A Comprehensive Bibliometric Analysis of Live Streaming Commerce: Mapping the Research Landscape

Systematic Literature Review

PRISMA is a solidified methodologies technique that assures research transparency and reproducibility (Liberati et al., 2009). It has been utilized in live streaming-specific investigations (Y. Li et al., 2020). The protocol comprises four key stages, with the present study utilizing a single source of information—Web of Science (WoS)—for identification. WoS is owned by Clarivate Analytics and includes several databases. This method may restrict document coverage, but it ensures uniformity in citation format, usage, and classification across research fields. In addition, WoS provides a sufficient number of documents with demonstrated scientific quality and offers data cleansing to rectify errors (Harzing & Alakangas, 2016).

Identification, an extensive array of search phrases was meticulously selected based on early study and encompassed a multitude of combinations of search terms: “live streaming,”; “live-streaming,”; “live streaming commerce,”; “live streaming e-commerce,”: “social live streaming”; “internet broadcast”; “network broadcast”; “webcast”; “twitch”; “twitch. tv”; “Youtube Live”; “Douyin live”; “Taobao live.” The resolute determination was made to encompass all published documents inscribed in the eloquent language of English. No categories of documents were omitted, thereby enabling the incorporation of papers germane to scrutiny burgeoning research domains.

The screening process was filtered by incorporating the keywords “commerce,”“e-commerce,” or “e-commerce” to ensure precision. All documents’ titles, abstracts, and keywords underwent a thorough review. Only those records that pertained to live streaming while addressing the specific topic of commerce or e-commerce were retained, whereas irrelevant ones were eliminated. This empowered us to pinpoint any novel items that could have been overlooked during the primary exploration sequence. For eligibility: the document had to be exclusively related to LSC, even if it was a component of a more extensive study. Secondly, it has been indexed in WoS and has undergone some form of peer review. Also, the full-text document should have been effortlessly retrievable. All references satisfying these rigorous requirements were successfully retrieved. The last step is Included: The full-text documents were meticulously scrutinized, except for those associated with living streaming that were not relevant to commerce or e-commerce, those with a vague primary subject matter, those lacking any research data, and those missing methodology from the text. Ultimately, 75 references were obtained through this process, which included records until April 6th, 2022. A comprehensive overview of the entire process is provided in Appendix Figure A1. Appendix Table A1 extracted from main documents about methodological approach, theoretical, and data sources.

Bibliometric Performance Analysis

Bibliometric performance analysis refers to the using of bibliometric methods and tools to evaluate the research performance, impact, and influence (De Bellis, 2009)). Following many prior studies (Arrigo, 2018; Cabeza-Ramírez et al., 2021), the study utilized several indicators, including indexing categories, document types, countries of origin, and journals within the domain under investigation. Most research was conducted in geographical contexts was conducted in China, followed by USA and Australia. Canada, Thailand, and Finland were simultaneously in the third spot based on the geographical contexts in prior studies (see Figure 1 and Appendix Table A1). Figure 2 illustrates an overview of publications concerns about LSC starting to increase in 2017. From 2019, the number of publications grew exponentially, especially with the pandemic rising. Although, there were mostly from the conference (Cai & Wohn, 2019; C. Chen et al., 2019; Z. Chen et al., 2019; Su, 2019; X. Wang & Wu, 2019). The last but the most stage, 2020–2022, represents the explosion stage. It took off during the covid-19 pandemic as widespread lockdowns and restrictions on physical retail stores. Until 2022 the number of publications is expected to decrease due to the present study tracking data only until April 2022. However, even in the middle of the year, publication rates are projected to be higher than those in 2020. The present study primarily analyses the documents issued during the last 5 years to identify and highlight the trends in LSC development.

OPEN IN VIEWER

Figure 1.

Geographical contexts investigated in LSC research studies.

OPEN IN VIEWER

Figure 2.

Number of documents and citation over time (from 2017 to 2022).

Figure 3 displays the most productive institutions, with Zhejiang University and Zhejiang Gongshang University being the primary ones in Zhejiang province. It is worth mentioning that Zhejiang province is considered the capital of live-streaming e-commerce(Qu, 2021). Additionally, the WOS categories indicated a gradual rise in the research areas involved. Intriguingly, as Figure 4 shows, the co-occurrence analysis of disciplinary categories divulges that LSC is a multifaceted and all-encompassing research topic encompassing numerous disciplines and garners significant scholarly interest. Notably, Table 1 ranks both business and computer science information systems at the top with an impressive count of 24 documents. Followed by computer science interdisciplinary applications with 17 documents, then management 12, Computer science theory methods 11. Besides, psychological multidisciplinary (6), computer science artificial intelligence (5), engineering electrical electronic (4), computer science cybernetics (3), computer science software engineering (3), and environmental science (3) related to LSC. Even though live streaming services are heavily commercialized, the business documents were found to be on par with other computer science information systems. This suggests a promising avenue for future research where innovative techniques from management or economics can be employed to supplement prior discoveries. The same holds for other uncharted domains like environmental science and software engineering in computer science.

OPEN IN VIEWER

Figure 3.

Most productive institutions WOS.

OPEN IN VIEWER

Figure 4.

Most productive institutions in WOS categories (no. publications ≥ 3).

OPEN IN VIEWER

Table 1.

General Profiles of Research on LSC.

Category	Items
WOS categories (No. publications ≥ 3)	Business (24); computer science information systems (24); computer science interdisciplinary applications (17); Computer science theory methods (11); management (12); information science library science (6); telecommunications (8); psychological multidisciplinary (6); computer science artificial intelligence (5); engineering electrical electronic (4); computer science cybernetics (3); computer science software engineering (3); environment science (3);
Document types	Articles (54); proceeding papers (19); early access (2)
Most prolific authors (no ≥ 2)	Lu, Jinxuan (2); Cai, jie (2); Assarut, Nuttapol (2); Wohn, Donghee Yvette (2); Zhou, Wanhua (2); Chen, Zhenjiao (2); Cheng,Sijia (2); Li, Qi (2); Nie, Kun (2); Guo, Lingyun (2); Sun,Yuan (2); Toehonen, maria (2); Wongkitrungrueng, Apiradee (2); Widowati, Retno (2); Shao, Xiang (2); Liu, xiaotong (2); Guo,Yue (2); Hong,Yilli (2); Hamari, Juho (2);
Most productive publication titles ≥ 3	Electronic commerce research and application 5; internet research 4; frontiers in psychology 3; proceedings of the 52nd annual hawaii international conference of system sciences 3;
Most productive institutions ≥ 3	Zhejiang university (4); ALIBABA Group (3); Beijing university of posts telecommunications (3); Chinese academy of sciences (3); Hebei university of technology (3); Tamkang university (3); Tampere university (3); university of international business economics (3); Zhejiang Gongshang university (3);
Most productive countries (≥3)	Peoples r china (58); USA (14); Taiwan 10; Thailand (3); Australia 4; Canada 3; Finland 3
Core references (no. of received citations (≥30)	197 Wongkitrungrueng and Assarut (2020); 174 Sun et al. (2019); 99 Kang et al. (2021); 90 Q. Zhao et al. (2018); 88 M. Hu and Chaudhry (2020); 84 Hou et al. (2019); 75 X. Xu et al. (2020); 72 Wongkitrungrueng et al. (2020); 69 Cai et al. (2018); 62 Zhou et al. (2019); 52 B. Lu and Chen (2021); 47 Y. Li et al. (2021); 37 (L. Guo et al., 2021); 31 (Addo et al., 2021); 30 (M. Zhang et al., 2022)

Science Mapping

Crafting a visual representation of the intricate connections between diverse publications, authors, journals, disciplines, or research areas is commonly referred to as science mapping. The application of scientific network analysis has proven to be an effective tool in vividly showcasing these links across various domains (C. Chen, 2017). The network analysis meticulously scrutinized the document’s content by examining keyword co-occurrence. This approach was considered fitting for the early identification of a given research domain’s intellectual structure, as emphasized by X. Chen et al. (2019). Additionally, keywords, titles, and abstracts play a critical role in selecting relevant themes Choi et al. (2011). The resulting network of relations between terms was visualized using VOSviewer (van Eck & Waltman, 2010) and Cite-Space (C. Chen, 2006); both two are open-source software that calculates centrality and strength to reveal the theme of the domain. Nodes represent keywords, with their size indicating importance, while links depict how often keywords appear together, with thickness representing link strength.

In order to ensure a comprehensive visualization of all identified thematic networks, a minimum of n = 2 occurrences was set for the entire domain, given that it is still emerging and only 75 documents were available. A thesaurus file was employed to refine further the data and group synonymous terms or singular/plural forms of certain words. For example, electronic sports, e-sports, and e-sport games were consolidated under the term e-sport. Additional parameters utilized in VOSviewer can be found in Appendix Table A2.

The Domain of Research on LSC

The previous discussion hinted at earlier research endeavors that sought to explore the boundaries of subject matter in order to establish the scope of investigation for LSC. As proposed by Shannon (1948), involves posing four fundamental inquiries: what, how, who, and where? The answers gleaned from these queries enhance comprehension of the phenomena under scrutiny. Subsequently, this literature review addresses the “what” and “how” aspects. The present study alluded to prior studies that aimed to examine the limits of topic areas in order to demarcate the LSC study domain. According to Shirmohammadi et al. (2021), it comprises asking four basic questions: what, how. The responses help to improve knowledge of the phenomena. The following review aids in answering the “what” and “how” questions.

Regarding the phenomenon’s essence, it is noteworthy that implementing of the PRISMA protocol resulted in a remarkable 75 documents. What is more interesting is that all these documents have one thing in common—they delve into the communication process arising from diverse definitions and meanings developed concerning LSC. Shannon (1948) defined the transmission of messages—these messages serve as sources of information that reach their intended destination. Once received, the audience or recipient can share information like engaging conversations, gifts, subscriptions, followers, and purchasing activity with each other or back to the initial source. Moreover, it is crucial to highlight that the transmission of messages involves intricate encoding and decoding mechanisms facilitated by various platforms that serve as intermediaries between senders and recipients. Thus, the present study divided the documents into three groups: source information (streamers), platform, and viewers. Firstly, platform in live streaming platforms are divided into types of platforms (Gong et al., 2020): There are three main ways in which live streaming is being incorporated into commercial activities: firstly, through live streaming platforms that integrate with commerce (such as TikTok); secondly, through e-commerce sites and marketplaces (like T-mall) or mobile apps that include live streaming features; and thirdly, through social networking sites (SNSs) that add live-streaming capabilities to facilitate selling (such as Facebook Live). The key benefit of these approaches is the ability for broadcasters to interact with products in real-time while transmitting their expressions and interactions directly to viewers and customers who may be located elsewhere. This creates a sense of direct synchronous communication between the broadcaster and customer, allowing for immediate feedback via written comments or purchase behavior.

In terms of “how,” the cumulative research is built using multiple theoretical frameworks. Although many publications did not offer the theory they used, they were mostly based on earlier research. It is feasible to identify well-established frameworks that are often employed. Overall, three specialized domains can be recognized from the 75 texts. The publications concentrated mainly on the platform’s usage of well-established concepts such as theory of reasoned action. (D. Li et al., 2018), attachment theory (Y. Li et al., 2021), use and gratifications theory (Jia et al., 2022), Integrating mean-ends chain of lifestyle theory, socialized charismatic leadership theory, and product uncertain theory (Z. J. Chen et al., 2019); social evolutionary game theory (Lv et al., 2022); rough set theory (S. H. Liao et al., 2021). These studies mostly focused on providing the development of applications, for example. Functions provided by live platforms make viewers engaged. Even some studies sought to improve transmission or platform solutions and do not need a firm grounding theory. Such studies come from computer science and engineering, such as F. L. Li et al. (2021, Z. Li et al. 2021).

The second group, which was focused on the receivers/reviewers/consumers, was underpinned by the widely-tested theories: SOR theory (Chandrruangphen et al., 2022; Lee & Chen, 2021; Ma et al., 2022; Ming et al., 2021; theory of para-social interaction (W. Zhang et al., 2021); elaboration likelihood model (Gao et al., 2021); social presence theory (T. Chen et al., 2021; Geng et al., 2020; M. Li & Hua, 2021; Z. Liu et al., 2020); self-efficacy theory (Cao et al., 2022; Gong et al., 2020); social cognitive theory (M. Li & Hua, 2021); information forging theory (X. Wang & Wu, 2019); social learning theory (M. Li & Hua, 2021). The last sub-domain group focused on the streamer employed well-established theories such as the signal theory (B. Lu & Chen, 2021), ground theory (Luo, Cheng, Zhou, Yu, & Lin, 2021); motive theory (W. Wang et al., 2021); social capital theory (P. Xu et al., 2022); social exchange theory (M. Zhang et al., 2021). Table 2 illustrates the comprehensive domain and the widely-accepted theory for enhancing the knowledge base of research on LSC.

OPEN IN VIEWER

Table 2.

Main Theories Applied.

Streamer	Receivers/viewers	Platform
Signaling theory Ground theory Motive theory Social capital theory Social exchange theory SOR theory	SOR theory Theory of para-social interaction Elaboration likelihood model Social presence Self-efficacy theory Social cognitive theory Information forging theory Social learning theory Affordance theory Theory of flow Attachment theory Charismatic leadership theory	Theory of reasoned action Information uses and gratifications theory Integrating mean-ends chain of lifestyle theory. Socialized charismatic leadership (SCL) theory Product uncertainty theory Social evolutionary game theory Rough set theory

Regarding the methodology, Table 3 shows the main approaches and methodologies applied in all documents. For the areas of consumer/viewer studies, quantitative research prevails, and many studies using structural equation models by adopting primary data (such as surveys and questionnaires). Sampling size range from 214 (Z. Liu et al., 2020) to more than 900 (W. Zhang et al., 2021). Regarding the area of streamers studies, also quantitative methods predominate; these studies adopted massive data by scraping data on live streaming platforms. Additionally, some of using structural equations model methods by exploring characteristics of streamers’ impact on purchase intention. The subdomains focused on the platform are characterized by using more heterogeneous methods for compiling information through application programing interfaces.

OPEN IN VIEWER

Table 3.

Main Methodology Applied.

	Consumers/viewers	Streamers	Platform
Qn	Primary data (survey) (Wongkitrungrueng & Assarut, 2020); (Hou et al., 2019);(Cai et al., 2018);(M. Hu & Chaudhry, 2020); (X. Xu et al., 2020);(Cai & Wohn, 2019); (X. Wang & Wu, 2019); (L. Guo et al., 2021); (Ming et al., 2021); (Lee & Chen, 2021);(W. Zhang et al., 2021); (Gao et al., 2021); (Z. Liu et al., 2020);(Chandrruangphen et al., 2022); (Cao et al., 2022); (M. Li & Hua, 2021); (Ma et al., 2022);(Y. Xu et al., 2021); (Y. Zhao & Bacao, 2021); (J. Guo et al., 2021); (H. Chen et al., 2021); (T. E. Hu et al., 2022); (Z. Wang et al., 2022); (Fengliang & Jianhong, 2021)	(Q. Zhao et al., 2018); (Y. Guo et al., 2022);(P. Xu et al., 2022); (Hsieh & Zhuo, 2021)	(Sun et al., 2019); (D. Li et al., 2018); (Y. Li et al., 2021); (M. Zhang et al., 2022); (Gong et al., 2020)
	Second hand data: (Kang et al., 2021) (Addo et al., 2021); (S.-H. Liao et al., 2022a); (Luo et al., 2021); (L. Li & Kang, 2020)	(Wongkitrungrueng et al., 2020); (Geng et al., 2020);(M. Zhang et al., 2021);(Törhönen et al., 2021); (Luo, Cheng, Zhou, Yu, & Lin, 2021);	(T. Chen et al., 2021); (S. H. Liao et al., 2021); (He et al., 2021); (W. Wang et al., 2021);
Qt	(Z. J. Chen et al., 2019); (C. Y. Lu et al., 2020);(J. Chen et al., 2020)	B. Lu and Chen (2021);(Jia et al., 2022)
M		(Jiang & Cai, 2021); (J. Liu et al., 2021); (L. Wang et al., 2022); (C. Chen et al., 2019)	(L. Guan et al., 2022); (Lv et al., 2022); (J. Liu et al., 2022);
Tc			(Zhanikeev, 2015); (Zhou et al., 2019); (F. L. Li et al., 2021; Z. Li et al., 2021); (Yu et al., 2021); (Kim & Park, 2012)

Thematic Structure of the Research Domain

In the upcoming session, the present study aims to conduct time-zone mapping by utilizing Cite-Space and keywords analysis techniques to identify and explore the primary emerging research topics on a yearly basis, while also examining their prominence. Furthermore, the present study will investigate the emergence of research hotspots by utilizing VOS-viewer to calculate the total strength of co-occurrence links between terms.

Dynamic Evolution of Research Topics

The cite-space analyze all 75 documents’ keyword and terms. Figure 5 illustrates the time-zone mapping of keywords; the current investigation has the potential to unveil the historical progression of research priorities and delve deeper into emerging trends in e-commerce live streaming. Before 2019, the main topics revolved around information communication, technology adoption, booming industry, and behavior intention. During this period, LSC was still a relatively new concept, and research focused on its effectiveness as a marketing strategy. Notable research points included streamer broadcasting motivation (Q. Zhao et al., 2018), platform adoption motivation (D. Li et al., 2018), and live-streaming behavior (Cai et al., 2018). From 2019 to 2020, primary research areas shifted toward motivation, purchase behavior, social presence, and customer engagement, with China emerging as a popular trend. Looking ahead to 2022, attitudes like trust and satisfaction took center stage while exploring new topics like credibility and impulse buying behavior. In 2022, the focus will be on celebrity endorsements type alongside customer engagement strategies and product quality aimed at creating attractive live-streaming interfaces.

OPEN IN VIEWER

Figure 5.

The dynamic keywords related to LSC research.

The current research was left with 466 keywords after using the VOSviewer thesaurus file to filter them. Only 97, however, matched defined threshold of at least two occurrences. After that, the program computed the co-occurrence connection strength between words and chose those with the greatest overall link strength. Figure 6 and Table 4 show the whole list, and theme clusters, linkages, and occurrences. To designate each cluster, sample terms were picked. While certain general phrases connected directly to the domain were included in these clusters, such as live streaming commerce, e-commerce, management, live stream, social media, and business, each instance had a distinct thematic focus. The present study analyzed the aims and results of all papers in our sample set to validate findings from this theme cluster analysis.

OPEN IN VIEWER

Figure 6.

Thematic clusters identified in the research domain on the LSC field through VosViewer.

OPEN IN VIEWER

Table 4.

Keywords Identified in Each Cluster.

Cluster/Color/Label	No. Keywords	Keywords (links, total link strength, occurrences)
C1/red / (Behavior)	19	Behavior (51, 78, 11); live streaming commerce (50, 83, 13); social presence (35, 45, 6); quality (30, 37, 5); engagement (30, 43, 5); word-of-mouth (29, 42, 7); intention (29, 31, 4); celebrity (13, 15, 2); Commerce (28, 33, 5); environments (15, 16, 2); evolution (10, 11, 2); information quality (19, 24, 3); involvement (7, 7, 2);Media (24, 31, 4); strategies (19, 25, 3); perspective (16, 16, 2); source credibility (13, 15, 2); social e-commerce (11, 11, 3); mechanism (6, 6.3);
C2/green/ (Credibility)	17	Live streaming (59, 101, 19); social commerce (54, 83, 13); social media (42, 58, 7); online (52, 85, 12);credibility (27, 32, 4); loyalty (24, 28, 4); Communication (15, 16, 2); content marketing (10, 10, 2); customer trust (16, 20, 3); E-Wom (15, 16, 2); image (18, 19, 2); influencer (18, 21, 2);perceptions (9, 9, 2); price (14, 15, 2); purchase behavior (10, 10, 2); self-disclosure (15, 18, 2); Facebook (16, 18, 2);
Cluster3/blue (Trust)	17	Trust (55, 99, 13); e-commerce (42, 58, 9); motivations (28, 37, 7); B2c-e-commerce (24, 31, 4); live streaming (22, 25, 5); utilitarian (18, 19, 3);cross-border e-commerce (12, 15, 2);e-commerce gamification (11, 13, 2); gift (11, 13, 2); information technology (9, 11, 2); live streamer (14, 14, 2);live streaming shopping (8, 8, 2); popularity (6, 6, 2); roles (13, 13, 2); shopping value (12, 12, 2); technology acceptance model (3, 3, 2); user acceptance (15, 18, 3);
Cluster4/yellow (Adoption)	14	Model (45, 62, 9); adoption (37, 51, 7); role (31, 39, 5); satisfaction (29, 33, 5); antecedents (28, 34, 4); Acceptance (16, 17, 2); flow theory (16, 16, 2); continuance intention (17, 17, 2); covid-19 (9, 9, 2);flow (10, 10, 2); flow experience (9, 10, 2); flow experience (9, 10, 2); mediating perceived value (13, 13, 2); psychological distance (15, 16, 2); system (12, 13, 2); technology acceptance (21, 23, 3)
Cluster5/purple (Costs)	11	Impact (67, 116, 15); business (24, 28, 3); costs (19, 20, 3); supply chains (17, 19, 2); pricing (17, 19, 2); Brand (17, 17, 2); channel (9, 10, 2); electronic commerce (17, 18, 3); self (16, 17, 2); supply chain (3, 3, 2); uncertainty (14, 17, 4);
Cluster 6/light blue (Information)	9	Information (44, 66, 9); interactivity (22, 25, 4); PLS-SEM (16, 17, 2); products (17, 17, 2); repurchase intention (17, 17, 2); similarity (14, 18, 2); stimuli (9, 10, 2); SEM (15, 16, 2); swift guanxi (19, 21, 3)
Cluster 7/orange (Purchase intention)	6	Purchase intention (55, 98, 13); technology (29, 34, 5); affordance (16, 17, 2); Consequences (16, 16, 2); customer engagement (15, 19, 4); experience (13, 14, 2)
Cluster 8/brown (knowledge)	4	Management (34, 39, 4); knowledge (23, 27, 3); adoption intention (16, 17, 2);

The thematic cluster associated with behavior (C1), credibility(C2), trust (C3), adoption(C4), costs (c5), information(C6); purchase intention(C7), knowledge (C8), shown Table 4. Subsequently, the main topics on LSC obtained in each cluster are detained. The first cluster covers topics related to the behaviors of consumers and streamers. This complex behavior, for example, the intention to purchase (Sun et al., 2019), continuous watching (Hou et al., 2019), engagement brand activity (M. Hu & Chaudhry, 2020), virtual gift (Z. Guan et al., 2022; Zhou et al., 2019), impulse buying behavior (Lee & Chen, 2021), adoption intention (Z. Liu et al., 2020), interaction behavior (McLaughlin & Wohn, 2021). Even streamer continuance broadcasting behavior (Q. Zhao et al., 2018). The cluster deeply connects with the antecedent of behavior. Thus, occurrences keywords involve the motivations of behavior, including social presence (Z. Liu et al., 2020) and information quality (M. Zhang et al., 2021).

The second cluster primarily focuses on credibility, particularly in the context of LSC. Research has shown that social influence factors like social presence, perceived expertise, and perceived similarity between streamers and viewers can significantly impact the perceived credibility of content and platforms (Y. Guo et al., 2022; P. Xu et al., 2022). Additionally, studies have explored how visual cues such as video quality and overall esthetics affect credibility (M. Zhang et al., 2021). The cluster also highlights price as a significant factor influencing purchasing behavior during live-streaming sessions. Future research should investigate pricing strategies like discounts or coupons to optimize both seller profitability and consumer benefit (Cao et al., 2022). as well as explore the development of an optimal/equilibrium price strategy (L. Guan et al., 2022; J. Liu et al., 2022).

The third cluster brings topics related to trust, covering the need for streamers to build and maintain audiences (D. Li et al., 2018; Wongkitrungrueng & Assarut, 2020). It also involves offering interactivity and facilitating communication with viewers, thereby the trust between them and their audience (Lai et al., 2021). The cluster provides a fresh view that distinguishes it from traditional e-commerce. Research in this area has explored the impact of gamification, such as vertical gifting, and rewards, on engagement, motivation, and retention of viewers in live-streaming platforms (T. E. Hu et al., 2022; Zhou et al., 2019). Utilitarians in LSC, such as convenience, efficiency, and access to a wide range of products, may be more likely to use continuous shopping or watching in the live streaming context (Cai et al., 2018).

The fourth cluster link to adoption. The topic is thus intricately linked with utilizing of advanced technological tools, predominantly on the streamer’s end. It provides a view of the antecedents of adopting new technologies or shopping channels. One subdomain explains that flow experience is relevant in the context of LSC adoption because it suggests that the more enjoyable and engaging the live stream is, the more likely consumers are to adopt it as a new shopping channel (C.-C. Chen & Lin, 2018; Z. Liu et al., 2020; Z. Wang et al., 2022); while another subdomain investigate the rising adopting during the COVID-19 (T. Chen et al., 2021).

The fifth cluster highlights the cost topic. Regarding the subject matter depicted by the chosen keywords, “uncertain,”“pricing,” and “supply chain.” Several studies addressed specific topics such as the type of streamers due to the unsustainability and exorbitant cost associated with celebrity hosts; Several companies have opted for their own branded broadcasters as a means of attracting followers (W. Wang et al., 2021). Additionally, streamers form a complete supply chain by selecting products, estimating sales, purchasing inventory, preparing it for sale, delivering it to customers, and even providing after-sale support (Jiang & Cai, 2021).

The sixth cluster pertains to the keyword “information.” The overarching theme revolves around information elements rooted in SOR (stimulus-organism-response) theory by adopting the SEM model (M. Hu & Chaudhry, 2020; Ming et al., 2021; X. Xu et al., 2020). This encompasses a co-created social experience that fosters dynamic interactions between streamers and viewers (Kang et al., 2021). Also, Live streaming offers a wealth of product information on quality and functionality, enabling consumers to better comprehend the products and dispel any doubts about their authenticity. Thereby enhancing transparency and building trust with consumers (Gao et al., 2021). The final two clusters comprise “purchase intention” and “knowledge.” Interestingly, the seventh cluster’s keyword is detected across all clusters as it represents the primary consequence during LSC. In the case of cluster 8, it can be inferred that streamers are driven by a strong urge to disseminate information about the available tools and sustain viewer engagement (P. Xu et al., 2022).