Bibliometric Analysis of Online Food Delivery Research Trends using CiteSpace

We analyzed the annual publication trend of existing literature on OFD, as shown in Fig. 2. Clearly, the number of publications exhibited an overall upward trend, rising from 2 articles in 2016 to 187 articles in 2022. This represents a nearly 100-fold increase over the span of 7 years, indicating growing attention from researchers towards OFD. Notably, there was a significant surge in publications after 2020. The emergence of new trends and abrupt changes in scientific literature may be attributed to internal and external factors, possibly related to the development of e-commerce and the outbreak of COVID-19. As of June 2023, 108 publications have been published in 2023, and more are expected in the future, with an estimated total surpassing 220 articles by the end of 2023.

Fig. 2.

The number of publications

The first literature related to OFD appeared in 2016 when Quevedo-Silva et al.[12] published the article titled "Intentions to purchase food through the internet: developing and testing a model" in the British Food Journal. In their study, the researchers utilized the Theory of Planned Behavior to assess customers' intentions to purchase food through the internet and proposed a more comprehensive model.

The analysis of published articles identified a total of 268 different journals related to OFD. Table 1 presents the top 10 journals with the highest number of published OFD articles. Notably, the British Food Journal ranks first with approximately 24 OFD articles. Regarding the journals' impact factors (IF), they range from 2.948 to 11.5. The International Journal of Hospitality Management stands out with the highest impact factor of 11.5.

Table 1.

Top 10 journal publication

Through the analysis of published articles, a total of 88 categories related to OFD were identified. Fig.3 presents the top 10 categories with the highest representation. The largest category is business, followed by management, hospitality leisure sport tourism, environmental sciences, and food science technology, among others. This indicates that research on OFD primarily revolves around the fields of business and management, gradually evolving into an interdisciplinary study with a high degree of integration.

Fig. 3.

Top 10 categories of publications

The articles on OFD are geographically diverse, originating from 165 countries or regions. Fig. 4 presents the top 10 countries or regions in terms of the number of published OFD articles. Notably, 141 articles are from China, which can be attributed to its large population and leading position in the development of e-commerce. China holds the largest OFD market globally, sparking significant research interest in the field. Other countries with substantial literature, exceeding 30 articles each, include the United States, India, the United Kingdom, Australia, and South Korea. According to Statista[13], the countries with the highest OFD market revenue rankings were China, the United States, the United Kingdom, India, Germany, and France. Despite ranking relatively lower in the global OFD market, Australia and South Korea exhibit high research activity and interest in the field.

Fig. 4.

Top 10 countries and regions of publications

The articles has been ranked in descending order based on the number of citations, and Table 2 displays the top 10 highly cited articles. Among them, the articles by Yeo et al.[4] has received 106 citations. This article examined the structural relationships between convenience motivation, post-usefulness, hedonic motivation, price-saving orientation, time-saving orientation, prior online purchase experience, customer attitude, and behavioral intention towards OFD services.

Table 2.

Top 10 cited articles 

Keyword co-occurrence network analysis was conducted on the articles (Fig. 5), wherein the node type selected was keyword. The resulting co-occurrence network consisted of 299 nodes (merged network nodes) and 1491 links (N=299, E=1491, network density=0.0335). The colorful citation tree-rings depicted in the graph represent the nodes, with each node labeled by a specific keyword, indicating the number of articles in the OFD domain that utilized that particular keyword. The thickness of the citation tree-rings corresponds to the quantity of articles associated with the respective keywords; thicker rings indicate a higher number of occurrences. The colors within the citation tree-rings signify the publication time of the literature, with larger areas of a particular color representing a higher volume of articles published within that time interval. The colored connecting lines (links) between nodes represent the relationships between keywords. The interconnections among nodes signify their intricate relationships.

Fig. 5.

Keyword co-occurrence network

By examining the centrality (centrality>0.01) and frequency (frequency>15) of high-frequency keywords, we identified the top 10 most frequent keywords. Table 3 displays these 10 keywords along with their centrality values, frequencies, and the year of their first appearance. The most frequently occurring keyword is “satisfaction”, which appears 51 times and was first mentioned in 2020.

Table 3.

Top 10 keywords

The Log-likelihood Ratio (LLR) algorithm was employed for keyword co-occurrence clustering analysis, resulting in the identification of 9 clusters. After removing an ineffective cluster, 8 clusters were retained (Fig. 6). The data indicated Modularity Q = 0.485, which falls within the acceptable range of 0.4 to 0.8, signifying that the network was divided into loosely coupled clusters. The mean silhouette value was calculated to be 0.6053, indicating that these clusters exhibit relatively good average homogeneity.

Fig. 6.

Network with clustered keyword

Each colored cluster represents a distinct group, and the 8 clusters, ranked from largest to smallest, are as follows: #0 online food delivery apps, #1 gig economy, #2 food safety, #3 customer satisfaction, #4 food environment, #5 digital platforms, #6 customer behavior, and #7 food waste. The high-frequency keywords within each cluster are depicted in black text.

Table 4 presents the size, silhouette value, average year of publication, and labels for the 8 identified clusters through Log-likelihood Ratio (LLR) algorithm-based keyword co-occurrence clustering analysis. The silhouette value indicates the homogeneity of the clusters, where silhouette>0.7 is considered to exhibit high homogeneity, and higher values indicate greater keyword consistency within the cluster.

Table 4.

Cluster details

(1) Cluster #0 food delivery apps, comprises 56 articles, with an average publication year of 2019. This cluster reflects research on topics such as customer technology acceptance, application quality, and customer value insights.

(2) Cluster #1 gig economy, includes 47 articles, with an average publication year of 2021. The gig economy refers to a labor market composed of freelancers and temporary workers who provide short-term or on-demand services to earn income. This cluster covers research on platform work, platform economy, algorithm management, and labor relations.

(3) Cluster #2 food safety, contains 37 articles, with an average publication year of 2020. Food safety pertains to ensuring food safety, reducing health risks, and preventing food poisoning during food processing, preservation, and sales. This cluster includes research on COVID-19 pandemic, loyalty, service advantage, trust, etc. After COVID-19, food obtained through OFD has become a potential source of transmission, making food safety a significant research topic.

(4) Cluster #3 customer satisfaction, consists of 37 articles, with an average publication year of 2020. Customer satisfaction refers to the percentage of customers whose experience with a company and its products or services exceeds a specified satisfaction goal. This cluster encompasses research on service quality, brand trust, and continuation intention.

(5) Cluster #4 food environment, comprises 32 articles, with an average publication year of 2020. Food environment refers to the physical presence of food that influences people's dietary behaviors, including the proximity to food stores, food store distribution, food services, and any physical entities that allow access to food. This cluster includes research on public health, green branding, microplastic pollution, etc.

(6) Cluster #5 digital platforms, include 28 articles, with an average publication year of 2020. This cluster covers research on collaborative consumption, Google maps, traffic conditions, etc.

(7) Cluster #6, customer behavior, consists of 21 articles, with an average publication year of 2018. Customer behavior refers to all activities related to the purchase, usage, and disposal of OFD services by individuals, groups, or organizations. This cluster includes research on influencing factors, application usability, various food choices, etc.

(8) Cluster #7 food waste, contains 12 articles, with an average publication year of 2017. Food waste refers to the amount of food that goes from residences, commercial establishments (restaurants), institutional sources, and industrial sources to landfills. This cluster encompasses research on health consciousness, food quality, user emotions, etc.

The timeline visualization provides a highly intuitive and accurate reference for understanding the evolutionary processes of various sub-domains. The timeline visualization depicts the clusters arranged along the top timeline axis from left to right. The main 8 clusters' timelines are shown in Fig. 7, with the top representing the timeline from 2016 to 2023, from left to right. Circular nodes represent highly cited keywords or keywords that triggered bursts of activity, and the labels of the keywords are placed below the circular nodes. The colored curves indicate the co-citation links added in the corresponding colored years. The clusters on the right are numbered in descending order, starting from 0, where cluster#0 is the largest cluster, followed by cluster#1 as the second largest cluster.

Fig. 7.

Timeline visualization

Horizontally observing Fig. 7, clusters #0, #1, and #3 exhibit highly concentrated burst nodes. Vertically observing Fig. 7, it can be observed that burst nodes appear in each cluster in the years 2020 and 2021.

According to the results of keyword burst detection, the top 10 keywords with the strongest bursts are arranged based on their appearance years (Fig. 8). These 10 keywords are: impact, acceptance, adoption, gig economy, commerce, labor, intention, big data, behavior, and model. The blue line represents the timeline, while the red line indicates the duration of the keyword bursts. A longer burst duration indicates a higher level of attention during a certain period. The keyword "acceptance" has the longest burst duration, starting in 2017 and ending in 2020. Notably, there were no burst keywords detected in the recent two years (2022-2023).

Fig. 8.

Top 10 keywords with the strongest citation burst

This trend encompasses three clusters: #0 online food delivery apps, #3 customer satisfaction, and #6 customer behavior. Researchers have extensively explored theoretical models to examine the factors influencing customers' willingness to adopt and continue using OFD services. These factors encompass trust, customer word-of-mouth, satisfaction, brand trust, loyalty, and service quality. Additionally, researchers have investigated how customer demographics, such as gender, age, or education level, moderate these factors.

The enhancement of customer experience in OFD has emerged as a central focus, encompassing efforts to improve the user interface, minimize waiting time, ensure food quality, and optimize pricing. Despite the significant growth of OFD, adoption rates remain relatively low in some countries, making it imperative to understand and promote the positive aspects of OFD to expand its market presence.

Furthermore, with the impact of COVID-19 gradually receding, it becomes essential to explore potential changes in customer usage intentions and behavior. Investigating how the decline of the pandemic affects customers' perceptions and preferences towards OFD services will be a pertinent area of investigation in future research.

This trend involves three clusters: #1 gig economy and #5 digital platforms. OFD work exemplifies the essence of platform work, setting it apart from traditional business models due to its creation of a two-sided market. Within the OFD system, value is generated by customers, platforms, restaurants, and delivery man concurrently. Delivery man, operating as freelancers, provide services to both customers and restaurants without formal employment by the platform. The algorithmic functions of OFD platforms have been subject to scrutiny for their potential exploitation of delivery man.

Despite these concerns, the role of the delivery man within the system cannot be underestimated, as any unavailability or irresponsibility on their part could result in losses for other stakeholders. Therefore, ensuring harmonious coexistence with all stakeholders is vital not only for safeguarding customer satisfaction but also for fostering higher profits. In future research, special attention can be focused on developing policies and regulations that address the concerns and well-being of delivery man. By doing so, the industry can strive to create a balanced and sustainable environment where all stakeholders can thrive.

This trend comprises three clusters: #4 food environment, #2 food safety, and #7 food waste. Food, as a crucial element within the OFD system, undergoes movement between restaurants, delivery man, and customers. The rapid and extensive distribution of food in OFD can pose the risks of widespread foodborne illnesses, especially in the aftermath of the COVID-19 pandemic. Concerns about food safety and food health have escalated, garnering increased attention from customers. Consequently, in future research, the focus on developing robust management systems for food safety is likely to be prominent.

Furthermore, the rapid expansion of OFD has resulted in heightened reliance on the platform by customers, contributing to food waste and pollution from single-use food disposables. Food waste poses significant threats to both society and the economy, while single-use disposables add to environmental hazards. Food waste occurs at various stages, including food processing, sales, and consumption. Restaurants may be affected by irregular acceptance of orders, leading to surplus inventory and waste during the processing stage. Additionally, OFD generates substantial plastic waste, necessitating sustainable waste management solutions.

Addressing these challenges, existing OFD platforms offer customers the option to opt-out of disposable cutlery. However, in practice, restaurants may still provide cutlery to customers even if they choose the eco-friendly option. Promoting environmental awareness among customers and restaurants emerges as a potential research area in future OFD studies. By encouraging responsible practices and sustainable approaches, the industry can mitigate its environmental impact and work towards a more eco-friendly future.