to pay for online grocery delivery services. The marginal effects for the preferred store were negative for the first level and positive for the other two levels of WTP. The implication is that rural consumers who prefer Shoprite are more likely to be willing to pay amounts higher than the premium amount, relative to those who prefer other stores (Shoprite, Woolworths, Pick n Pay, Makro, Spar, and others). This finding is attributable to the fact that Shoprite is more accessible to rural households, through its strong presence in rural towns [38].

Surprisingly, the variables regarding COVID-19 had no significant influence on rural consumers’ WTP for online grocery delivery services. This casts a doubt on whether rural consumers would be willing to order groceries online after COVID-19 has subsided – when lockdown restrictions are lifted, and consumers are vaccinated and free from fears of COVID-19. These results are contrary to the predictions that online grocery shopping would continue after COVID-19 [4], as grocery shopping is a recurring and habitual process, which cannot be changed easily [31].

6. Conclusion

COVID-19 restrictions have compelled certain food retail stores in South Africa to rebrand and restructure their business models and to partner with third parties for providing online grocery delivery services. However, as in other countries, online grocery delivery services are confined to urban consumers. As such, this study examined the willingness to pay (WTP) of rural consumers for online grocery delivery services and the factors affecting their WTP. To achieve this, the contingent valuation method was used to examine WTP, as it measures rural consumers’ WTP for services, without them ever having directly used or paid for those services. The findings provide information on rural consumers’ awareness, preferences, and WTP, which would assist the retailers in developing marketing, pricing, payment, and delivery strategies that are appropriate for the rural population.

In terms of awareness, the majority of the rural consumers were aware of online grocery delivery services (70%). Therefore, the food retailers who are intending to expand their services to rural areas should further raise awareness about their services in order to reach those rural consumers who are unaware of their online delivery services.

Concerning preference, there were mismatches between what rural consumers prefer and the services that retailers are providing. For instance, most rural consumers have identified Takealot, an e-commerce business, as their most preferred delivery service (57.1%), rather than retailers’ online delivery services, due to the popularity of Takealot in South Africa, including in rural areas. Moreover, most of the rural consumers stated that their preferred method of payment is cash, which is one of the payment methods accepted by Takealot, and not by the food retail stores (Checkers, Pick n Pay and Woolworths), or by the wholesaler, Makro. Therefore, the food retail stores should consider partnering with Takealot for the delivery of groceries to rural consumers and accepting cash payment.

It is acknowledged that the retailers would expand their services to rural consumers if it were profitable for them to do so. The findings showed that rural consumers who prefer Takealot are more likely to be willing to pay amounts higher than the premium amounts (i.e. more than the R35.00 that the retailers are charging for delivery). This shows that it would be economically viable for retailers to expand their services to rural consumers through partnership with Takealot.

Another notable mismatch relates to the store that most consumers would like to order from—the majority would like to order from Shoprite, although it does not offer grocery delivery services, as of this writing. Given this, the Shoprite Group should consider remodeling its business model by extending its sister company’s delivery services (i.e. Checkers’ Sixty60) to rural consumers. It would be viable to do so, as the empirical results showed that rural consumers who would like to order from Shoprite are more likely to be WTP amounts higher than the premium amount (i.e. more than R35.00 that the retailers are charging for delivery).

Another mismatch is in terms of WTP. In particular, while most of the rural consumers were WTP for online grocery delivery services (95%), the average amount that they were WTP (R24.50) was less than the amount charged by retailers (R35.00). Thus, the food retailers who are intending to expand their services to rural areas should consider reviewing their pricing strategies to suit rural consumers.

After determining the rural consumers’ WTP, the factors influencing their WTP were examined. The results provide food retailers with a better understanding of which class of rural consumers to target, should they expand their services to the rural areas. More specifically, the retailers should target high-income earning young consumers, with higher level of education and larger household sizes.

It is worth noting that the COVID-19-related variables had no significant influence on WTP. This casts a doubt on whether rural consumers would order groceries online after COVID-19 has subsided—when lockdown restrictions are lifted and consumers are vaccinated and free from fears of COVID-19.

The study was conducted with certain delimitations, which are explained in terms of the four areas for future research. The first area involves extending the research to other rural areas across South Africa as to enable a generalization of the results. The second area requires the inclusion of other online grocery delivery services, which were unaccounted for in this study. The third area involves examining WTP according to the individual lockdown levels experienced (i.e. Alert Level 5 through to Alert Level 1), as WTP could have changed as the lockdown restrictions were relaxed, as we moved down through to the lower levels of lockdown. The fourth area requires the use of valuation methods, which capture consumers’ actual behavior rather than the CVM, which captures hypothetical behavior [44]. In other words, methods that captures actual WTP instead of what consumers would pay, if they were to use or pay for the services.