Online retailing has grown at an unprecedented level in recent years. Globally, online retail sales are expected to grow from $4.248 trillion in 2020 to $7.391 trillion in 2025 (eMarketer 2022:2). This is a 74% expected increase in online retail sales within a period of 5 years. The majority of global online retail sales are expected to be from China. In 2022, eMarketer (2022:6) expects China to contribute 50.2% towards global online retail sales. The incredible growth of online retail in China can be attributed to an effective and efficient delivery system, among other factors (Hongfei 2017:15; Zandi, Torabi & Mohammad 2021:11). Despite the existence of a good delivery system in China, customer complaints have been on the increase, necessitating the establishment of an Internet court to handle electronic commerce (e-commerce) disputes (Du & Yu 2018:2). The e-commerce complaints reported in China include refund issues, after-sales service problems, dispatch problems, receiving wrong products, difficulties with return and/or exchange, invoice problems and delayed delivery (Statista 2021:1). This shows that even though China’s online retailers have a good delivery system, some customers were dissatisfied with the last-mile delivery. Nguyen, Leeuw and Dullaert (2018:9) define last-mile delivery as the final leg of a supply chain in which goods are delivered to customers.

In Kenya, Vision 2030 lists the retail sector among the six priority sectors that are expected to transform the country to middle-income status by the year 2030 (Government of Kenya 2007:24). This was because retailers are found in different parts of the country, creating many job opportunities to citizens. It is estimated that Kenya’s digital economy (i.e. e-commerce, digital media, e-services and e-travel) generated $1.124 billion revenue, of which e-commerce contributed $0.662 bn in 2019 (Statista 2020:63). This is a 58.8% contribution to the country’s digital revenue. Furthermore, the e-commerce revenue in Kenya is projected to have an annual average growth rate of 25.5% to reach $2056.7 million in 2024 (Statista 2020:65). This means that Kenya still has the potential to generate more revenue from e-commerce. Unlike China, Kenya’s delivery service was found to inhibit the growth of the online retailing subsector (Communication Authority of Kenya 2015:27). The Nielsen report (2019) on beating the odds in the consumer and retail landscape in Kenya also concluded that 70% of Kenyan online shoppers are not willing to shop online again, mainly because of delivery issues. From the report by Nielsen (2019), of great interest to researchers is that there are fewer online customers willing to shop online again (i.e. ‘users’) compared with the ones who are not willing to shop online again (i.e. ‘nonusers’). However, Nielsen (2019) did not investigate the extent to which online customers were dissatisfied with the delivery service in Kenya.

A search for journal articles on customer satisfaction with last-mile delivery service in Kenya failed to give useful results in the Emerald and Scopus databases. However, a further search on Google Scholar found two master’s studies carried out in Kenya that addressed online service quality and customer satisfaction in online retailing (Indeche 2017:8; Onyango 2018:8). The studies consider online service quality to include all activities performed by an online customer on the online retail store before checking out. This means that the two studies did not examine customer satisfaction with the last-mile delivery service. Therefore, there is a need for this study to investigate the problem and more specifically: (1) to establish the extent of customer satisfaction with the different elements of last-mile delivery service offered by online retailers in Kenya and (2) to establish the difference between ‘users’ and ‘nonusers’ in levels of customer satisfaction with last-mile delivery service in Kenya.

This study is based on the expectation confirmation theory developed by Oliver (1977) to explain postpurchase behaviour of individuals. According to this theory, customer satisfaction (i.e. a postpurchase behaviour) is determined by the initial expectations of a service (e.g. last-mile delivery service) before purchase and the perceptions about performance following the use of that service. Furthermore, Collins (2015:16) argues that different last-mile delivery and pick-up alternatives provide different levels of utility, and consumers will choose the option that will maximise their utility. The expectation confirmation theory assumes that if the service performs at or above initial expectations (i.e. positive disconfirmation), the customer is likely to be satisfied. In contrast, if the service falls short of initial expectations (i.e. negative disconfirmation), the customer is likely to be dissatisfied (Oliver 1980:1). However, initial studies conceptualised the construct of satisfaction differently. Some studies considered satisfaction in terms of attitude, while others considered it as experience-specific (Bhattacherjee 2001:5). The author asserts that attitude differs in terms of its ability to predict customer satisfaction. For this reason, this study considers last-mile delivery experience as a better predictor of customer satisfaction. In this manner, online customers are expected to be satisfied with the different elements of last-mile delivery service they experience if the service is performed at or above initial expectations or be dissatisfied if the service fall short of initial expectations.

This study intended: (1) to establish the extent of customer satisfaction with the different elements of last-mile delivery service offered by online retailers, and (2) to establish the difference between ‘users’ and ‘nonusers’ in levels of customer satisfaction with last-mile delivery service in Kenya. To achieve this, a quantitative approach was utilised. A descriptive survey was used to provide numeric description of opinions of a population by studying a sample (Creswell & Creswell 2018:245). The target population was divided into two categories, consisting of the 258 353 ‘users’ and an unknown number of ‘nonusers’ (Kenya National Bureau of Statistics 2019). The targeted population were residents in Nairobi County, whereby the ‘users’ have access to the Internet and have bought goods online within the last 12 months, and the ‘nonusers’ have access to the Internet but have not bought goods online within the last 12 months. A sample of 384 ‘users’ and 384 ‘nonusers’ was obtained from the target populations of 258 353 ‘users’ and the unknown number of ‘nonusers’, respectively. This was based on Saunders, Lewis and Thornhill’s (2019:300) assertion that for most studies in business and management, researchers estimate the target population characteristics at 95% certainty to within ± 3% to 5% of its true value. Social media influencers were used to recruit the targeted ‘users’ and ‘nonusers’ because they have a huge number of online followers. This was because of the online retailers’ unwillingness to provide access to the contact details of their online customers in Nairobi County, citing confidentiality reasons. In addition, the use of social media influencers to recruit the targeted ‘users’ and ‘nonusers’ was because of the coronavirus disease 2019 (COVID-19) pandemic. The government had banned physical interactions in Kenya, making it impossible to physically identify the online customers through street intercepts (Government of Kenya 2020:2).

A structured questionnaire was administered to online customers aged 18 years and above (i.e. ‘users’ and ‘nonusers’). The questionnaire had a screening question asking online retail customers the last time they shopped online to distinguish ‘users’ from ‘nonusers’. A link to the online research questionnaire that was developed using Google Forms was distributed by the social media influencers through their social media applications, such as Twitter and Facebook, for a period of 8 weeks, starting from 01 July 2021 to 25 August 2021. The online research questionnaire comprised 13 items on satisfaction with the different elements of last-mile delivery service used in this study.

Validity was ensured in this study by pretesting the structured questionnaire on 10 online customers to ensure all instructions and questions were clear (Leedy, Ormrod & Johnson 2019:130). Furthermore, the structured questionnaire was developed from prior similar studies for purposes of comparison (Mentzer, Flint & Hult 2001:8; Nguyen 2020:8; Wolfinbarger & Gilly 2003:6). Cronbach’s alpha was used to test the reliability of the analysed data, whereby the preferred value is at least 0.70 for multi-item constructs (Pallant 2016:23; Saunders et al. 2019:518). The Statistical Package for the Social Sciences (SPSS) version 28 was utilised to generate descriptive statistics for objective one. Composite means, an independent sample t-test and the mixed model analysis of variance (ANOVA) approach were used to compare the satisfaction scores between ‘users’ and ‘nonusers’ required for objective two.

A total of 407 ‘users’ and 60 ‘nonusers’ responded to the online research questionnaire. This translated to a 105.99% and a 15.63% sample response rate for ‘users’ and ‘nonusers’, respectively. It was observed that the response rate for ‘nonusers’ was very low in the first 4 weeks compared with the response rate of ‘users’. This can be attributed to the fact that ‘nonusers’ are no longer shopping online because of various reasons; thus, they were not attracted to anything to do with online shopping, which was the aim of this study. The initial low response rate forced the researcher to extend the data collection period from 4 to 8 weeks, which ensured achievement of a 15.63% response rate for ‘nonusers’, whereas the response rate for ‘users’ surpassed the targeted sample size. In comparison, the response rates in this study are more favourable than those of other similar studies. For instance, Griffis et al. (2012:6) study in the USA reported a 47.3% response rate, and Brink’s (2018:53) study in South Africa reported a response rate of 15.64%.

The 13 items on the online research questionnaire were ticked by each respondent based on a five-point Likert-type scale, ranging from 1 = very dissatisfied to 5 = very satisfied. Table 1 shows Cronbach’s alpha for all the scales with values above 0.70. In addition, the interitem correlation was inspected, and none was below 0.3, indicating that each item had a high correlation with the total score (Pallant 2016). This implies that the questions combined in the scale are internally consistent in their measurement.

To establish the extent of customer satisfaction with the different elements of last-mile delivery service offered by online retailers, most ‘users’ were satisfied with options for delivering items (54.4%). In addition, most ‘users’ were dissatisfied with the ease of locating the returns procedure on the retailer’s website (41.5%). Most ‘nonusers’ were satisfied with the accuracy of the bill from the retailer (6.9%). Moreover, most ‘nonusers’ were dissatisfied with the clarity of the returns policy, for example, refunds and/or replacements (7.3%), as shown in Table 2. This implies that many ‘users’ were satisfied with delivery options, whereas a majority of ‘nonusers’ were satisfied with delivery fees. However, a majority of both ‘users’ and ‘nonusers’ were dissatisfied with returns.

The mixed model ANOVA was used to establish the difference in the levels of customer satisfaction with elements of last-mile delivery service for ‘users’ and ‘nonusers’. The results in Table 3 reveal that ‘users’ were most satisfied with delivery options (mean of 3.5545) and least satisfied with returns (mean of 2.8731). However, ‘nonusers’ were most satisfied with delivery fee (mean of 3.2389) and least satisfied with returns (mean of 2.6056). In addition, ‘users’ were more satisfied with delivery options, delivery tracking and returns compared with ‘nonusers’. In contrast, the ‘nonusers’ were more satisfied with delivery timeliness and delivery fee compared with the ‘users’ (Table 3).

Table 4 shows an independent-samples t-test conducted establishing that there was a statistically significant difference in the score of satisfaction with delivery options for ‘users’ (M = 3.5545, s.d. = 1.37127) and ‘nonusers’ (M = 3.1111, s.d. = 1.16391; t [85.099] = 2.688), p = 0.009 (two-tailed). The magnitude of the differences in means (mean difference = 0.44335 CI: 0.11546 to 0.77125) was very large (Table 4). However, there was no statistically significant difference in the score of satisfaction with the other elements of last-mile delivery service.

Collins (2015) argues that different last-mile delivery and pick-up alternatives provide different levels of utility, and consumers will choose the option that will maximise their utility. This study revealed that most ‘users’ were satisfied with delivery options. This implies that the delivery options (e.g. at home, office, pick-up station) offered by online retailers in Kenya meet or exceed the expectations of online customers who have bought goods online within the last 12 months. This finding supports Mofokeng’s (2021) study, which found that satisfaction for online customers in South Africa was influenced by delivery mode. Dholakia and Zhao’s (2010) study in the USA also found that having a variety of shipping options after checking out from an online store contributes positively to satisfaction. The prior studies were conducted in countries that are e-commerce leaders in their respective regions. Thus, ensuring that online customers are satisfied with delivery options will likely lead to increased online shopping in Kenya.

The results that most ‘users’ and ‘nonusers’ were dissatisfied with returns imply that online retailers in Kenya have failed to meet or exceed the returns expectations of online customers. This means that the returns policies offered by online retailers either lacked clarity or customers find it very difficult to locate the returns procedure on the retailers’ portals. In addition, customers are dissatisfied with the options for returning items. This finding supports the results of Statista (2021), which found that refund issues and difficulties with return or exchange were among the e-commerce complaints reported in China, despite the existence of an effective and efficient delivery system. Thus, online retailers in Kenya should improve the way they offer returns to online customers. This can be justified by Jones’s (2017) study, which established that returns contribute to customer satisfaction for online customers in the USA.

The mixed model ANOVA revealed that the difference in satisfaction with elements of last-mile delivery service between ‘users’ and ‘nonusers’ was largest on delivery options and least on delivery timeliness. This means that the online customers’ expectations of delivery options are very different between ‘users’ and ‘nonusers’ as compared with the expectations on delivery timeliness. This result supports the study by Vakulenko et al. (2022) in Sweden, which found that even though the availability of different delivery options can lead to satisfaction, the degree of satisfaction may vary across the different groups of customers. The authors found that online customers in urban residential areas in Sweden were more satisfied with delivery options compared with those in rural residential areas. This study also found that ‘users’ were more satisfied with delivery options, delivery tracking and returns compared with ‘nonusers’. Therefore, online customers in Kenya continue buying from online retailers who meet or exceed expectations on delivery options, returns and delivery tracking. This is validated by a study by Cao et al. (2018) that found customer satisfaction for online shoppers in China and Taiwan was partially predicted by tracking. However, this contradicts the Dholakia and Zhao (2010) study in the USA that found order tracking failed to significantly contribute to customer satisfaction in the second year as compared with the first year.

The findings also revealed that ‘nonusers’ were more satisfied with delivery timeliness and delivery fee compared with the ‘users’. This implies that the last-mile delivery expectations of online customers in Kenya differ across the different groups of online customers. This result concurs with the Nguyen et al.’s (2019) study in the Netherlands that found the existence of clusters of customers who value different delivery attributes. Some customers, referred to as ‘price-oriented’ customers, considered delivery fees as the most important delivery attribute, while the ‘value for money–oriented’ customers value price as well as the convenience of shopping goods online (Nguyen et al. 2019). This justifies the need for online retailers in Kenya to understand the last-mile delivery expectations of all groups of online customers to satisfy their needs. An independent-samples t-test found that there was a statistically significant difference in the score of satisfaction with delivery options for ‘users’ and ‘nonusers’. The difference in scores of satisfaction on the other elements of last-mile delivery service was not statistically significant. Thus, it can be deduced that there is a known difference in the levels of satisfaction on last-mile delivery options for ‘users’ and ‘nonusers’.

The purpose of this study was to establish the extent of customer satisfaction with the different elements of last-mile delivery service offered by online retailers in Kenya. In addition, it was meant to establish the difference between ‘users’ and ‘nonusers’ in levels of customer satisfaction with the last-mile delivery service. This follows the report by Nielsen (2019) that indicated there are fewer online customers willing to shop online again (i.e. ‘users’) compared with the ones who are not willing to shop online again (i.e. ‘nonusers’), mainly because of delivery issues. However, the report failed to indicate the extent to which the online customers were satisfied or dissatisfied with delivery. Therefore, this study provides online retailers in Kenya with a better understanding of the last-mile delivery expectations of online customers. In a country where the majority of online customers are not willing to buy online again because of delivery issues (Nielsen 2019), it is required that online retailers focus on providing the different elements of last-mile delivery service that are excellent to retain current customers as well as attract new ones. This study revealed that most ‘users’ and ‘nonusers’ in Kenya are dissatisfied with returns. In addition, it was revealed that there is a statistically significant difference in the level of satisfaction with delivery options for ‘users’ and ‘nonusers’.

The finding that most ‘users’ and ‘nonusers’ are dissatisfied with returns justifies the need for online retailers to improve the way they offer returns to their customers. This may include offering more options for returning the ordered goods, making it easy to locate the returns procedure on the retailer’s website, and having a clear returns policy. The finding that there is a significant difference in satisfaction with the delivery options between ‘users’ and ‘nonusers’ requires online retailers to understand the specific expectations of online customers in relation to the last-mile delivery options. This may include offering a wide variety of delivery options, making it easy for online customers to reach the order pick-up points and using trusted delivery personnel. Thus, it is important for online retailers to selectively offer delivery options that meet the needs of ‘users’ and ‘nonusers’. The online retailers should further investigate the specific delivery option needs of ‘users’ and ‘nonusers’. This will ensure that there is continued growth in the online retail subsector in Kenya.

This study advances an understanding that last-mile delivery service comprises various elements, including delivery options, delivery timeliness, delivery fee, delivery tracking and returns (Cao et al. 2018; Holdorf & Haasis 2014; Hu et al. 2016; Nguyen et al. 2018). Little is known of prior research that has comprehensive empirical evidence of customer satisfaction with the identified elements of last-mile delivery service. Furthermore, little is known of any comparative study done before that examines the difference between ‘users’ and ‘nonusers’ in levels of customer satisfaction with last-mile delivery service.

Although this study provided new perspectives on customer satisfaction with elements of last-mile delivery service in Kenya, it has some limitations. For instance, online customers from other parts of the country were excluded from this study. Future research can consider expanding the geographical scope to include other counties in Kenya, as well as rural areas, to determine customer satisfaction with the different elements of last-mile delivery service. The elements of last-mile delivery service were also limited to delivery options, delivery timeliness, delivery fee, delivery tracking and returns. This was based on the review of prior similar studies. Thus, future research can be extended to include other emerging elements of last-mile delivery service. Findings were also limited to the online customers, following a quantitative approach. Future research may apply a mixed-methods approach incorporating both online customers and online retailers to investigate customer satisfaction with elements of last-mile delivery service to gain more insights into the concepts in this study. This study was also limited by the low response rate for ‘nonusers’. Future research may be extended to include more ‘nonusers’ by either considering those located in other parts of the country or using alternative methods of data collection.