A bibliometric analysis database is gathered from the Scopus database. The Scopus database is regarded as the most systematic and scientific database for bibliometric analysis ([22]). Scopus has been recognized as the best database for bibliometric analysis in addition to the aforementioned claim ([8]; [33]). Therefore, it can be claimed that Scopus is the most extensive database that includes a wide range of information on articles and that papers must meet strict criteria in order to be included in this database. The principal database for this study was Scopus, which was chosen for its exhaustive coverage, rigorous quality standards, and robust analytical features. Scopus provides a more comprehensive yet meticulously curated compilation of scholarly journals, conference proceedings, and publications from a variety of disciplines in comparison to Google Scholar and the Web of Science. Scopus’ Content Selection and Advisory Board (CSAB) enforces rigorous selection criteria, which guarantees the credibility of indexed materials, in contrast to Google Scholar, which contains unfiltered and non-peer-reviewed sources. Although the Web of Science also enforces rigorous inclusion standards, its coverage is more limited and primarily Western-centric. Moreover, Scopus offers sophisticated citation metrics, including CiteScore, SNIP, and SJR, which provide a more comprehensive understanding of the impact of research in comparison to the traditional Impact Factor of the Web of Science. Its suitability for systematic academic research is further enhanced by its frequent updates, user-friendly interface, and integration with tools like Mendeley. Consequently, Scopus is the most dependable and well-balanced platform for the retrieval of scholarly literature that is both globally inclusive and of high quality. For the best search, a number of keywords are identified, such as “online food delivery apps,” “food delivery apps,” and “online food delivery.” “Online food delivery apps” was the best keyword to use. Apps for online meal delivery will be examined in this article. When the term “online food delivery apps” was used to search the Scopus database, 328 references were found.

Document type screening was conducted to include journal articles, review papers, and conference papers. These document types were classified as signifying peer-reviewed and citable academic contributions. Non-scholarly elements, including book chapters, editorials, and notes, were eliminated to ensure analytical accuracy and research quality.

When there are a lot of language papers, the first non-filter is utilized for the language filter. The selected papers were then those that were printed in journals. 328 articles were ultimately chosen for study. Since further filtration may lower the number of articles and may affect the bibliometric analysis, no sorting criteria other than these two were used.

A duplicate elimination procedure was executed using both automatic detection on the Scopus platform and human verification of author names and publication titles. This process guaranteed the removal of superfluous records and the maintenance of data integrity. Subsequent to these filtering methods, a total of 328 unique and valid records were preserved for bibliometric analysis. No exclusion parameters such as publication year, source title, or topic category were used, providing full coverage of the scientific environment for Online Food Delivery Applications. This thorough screening approach improves the reliability and representativeness of the dataset used for future performance and science mapping evaluations.

One performance analysis approach is total publication. The overall number of publications on the study topic is taken into account. The 328 gathered findings from the articles released each year are included in the yearly scientific production. The papers were arranged in groups based on the year they were published in the journals. It is possible to conclude from Figure 1 that OFDA research is growing annually. Although there was just one publication in 2013, by 2017, there were seven, and by 2023 and 2024, there were sixty-five and sixty-nine publications, respectively. Forty-one documents are counted until July 2025. The phrase “active years of publishing” refers to the number of years that research has been done in the field to examine the phenomena. Groundbreaking research is still being conducted, and the active year begins in 2013. Considering the trend, this issue has the potential to be studied further because there are still comparatively fewer studies conducted annually.

Figure 1. Shows year-to-year publications from 2013 to 2026.

The journals that publish the most articles on a certain subject are considered promising. R software was used to identify promising journals based on Bradford’s law. Table 1, which lists the top 10 promising journals in the field of OFDA, shows the most promising journals. With 13 publications in the subject of OFDA, the British Food Journal has an advantage over other journals. With nine publications in the subject of OFDA, ranking second overall, the Journal of Retailing and Consumer Services has an advantage over other journals. With a total of six publications in the subject of OFDA, the ACM International Conference Proceeding Series, International Journal of Contemporary Hospitality Management, Journal of Foodservice Business Research, and Sustainability Switzerland have an advantage over other journals. Bradford’s law was chosen for examination in publications that showed promise. The outcome was a graph that showed the number of papers published by a journal together with the source name. As can be seen in Table 1, the graph was later shaped into a table. By identifying the most pertinent journals that are ahead of the curve in producing information pertaining to OFDA, Bradford’s legal analysis will assist scholars to swiftly uncover and choose a few journals that will support their study of OFDA and future research. A journal’s potential to impact future scholars in a certain topic increases with the number of papers it publishes in that field.

In the realm of research, the nations with the highest number of published papers and citations are regarded as dominating nations. The database is examined using R software for the analytical task. The results are interpreted using citations and scientific production in order to identify the leading nations in the OFDA discipline. Both the quantity of papers and the number of citations are used to determine which nations are dominant. The top ten nations are chosen to be examined from both angles. By being aware of the nation that dominates according to a scan of numbers, India leads all other nations in terms of both average article citations (17.90) and documents (839). With just 471 papers, Finland has high citations (157.00), and Norway ranks second in terms of average article citations, which is a surprising aspect. Despite having more published texts, the USA lagged behind numerous nations, including China, Finland, Korea, and others, in terms of citations. Italy is in a similar situation. Compared to its Chinese equivalent, this country obtained more citations with fewer papers. Information from Figure 2 and Figure 3 was transformed into a tabular format in Table 2 to facilitate the interpretation of the analysis. Consequently, it may be said that Finland is at the forefront of OFDA research on citations. Prominent nations in the field include Brazil, Portugal, Italy, and Norway.

Table 1. Shows the name of the journal along with the number of papers published.

Figure 2. Shows a world map depicting the number of documents published by each country.

Figure 3. Shows the number of citations received by each country’s documents.

Table 2. Shows the countries’ names, the number of documents published, and the number of citations received.

The quantity of papers each author publishes determines which writers are the most pertinent. R software thus calculated it by counting the number of OFDA-related publications they had published. According to the data, an author’s significance increases with the number of papers they have written. The top ten most pertinent writers are displayed in Figure 4 and each with 29 papers apiece. The top ten writers cited can help readers understand their work and what more needs to be done. With four documents each, DHIR AMANDEEPS and MAI TAI XUAN have made noteworthy contributions.

The writers with the most citations in their field are considered to be the most influential. Only the citations serve as its foundation. As a result, the total number of documents that an author gets was used to identify significant writers. The most prominent writers will have the most documents in their account, while the least important authors will have fewer citations. Clearly at the top of the list are authors like Dhir Amandeep (827 citations) and Kaur Puneet (790 total documents). However, writers like Ray Arghya and Bala Pradip Kumar rank third and fourth, respectively. Figure 5 displays the remaining writers.

Figure 4. Shows the authors’ names and the number of papers published by them.

Figure 5. Shows authors’ names and numbers of total document.

Citation analysis is a scientific mapping approach. When one publication is mentioned by another, it links to the original work ([22]). Two criteria are used for citation analysis: (1) global citation and (2) local citation. The quantity of citations an article obtains from readers may be used to gauge its impact in citation analysis.

4.6.1. Most Globally Cited Documents

The articles with the most citations without any filtering, i.e., topic domain, are referred to as the most globally cited texts ([22]). In other words, global citations are those that a publication receives regardless of whether it has been cited inside or outside its topic domain. The ten most frequently cited papers worldwide might also be considered as having a significant impact on other writers’ decisions to include citations in their works. Both papers pertaining to OFDA and those unrelated to OFDA reference these records. Figure 6 displays the ten most influential papers, and Table 3 analyzes them. The top ten most internationally cited publications that were examined from Figure 6 are included in Table 3 along with their article names, authors, and citation counts.

Figure 6. Shows the most influential papers in terms of global citation.

Table 3. Shows the article name, authors’ names, and citations of the top 10 globally cited documents.

4.6.2. Most Locally Cited Documents

Documents that are discussed within the topic domain are known as locally cited documents; for example, one article may acquire citations from another article in the same field ([22]). In other words, local citations are those that are specific to the field in which the materials are found. For instance, one publication affiliated with OFDA is mentioned in another that is likewise related to OFDA. As a result, the majority of locally cited papers examine works that are often referenced or mentioned in the field. One can rely on the papers displayed in Figure 7 to gain understanding about the topic matter. These documents are especially pertinent in the subject of OFDA and can be considered trustworthy sources to obtain early paperwork. It should be highlighted that, for the reasons stated in their definition, local citations are always lower than global citations. The analysis ofFigure 7 is completed in Table 4.

Figure 7. Shows documents receiving a total number of local citations.

One method of scientific mapping is co-citation analysis. When two references are mentioned together in a third work, it is assumed that they are related in some way or share a similar content structure. To uncover the conceptual framework of a particular topic of study, co-citation analysis is utilized ([22]). Based on the clusters created, this approach also assists us in identifying the most important publications. Each cluster has a theme and is based on a certain foundation. After grouping the documents into clusters, the co-citation analysis identifies the publications that are most related to each subject. Researchers can learn more about the article based on their interests thanks to this analysis. Additionally, future scholars might gather literature on a specific topic by consulting the related publications. The VOSViewer program employs a co-citation approach for the analysis. The established bibliometric technique ([13]; [53]; [43]) asserts that a minimum citation threshold must be enforced to exclude weak or accidental connections between sources. Only publications that have substantial intellectual influence and receive consistent recognition in the literature are included, according to a minimum citation criterion of five. To achieve a compromise between network density and interpretability, bibliometric tools such as VOSviewer and Bibliometrix often use this method. Only sources referenced five or more times are included in the study, enabling a focus on the core concepts of a subject without being obscured by seldom-cited articles that provide noise rather than insight. Only publications that have at least five citations in published articles are chosen. Just 156 out of 25,030 reach the citation limit. Figure 8 was examined by assigning weights to the links.

Seven clusters are created in total. By giving the connections more weight, you may show how many additional papers are related to the papers in the reference list. A paper is better suited for study if it has a greater number of connections (Table 5).

Figure 8. Shows a map of co-citation analysis based on the authors’ name.

Table 4. Shows the article name, authors’ names, and citations of the top 10 locally cited documents.

Table 5. Shows the interpretation of the co-citation map.

The data in Figure 9 shows how many documents were written by 10 distinct people. With a count of four, the chart indicates that Dhir, A. has the most papers. Each of the following writers has written three documents: Abbasi, A.Z., Allman-Farinelli, M., Bennett, R., Gibson, A.A., Hakim, M.P., Jia, S.S., Kaur, P., Nguyen, T., and Partridge, S.R. This suggests that one author in this dataset is somewhat more productive than the others in terms of the number of papers they have released.

The cooperation data is shown in Table 6 and includes many author pairings that have co-authored precisely three times. There are three partnerships between the couples Bennett R. and Zorbas C., and Hakim M.P. and da Cunha D.T. Allman-Farinelli M., Gibson A.A., Jia S.S., Partridge S.R., and Phongsavan P. have a broader network of cooperation, since every pair that might possibly be in this group has cooperated precisely three times. This suggests a very cohesive and effective research team.

Figure 9. Shows a documents written by author.

Table 6. Shows the interpretation of the co-citation map.

A further science mapping method that uses “author keywords” is co-occurrence analysis. Because the study aims to concentrate on the author’s preferred method of conducting research, it contains terms that the author has utilized as keywords. Words that occur together are used to create themes or clusters in co-word analysis ([22]). Because only those keywords are desired, only those that appear in at least five articles are used for the analysis.

These are frequently employed by several writers to assess how a field of study is operating, and only powerful words may be examined. The threshold limit is met by 40 words. Figure 10 was used to perform co-occurrence analysis, forming seven clusters. When it comes to online meal delivery and food delivery applications, the larger the circle, the more powerful the word. In a cluster (1) of red highlight words, the word “food delivery” occurred 27 times, whereas the terms “customer satisfaction” and “mobile application” appeared 10 times. The terms consumer behavior (8) and continuous intention (7) are included in cluster (2) of the color green. Words like “e-commerce” (13), “online food delivery” (36), and “food delivery app” (11) in a blue cluster (3) are highlighted. Words like “mobile applications” (which appeared 12 times) and “food environment” (9 times) were included in the gold-colored cluster (4). The purple cluster (5) contains the terms COVID-19 (27) and meal delivery apps (36). Cluster number six is sky blue and includes the terms “gig economy” (8) and “Malaysia” (5). The orange-colored cluster (7) draws attention to terms like sentiment analysis (5) (Table 7).

Figure 10. Shows a map of the co-occurrence of keywords.

Table 7. Shows various keywords formulated through co-occurrence analysis.

Thematic Analysis

Through the use of co-occurrence analysis, thematic clusters were created. Co-occurrence analysis is a scientific mapping approach that groups words that occur together often to create clusters of different topics ([22]) in Figure 11. Every term was chosen in order to recognize the many themes that emerged. This is due to the fact that all keywords provide a more accurate image for theme creation, which considers the words used in the abstract, title, and keywords. According to other research on co-word analysis ([13]; [11]; [9]), a minimum occurrence criterion should be used to make sure that the keywords that are found reflect recurrent and thematically stable research trends. In medium-sized datasets (200 - 400 documents), selecting 10 occurrences is seen as the best way to capture concepts that appear frequently while excluding isolated or author-specific phrases. This level makes it easier to see clusters and themes in VOSviewer and guarantees thematic robustness. Because the coverage of all keywords is greater than that of the author’s keyword, words that occurred in documents at least ten times were picked for the thematic analysis. Words that appeared in documents more often were also considered. Because a keyword becomes more significant in a field if it appears more than ten times in various papers, we must give it a weight of at least ten times before it appears in papers. Second, a hit-and-trial approach was used with varying weights; findings were obvious and informative when the weight was ten times. A total of 37 words satisfied the baseline, which assigned weight to keyword occurrences. Three clusters in all were created. The first theme discusses OFDA’s function as a tool for food marketing. The second theme looks at how consumers behave and feel about OFDA. The third theme focuses on how customers assess food delivery according to OFDA. The same idea applies to analysis here as it does to co-occurrence; that is, a word is given more weight the larger the circle, which is visible.

Figure 11. Shows a map of co-occurrence analysis to form themes through all keywords.

Thematic evolution is used to determine the direction of the trend, the diversification of the subject over time, and the paradigm change. Time is broken down into groups in theme progression in order to examine changes over time. Additionally, this research aids in comparing the time windows that the author determined to be appropriate. Although time can be shortened appropriately, time has been divided into two halves for analysis. Thirteen years of research time are split into two periods of ten and three years between 2013 and 2026. This is done to examine the work completed in the first and second halves of the section and to see how the subject of study has changed. Thematic evolution is examined using author keywords.

A paradigm change in the way the OFDA research was carried out in terms of OFDA keywords and subjects was noted after examining the data in two-time windows. First, the term “meal delivery apps” was used more frequently between 2008 and 2015. Food delivery, one of the well-known OFDA systems, has recently been the subject of extensive human research. Similarly, food delivery now includes key concepts like sentiment analysis, food applications, food delivery apps, and customer experience. Food delivery takes up the majority of the 2024-2026 time window. Since the aforementioned terminologies are used to study the value of food delivery, they may have developed into food delivery. Menu delivery expanded to include menu choices and a health star rating. There has undoubtedly been a shift in perspectives on the research domain’s time slots and emphasis points. It is evident from Figure 12 that the focus of study has evolved throughout time. In contrast to prior years, new ideas and keywords have been employed for research objectives in recent years.

Figure 12. Shows a map of co-occurrence analysis to form themes through all keywords.

The conceptual framework of this study combines components from the Technology Acceptance Model (TAM), the Theory of Planned Behaviour (TPB), and the SERVQUAL model to create a cohesive theoretical framework for comprehending consumers’ purchase intentions in online food delivery (OFD) applications. These models offer complementary perspectives on the ways in which user acceptability and behavioural intention are influenced by technological, behavioural, and service-quality dimensions. The explanatory power of the model is thereby enhanced by the theoretical linking of each of the identified factors, including time-saving orientation, price-service orientation, social influence, e-service quality, e-security, e-trust, and attitude, to specific constructs within these frameworks. The orientation towards time-saving is intricately connected to the perceived utility construct in the Technology Adoption Model (TAM), since consumers’ conviction that Online Food Delivery (OFD) applications improve their efficiency and convenience is a fundamental factor in technology adoption. When customers recognise that an OFD platform facilitates time and effort savings relative to conventional food buying methods, their propensity for adoption and continued use escalates. Price-service orientation pertains to both the responsiveness and value-for-money aspects of the SERVQUAL model, as well as the attitude construct within the Theory of Planned Behaviour (TPB). Perceptions of equitable pricing and acceptable service cultivate positive assessments of the platform, which influence users’ favourable dispositions and reinforce their decision to use OFD services. Social impact is most well comprehended via the subjective norm element of the Theory of Planned Behaviour (TPB), which asserts that people’s behavioural intentions are influenced by perceived social pressure from family, friends, and online groups. Within the realm of OFD, recommendations, social media presence, and online reviews significantly impact users’ choices to embrace and persist in using these platforms. E-service quality pertains to many elements of SERVQUAL, including reliability, assurance, and tangibility, which represent users’ assessments of the service provider’s consistency, dependability, and interface design. A well-designed, dependable, and responsive OFD application improves user happiness and strengthens behavioural intention.

E-security is a crucial factor, reflecting consumers’ trust in the safety and integrity of online transactions. This construct intersects with confidence in SERVQUAL and perceived ease of use in TAM, as consumers are more inclined to embrace and persist in using systems they see as safe and user-friendly. E-trust serves a dual function in shaping behavioural intention: in the Theory of Planned Behaviour (TPB), it relates to perceived behavioural control, as trust enhances users’ sense of control over online transactions; in the Technology Acceptance Model (TAM), it corresponds to perceived usefulness, as reliable systems are regarded as more advantageous and trustworthy. Therefore, attitude operates as a mediating variable in both the Technology Acceptance Model (TAM) and the Theory of Planned Behaviour (TPB), acting as a significant predictor of behavioural intention. Users who see OFD apps as beneficial, secure, reliable, and of superior service quality are inclined to develop favourable views, which directly correlate with heightened purchase intentions. This research elucidates how technical acceptability, behavioural factors, and service quality characteristics all impact customers’ purchase intentions for online meal delivery platforms.

This study outlines a number of important advantages for professionals and scholars who are interested in online food delivery apps. The thorough depiction of the research ecosystem, which aids in identifying important journals, prominent authors, and prevalent trends, is one of the main benefits. The study identifies well-known journals such as the Journal of Retailing and Consumer Services and the British Food Journal, as well as prominent writers like Dhir Amandeep and Kaur Puneet, offering a firm foundation for further investigation. The data also identifies regional research hotspots, indicating that nations like Finland, Norway, and India are at the forefront of the subject in terms of citations and published articles. This geographic perspective aids in comprehending the worldwide dynamics of OFDA research. The study also identifies the primary research themes, including consumer behavior, satisfaction, and the substantial effects of the COVID-19 epidemic on the industry, by clustering commonly used terms. The charting of theme progression over time is a very useful addition. According to the data, there has been a noticeable change from broad phrases like “food delivery applications” to more focused and modern ideas such as “sentiment analysis” and “customer experience.” A roadmap for future research directions is provided by this evolution mapping, which shows how the subject has diversified and what new fields of study have emerged.

Customers’ inclination to use online food delivery applications is significantly influenced by their attitude, price-service orientation, social influence, e-security, e-trust, and time-saving orientation ([57]; [15]; [5]; [19]). Customers’ propensity to utilize online food delivery applications is significantly influenced by a number of factors, including attitude, price-service orientation, social influence, e-service quality, e-trust, and time-saving orientation ([20], [46]; [15]; [5]; [36]). Customers’ propensity to utilize online food delivery apps is significantly influenced by a number of factors, including attitude, price-service orientation, social influence, e-service quality, e-security, e-trust, and time-saving orientation ([20]; [55]; [57]; [10]; [23]; [31]; [5]; [45]; [54]; [12]; [19]; [26]; [6]; [50]; [34]). Customers’ inclination to utilize online meal delivery applications is significantly correlated with the quality of the e-service ([35]; [16]). Customers’ inclination to utilize online meal delivery applications is significantly correlated with price-service orientation, time-saving orientation, and the E-trust SERVQUAL model ([7]; [18]). By employing bibliometric analysis of the Scopus database to describe the trends of publications on online delivery services, this study sought to add to the earlier findings. This study looked at a range of publications in terms of accessibility, language, subject matter, and source title in addition to the most often cited works, publishing patterns, and authors’ keywords.