To ensure the transparency and reproducibility of the review, this study developed a clear protocol covering data sources, search strategy, inclusion and exclusion criteria, screening procedures, deduplication, data extraction and quality assessment. The specific procedures are reported as follows.

Data sources. Chinese-language literature was retrieved from the China National Knowledge Infrastructure (CNKI), the most comprehensive academic database in China, and served as the basis for the bibliometric analyses presented in Figures 2-5. As international research on user satisfaction with short-form video platforms is comparatively limited and dispersed, the relevant international (TikTok) studies were identified through supplementary database searches and manual tracing of the reference lists of the retrieved domestic studies, and were incorporated into the narrative synthesis in Sections 2-5.

Search strategy. The literature search was completed in 2026. For the CNKI database, subject-term retrieval was adopted, combining platform-related keywords (such as “抖音” and “短视频”) with satisfaction-related keywords (such as “满意度” and “用户满意度”) by means of the Boolean operator “AND”. As Douyin was officially launched in 2016 and relevant research began to emerge from 2017, the retrieval time span was set from 2017 to 2026, consistent with the publication trend shown in Figure 2. The retrieved records were restricted to Chinese- and English-language academic literature.

Inclusion criteria. Studies were included if they: 1) took Douyin, TikTok or

Figure 1. Technology road map.

other short-form video platforms as the research object; 2) explicitly examined user satisfaction or its influencing factors; 3) were peer-reviewed journal articles or formally completed degree theses; 4) reported original empirical data or systematic theoretical analysis; and 5) were written in Chinese or English within the specified time span.

Exclusion criteria. Records were excluded if they: 1) were non-research documents, such as news reports, editorial notes, conference abstracts, book reviews or promotional materials; 2) mentioned short-video platforms only incidentally without treating user satisfaction as a core research variable; 3) were duplicate publications of the same study; or 4) lacked an accessible full text.

Screening and deduplication. All retrieved records were exported and managed using EndNote. Duplicate records were first removed automatically with the built-in deduplication function of EndNote and then checked manually. Two reviewers independently screened the titles and abstracts of the remaining records against the inclusion criteria, and disagreements were resolved through discussion or, where necessary, adjudication by a third reviewer. The full texts of the potentially eligible records were subsequently retrieved and assessed. After the removal of duplicates and full screening, a total of 193 records were ultimately included for the bibliometric analysis presented in Figures 2-5, whilst the international studies identified through supplementary search and citation tracing were incorporated into the narrative synthesis.

Data extraction and synthesis. A standardised data extraction form was used to record the publication year, research platform, theoretical framework, research method, sample characteristics, core variables and main findings of each included study. Owing to the considerable heterogeneity of the included studies in terms of theoretical frameworks, survey samples and econometric methods, a narrative synthesis approach was adopted. The included studies were first classified by platform (TikTok versus Douyin) and then by research theme (algorithmic recommendation, content quality, platform services, risk perception and psychological exhaustion), after which the findings were compared and integrated to identify points of convergence and divergence. To avoid conflating evidence across platforms, the following rules were applied consistently: i) findings derived from studies of Douyin were treated as evidence for the domestic platform; ii) findings derived from studies of TikTok were treated as evidence for the international platform; iii) a conclusion was described as cross-platform only when it was supported by evidence from both platforms, and platform-specific conclusions were extended to the other platform only tentatively and were explicitly labelled as such.

Quality assessment. The methodological quality of the included empirical studies was appraised using the Joanna Briggs Institute (JBI) critical appraisal checklists for cross-sectional studies, adapted to the specific design of each study (for example, questionnaire surveys, structural equation modelling and fuzzy comprehensive evaluation). Each study was rated as being of low, moderate or high risk of bias. Studies judged to be of critically low quality were excluded from the narrative synthesis, and the quality ratings of the remaining studies were used to calibrate the confidence placed in individual findings.

Before synthesising the relevant literature, it is necessary to clarify the concept of user satisfaction and to distinguish it from several related constructs with which it is frequently conflated in the reviewed studies—namely continuance intention, loyalty, stickiness, burnout and churn. The definitions below, and the distinctions drawn between them, are applied consistently throughout the review.

User satisfaction. In line with the ACSI-based analyses in the reviewed literature ([11]), user satisfaction is defined in this study as the overall evaluative judgment that users form after using a short-form video platform, reflecting the degree to which their prior expectations and needs are fulfilled by the platform’s algorithmic recommendations, content quality, platform services and related experiential dimensions. It is therefore a post-usage, attitude-like evaluation, rather than a behavioural intention or a behavioural outcome.

Continuance intention. Continuance intention refers to users’ subjective willingness or intention to continue using the platform in the future, and is a typical behavioural-intention variable ([25]; [20]; [10]; [17]). It is conceptually distinct from satisfaction: satisfaction captures an evaluative judgment of the past and present experience, whereas continuance intention captures an expectation of future behaviour.

Loyalty. Loyalty denotes users’ deep commitment to a platform, manifested in repeated use, a positive attitude and resistance to switching to competing platforms; in the reviewed literature it is often operationalised through customer loyalty, retention and recommendation behaviour ([28]; [27]). Loyalty is a broader and longer-term construct than satisfaction, and satisfaction is generally regarded as a necessary but not sufficient condition for loyalty.

Stickiness. Stickiness refers to the extent to which users repeatedly visit, remain on and habitually return to a platform, reflecting the strength of the platform’s hold over users’ browsing time and attention; in the reviewed literature it is closely associated with flow experience and user retention ([9]). Whereas satisfaction is an evaluative judgment, stickiness is primarily an engagement-related behavioural indicator.

Burnout. Burnout (also referred to as fatigue in the reviewed literature) denotes a negative psychological state that develops through prolonged or intensive platform use, encompassing mental fatigue, information overload, social overload and the perception of wasted time ([29]). It is an experiential cost that coexists with positive evaluations in short-video use, and is conceptualised in the reviewed studies both as an antecedent associated with lower satisfaction and as an outcome of prolonged use ([13]; [29]).

Churn. Churn refers to the actual behaviour of discontinuing platform use, such as uninstalling the application or abandoning the platform ([6]). The reviewed literature distinguishes churn from the intention to discontinue (a behavioural intention) and from burnout (a psychological state); churn represents the final behavioural manifestation in the chain from dissatisfaction and burnout to discontinuance ([13]; [6]).

In summary, the five related constructs and user satisfaction can be ordered along a conceptual chain: user satisfaction is a post-usage evaluative judgment; continuance intention, loyalty and stickiness are behavioural intentions or engagement states that satisfaction is hypothesised to promote; burnout is a negative psychological state that is hypothesised to erode satisfaction; and churn is the eventual behavioural outcome. Accordingly, throughout the review the findings of each study are interpreted according to the specific construct that the study actually measured, and conclusions concerning one construct are not extended to another unless the source study itself establishes such a link.

Within the research framework on user satisfaction with short-form video platforms, the international academic community has taken the lead in conducting multi-dimensional explorations centred on TikTok, with the overall scope of research covering eight key areas: algorithmic mechanisms, perceived duration, parallel psychological costs, social interaction experiences, short-form video e-commerce services, advertising effectiveness, negative user behaviour, and customer retention ([12]; [19]). As the research framework has continued to mature, relevant overseas studies have gradually developed a mixed empirical research paradigm centred on SEM and fsQCA; furthermore, fuzzy comprehensive evaluation, grounded theory and multi-dimensional combination evaluation models have also found very widespread application in this field.

Through extensive empirical analysis, the international academic community has gradually reached a consensus that there is a clear trade-off between users’ perceived usage duration and the psychological cost of multitasking; consequently, negative psychological experiences such as the perception of wasted time and mental fatigue are associated with progressively lower enjoyment of the platform ([29]). Furthermore, privacy risks and information overload are also key correlates of negative user behaviour; both are primary antecedent variables negatively associated with user satisfaction ([13]). Algorithmic recommendation mechanisms exhibit a distinct duality in short-video usage contexts: whilst they are associated with an improved content-matching experience for users, they are also associated with heightened concerns regarding privacy breaches; moreover, platform reputation plays a significant moderating role between these two sets of positive and negative associations ([17]). Overseas empirical research focusing on the short-video e-commerce sub-sector further confirms that content quality and social connections can influence customer loyalty by enhancing user stickiness; consequently, user satisfaction is regarded as the cornerstone for ensuring long-term user retention ([27]; [9]). In addition, the international academic community has conducted extensive, specialised research into social media advertising, focusing on analysing the various factors influencing digital customer experience ([27]), negative user interactions ([13]) and complex user loyalty ([28]), thereby further enriching the research dimensions of short-video user satisfaction.

Research in this field can be divided into three phases (Figure 2). During the embryonic phase (2017-2019), the total number of annual publications was fewer than 20, with only a small amount of basic research on online videos and virtually no empirical studies specifically on Douyin. The period of rapid growth (2020-2024) saw Douyin’s commercialisation gradually mature, whilst the user base for short videos expanded and the annual number of publications rose considerably overall, despite a slight decline between 2021 and 2022; literature from this phase accounts for approximately 51 per cent of all relevant research outputs. The period of refined and in-depth exploration (2025-2026) saw a slowdown in the growth

Source: China National Knowledge Infrastructure (CNKI).

Figure 2. Trend chart showing the volume of posts on the Douyin platform regarding satisfaction with China National Knowledge Infrastructure (CNKI) from 2017 to 2026.

rate of publications, but research topics became more specialised. Papers on cutting-edge themes such as AIGC short videos, advertising disclosure and situational narrative emerged in large numbers, reflecting how, against a backdrop of intense competition within the industry, the focus of academic research shifted from basic satisfaction assessments to debates over the mechanisms of emerging technologies and commercial scenarios. These changes in publication trends indicate that shifts in academic research priorities evolve alongside the development of the short-video industry’s market landscape. Early research focused on basic user experience, whilst in recent years the focus has shifted to complex user perceptions characterised by the coexistence of multiple contradictions; this also provides the historical context for the emergence of significant research divergences within the academic community.

The scope of research into satisfaction with domestic short videos continues to expand, with econometric methods such as multiple regression ([7]), correspondence analysis ([21]) and fuzzy comprehensive evaluation ([3]) being widely adopted. With the industrialisation of AIGC, the increasing standardisation of short-video advertising regulation, and the widespread adoption of scenario-based short videos, domestic research has given rise to three emerging sub-directions: Firstly, specialised studies on AIGC short-video satisfaction have introduced new evaluation dimensions, including the originality of AI-generated content, the authenticity of information, and the quality of visual storytelling ([16]); secondly, empirical studies on advertising disclosure based on persuasion theory have been conducted to verify the interactive effects of advertising transparency and users’ uncertainty-avoidance traits on user experience and consumption willingness ([18]); thirdly, modelling and analysis of scenario-based short videos, grounded in scenario theory and communication theory, have confirmed that scenario-based narration, emotional rendering and story-driven expression evoke emotional resonance among users, whilst narrative logic, visual presentation and plot plausibility directly determine user satisfaction levels ([2]).

Academic circles have conducted extensive discussions on user churn in the short-video sector. Relevant studies indicate that information overload, upward social comparison, privacy concerns and usage costs are key correlates of users’ intention to abandon platforms ([6]). The conclusions drawn from different survey samples vary in their focus; large-scale, nationwide surveys tend to emphasise objective usage and external disruptive factors, thereby confirming that actual usage duration, perceived content value and the level of advertising disruption are key correlates of satisfaction ([21]). However, regional surveys adopt a more nuanced perspective; most such studies conduct comprehensive assessments by categorising short-video user satisfaction into four main dimensions: media experience, social interaction, security and safety, and functional services ([23]). In addition to platform experience factors, demographic characteristics also give rise to certain user differences. Age, educational attainment and income level not only significantly differentiate levels of user satisfaction but are also associated with users’ willingness to churn and their content preferences; however, the heterogeneous associations involving gender are generally weaker ([3]). Building on this, He Qian’s study further validated that gender differences have a distinct impact on user experience in terms of perceptions of advertising intrusion and platform security ([7]). Compared to studies focusing on external characteristics, some research draws upon classical models to delve deeper into core driving mechanisms. For instance, Li Ruiyu conducted an empirical analysis based on the ACSI model, thereby confirming that the level of expectation fulfilment and perceived value can positively drive an increase in user satisfaction ([11]). As research has progressed, the domestic academic community has moved beyond the analysis of single influencing factors to focus progressively on the interplay between negative user interactions, platform fatigue, and the relationship between user retention and satisfaction, thereby accumulating a wealth of multidimensional empirical research findings ([13]; [9]).

Research topics related to short videos exhibit a distinct tiered distribution (Figure 3).

The first tier comprises core foundational topics. The number of publications on “short videos” and “short video platforms” is far ahead of the rest, at 62 and 36 respectively; these are the most mainstream research directions in the field. Most related studies conduct empirical analyses based on classical models such as the Technology Acceptance Model (TAM) and Expectation-Confirmation Theory, forming the foundation of research in this field. Among these, Liu Mingzheng ([14]) utilised the D&M Information Systems Success Model to establish an evaluation system for user-generated content (UGC) satisfaction in short videos, providing a standardised indicator framework for the comprehensive assessment of short video platforms.

The second tier comprises long-standing hot topics. Keywords such as “mobile short videos”, “research on influencing factors”, “Douyin short videos”, “reading promotion” and “new media” have maintained a stable volume of publications.

Source: China National Knowledge Infrastructure (CNKI).

Figure 3. Distribution of the main research themes.

Among these, “research on influencing factors” continues to attract academic attention, echoing the six major points of contention identified in this study; areas such as marketing strategies, communication effectiveness and user experience have likewise formed stable research clusters.

The third tier comprises niche and emerging topics. The number of publications on keywords such as “optimisation strategies” and “short-form video marketing” remains relatively low. Based on an analysis of the research landscape in this field, cutting-edge topics such as AIGC and situational short videos have seen rapid growth in the volume of literature in recent years; however, the overall research volume remains small, with studies on each topic being rather fragmented and no unified analytical framework having yet been established.

A comprehensive review of existing research reveals distinct differences in research focus between domestic and international academia. Whilst overseas research largely centres on TikTok, focusing on issues such as users’ psychological costs and usage fatigue, as well as user churn and e-commerce retention, domestic research is more closely aligned with the practical development of local platforms, concentrating primarily on content operations and advertising communication on short-video platforms, as well as explorations of local user behaviour.

Looking at the overall distribution of research platforms (Figure 4), there were a total of 19 papers related to short-video platforms; this category represents the most frequently cited secondary theme, thereby fully confirming that short-video platforms are the core research platforms in this field. Focusing on the current state of application of traditional theories reveals that there are three studies each centred on the creator’s perspective and grounded theory. In addition, classic user behaviour theories such as the flow experience and the Five-Factor Model, as well as personalised recommendations and perceived usefulness, each has two research

Source: China National Knowledge Infrastructure (CNKI).

Figure 4. Distribution of sub-themes.

studies. Consequently, current empirical research largely relies on established models such as the Technology Acceptance Model (TAM), flow theory and grounded theory for analysis. Chen Xi’s study, which analyses the psychological needs of Douyin users through the lens of use and satisfaction ([5]), represents a classic example of user psychology research. In contrast, when examining research progress on cutting-edge theories, emerging theoretical and technological perspectives—such as artificial intelligence and usage motivation, which each have two corresponding research papers, and the 4I theory and user profiling, which each have one—remain relatively limited. This also indicates that the overall number of studies combining cutting-edge theories with real-world short-video scenarios remains relatively low; consequently, there remains a significant research gap in this field regarding emerging research perspectives.

Short videos possess cross-media and cross-disciplinary characteristics; there are marked differences in research focus across different disciplines, and the disciplinary distribution is characterised by extreme imbalance (Figure 5).

With 91 articles, the field of journalism and media studies accounts for 47.15 per cent of the total, making it the dominant field in short video research; the inherent media and communication attributes of short videos determine the leading role of the journalism and communication field in this research.

Library and Information Science and Digital Libraries (15 papers, 7.77 per cent) and Business Economics (11 papers, 5.70 per cent) form the second tier, with library knowledge dissemination and the commercial operation of short videos being the two most popular cross-disciplinary research areas.

Source: China National Knowledge Infrastructure (CNKI).

Figure 5. Distribution of academic disciplines.

Relevant literature published in higher education (10 papers, approximately 5.2 per cent) stood just above the five per cent threshold, whereas disciplines such as computer science, trade and economics, and sociology each accounted for less than 5 per cent of the total. Interdisciplinary research integrating fields such as psychology, tourism, public administration and medicine is extremely rare, indicating that there is ample scope for further exploration of the interdisciplinary applications of short videos.

Scholars both in China and abroad have conducted extensive empirical research into user satisfaction with the Douyin and TikTok platforms. Drawing on diverse theoretical frameworks, survey samples and econometric analysis methods, these studies have yielded a comprehensive body of research findings. Whilst the academic community has reached a degree of consensus on certain fundamental aspects, no unified conclusions have yet been reached regarding the intensity of the influence of various variables, the mechanisms by which these influences are transmitted across different contexts, or the specific pathways through which multidimensional psychological costs operate (Table 1).

Overseas research has largely focused on TikTok, examining the intrinsic links between users’ subjective experiences and psychological perceptions. Academic research, utilising a hybrid model combining SEM and fsQCA, has confirmed that there is a significant positive correlation between users’ perceived usage duration and their satisfaction with the platform. However, as users’ usage duration continues to increase, multiple psychological costs—such as privacy anxieties,

Table 1. Overview of the literature review on user satisfaction with Douyin and TikTok.

Source: China National Knowledge Infrastructure (CNKI).

health concerns, social pressure and platform fatigue—accumulate, leading to a usage state characterised by the coexistence of positive experiences and negative perceptions; this is a typical feature of the current short-video user experience ([29]). Among these, privacy risks and usage fatigue are the core factors reducing user satisfaction; however, health perceptions and social anxiety do not directly influence satisfaction. This conclusion also provides a reasonable explanation for the paradoxical phenomenon whereby users continue to use the platform despite experiencing negative effects ([29]). Short-video algorithms possess dual attributes: whilst they can optimise content matching accuracy and enhance the user browsing experience, they also exacerbate users’ concerns regarding privacy breaches, with platform reputation serving as a moderating buffer between these positive and negative effects ([17]). In short-video e-commerce and advertising contexts, content quality and social connections can enhance consumer loyalty by increasing user retention, whilst advertising transparency and disclosure practices directly influence users’ attitudes towards the platform and their behavioural decisions ([27]; [18]). Furthermore, flow experiences, platform usability, ease of use and a positive community atmosphere are key positive factors in sustaining long-term user engagement ([5]; [15]).

Research within the domestic academic community focusing on the Douyin platform is more closely aligned with the characteristics of the domestic industry’s development, with research dimensions covering multiple levels including content algorithms, platform services, advertising dissemination, and user psychology and demographic characteristics. Most studies confirm that content quality, the effectiveness of algorithmic recommendations, and basic platform services are the three core independent variables influencing short-video user satisfaction ([7]; [21]; [3]). Liu Mingzheng ([14]), through a multi-platform comparison, further corroborated that information quality is the most significant dimension determining user satisfaction with short-video UGC, followed by system quality and service quality. Early domestic research primarily relied on the ACSI model for analysis, verifying the positive impact of expectation fulfilment, perceived quality and perceived value on user satisfaction ([11]). Subsequent studies, combining logistic regression models with large-scale nationwide surveys, further highlighted the heterogeneous effects arising from advertising intrusiveness, perceptions of platform security, and demographic characteristics such as age and educational attainment ([7]; [21]). As new industry trends—such as AIGC technology, scenario-based short videos and advertising governance—continue to mature, domestic research has gradually expanded to adopt entirely new analytical perspectives. The originality, authenticity and narrative quality of AI-generated content have progressively become new benchmarks for evaluating the user experience on short-video platforms ([16]). Scenario-based short videos can shape users’ emotional perceptions through scene-building and emotional rendering, thereby directly influencing their satisfaction ratings ([2]). At the same time, advertising transparency can interact with users’ uncertainty-avoidance traits, ultimately influencing their consumption intentions and behavioural choices ([18]). Information overload, social comparison, privacy concerns and usage costs can trigger users’ intention to churn, implying that users’ overall evaluation of a platform is the result of the combined effects of implicit psychological strain and explicit user experience ([13]; [6]).

From the perspective of theoretical frameworks and research methods, the Technology Acceptance Model (TAM), Expectation Confirmation Theory, Flow Theory, Scenario Theory and Persuasion Knowledge Theory constitute the core theoretical underpinnings of this field. Traditional questionnaire surveys, multiple regression analysis and structural equation modelling serve as the fundamental empirical tools in this field, whilst grounded theory, fuzzy comprehensive evaluation and the CRITIC-VIKOR combined evaluation method are gradually being adopted, effectively enhancing the precision of relevant empirical research ([12]; [19]). Although existing research findings are abundant, there remain some notable common shortcomings. Most studies focus solely on the independent effects of individual variables, lacking configurational analyses of the synergistic interactions among multiple factors. The academic community has not explored the interaction mechanisms between subjective usage duration and multidimensional psychological costs in sufficient depth; furthermore, emerging research topics—such as AIGC applications, the dissemination of contextual content, and advertising regulation—remain fragmented and have yet to form a systematic and comprehensive analytical framework. Differences in sample selection criteria and econometric models across studies have led to divergent empirical conclusions; moreover, few studies have provided a reasonable interpretation of these discrepancies in light of the emerging content formats of short videos. Furthermore, the standardisation of sampling procedures in some surveys ([15]) requires improvement, and traditional research tools struggle to accommodate the diverse and unstructured nature of short video data, resulting in certain research limitations.

Overall, existing literature both domestically and internationally has established a foundational research framework for short-video user satisfaction, with the academic community generally acknowledging that user satisfaction is the result of the coupled effects of multiple factors. However, significant controversies and research gaps persist regarding the intensity of variable influences, the transmission pathways in specific scenarios, and the precise mechanisms underlying multidimensional psychological costs. The short-video industry is currently undergoing continuous iteration and upgrading towards intelligent, scenario-based and e-commerce-oriented models. AIGC technology, new content formats and diversified commercial monetisation models are constantly reshaping the user experience; consequently, this field still offers ample scope for research and theoretical expansion.

Personalised recommendation algorithms form the core technological backbone of the Douyin platform. Existing research confirms that these algorithms are capable of achieving precise content matching based on user behaviour tags, thereby effectively optimising users’ basic browsing experience ([23]). However, following prolonged implementation, such algorithms are prone to fostering issues such as information silos, content homogenisation and information redundancy; there are also potential negative implications arising from the platform’s excessive collection of users’ private data ([24]; [13]). As AIGC technology is gradually integrated into the platform’s content distribution process, the overall volume of short-form video content has increased significantly, which may further aggravate the problem of content overload on the platform.

The Douyin algorithm must carefully balance the proportion of ordinary short videos, narrative-driven content and commercial advertisements in its recommendations; any imbalance in the frequency of recommendations or the composition of content is likely to be associated with user resistance ([2]). It is now widely accepted in academic circles that the impact of algorithms on user satisfaction follows an inverted U-shaped pattern. However, different empirical studies have yet to reach a unified conclusion, and there remains a divergence of opinion within the academic community as to whether positive gains or negative consequences play the dominant role in the impact of algorithms.

The academic community has yet to reach a consensus on the mechanisms through which social interaction features on platforms influence user satisfaction; such features encompass mainstream forms of interaction such as “likes”, comments, shares and private messages. Some empirical studies suggest that a harmonious and friendly atmosphere for platform interactions is positively associated with users’ sense of belonging, enjoyment of use and satisfaction ([23]). However, other studies have suggested that high-frequency social interactions can lead to pressure to maintain one’s online image and anxiety arising from social comparison; such psychological burdens may inhibit users’ overall satisfaction levels ([7]). Furthermore, results from econometric analyses indicate that there is no significant correlation between users’ platform interaction behaviour and their overall satisfaction ([21]). As new interaction models—such as AIGC-driven virtual interactions and narrative-based short videos—become increasingly widespread, differences in users’ perceptions of these emerging features are becoming more pronounced, thereby further amplifying the divergent conclusions of previous research ([16]; [2]). Based on actual user feedback, publicly visible interaction records are readily associated with concerns regarding users’ social image, whilst prolonged exposure to high-frequency social scenarios is also associated with social overload—a factor frequently reported as a reason why users choose to uninstall the platform ([13]; [6]).

Advertising-related issues currently represent the area of greatest disagreement within this field of research. Traditional empirical studies have established a relatively unified consensus: unrestrained push notifications or hard-sell advertisements that are awkwardly inserted disrupt users’ normal browsing rhythm, thereby significantly diminishing the overall user experience. In contrast, product-promoting advertisements that are natively integrated into the plot and context often achieve higher levels of user acceptance ([30]; [7]; [21]). With the formal implementation of policies mandating the labelling of advertisements, entirely new research controversies have emerged in this field. Some argue that clear advertising labels can enhance the transparency of a platform’s commercial activities, effectively building user trust and thereby optimising overall satisfaction ([18]). However, opposing research perspectives suggest that explicit advertising prompts trigger users’ defensive mechanisms, easily inducing resistance and consequently lowering users’ evaluations of the platform. The covert placement of advertisements within situational short videos has further exacerbated the polarisation of user evaluations ([2]). Furthermore, users’ inherent personality traits related to uncertainty avoidance serve to further amplify the empirical discrepancies between these two lines of research ([18]).

Content is the core medium of short-video platforms, and related controversies continue to evolve alongside changes in content formats. In the realm of conventional user-generated content (UGC) short videos featuring real people, entertainment value and thematic diversity are positively associated with user satisfaction, whilst homogenised and vulgar content is negatively associated with user ratings, forming a general consensus ([23]; [21]). Following the widespread adoption of AI-generated content (AIGC), flaws such as illogical narratives and a lack of originality in AI-created works have led to a severe polarisation in user evaluations ([16]). Scenario-based short videos have emerged as a new source of controversy: whilst some users recognise that a complete narrative, immersive settings and emotional rendering provide emotional value, others criticise the plots for being exaggerated, with content deliberately creating anxiety and peddling negative emotions. Various quality defects in content are likely to trigger negative interactions such as poor user reviews and reports; this conclusion has been consistently validated by multiple empirical studies ([13]).

Data collection by platforms and the protection of personal information are key issues of concern to both users and the academic community. Most empirical studies define privacy concerns as a significant psychological burden; this negative perception has a significant adverse impact on users’ overall satisfaction ([23]; [7]; [13]). Whether it be AIGC content creation, targeted advertising or behavioural data collection through contextual videos, every aspect of platform operations involves the large-scale collection of users’ personal data, and the potential risk of data misuse continues to heighten users’ privacy concerns ([16]; [18]; [2]). The primary reason for conflicting findings in privacy-related research lies in the highly concealed nature of perceived privacy risks. Users who have never experienced a data breach tend to pay relatively little attention to such risks and have a generally low perception of them, whereas those who have experienced a data breach or online fraud develop strong feelings of aversion towards privacy violations. The differing perceptions between these two groups directly result in conflicting conclusions in existing empirical research ([24]).

There is a clear bidirectional relationship between user engagement time and satisfaction. Browsing within a reasonable timeframe can lead to a positive enhancement of the user experience; however, as engagement time continues to increase, negative psychological effects such as a sense of wasted time and mental fatigue gradually emerge, causing the positive experience and negative feelings to cancel each other out ([7]; [21]; [29]). Demographic characteristics exert a consistent heterogeneous influence; age, educational attainment and income level can significantly distinguish between users’ levels of satisfaction, their propensity to churn, and their content preferences ([21]; [3]; [6]). Among these, users with higher educational qualifications have higher expectations regarding the quality of AIGC content, the compliance of advertisements, and the production standards of situational short videos; consequently, discrepancies in their evaluations are more pronounced ([2]). In contrast, the gender variable has a relatively limited effect on distinguishing overall user satisfaction. Overall, Douyin users generally exhibit a composite usage state characterised by high satisfaction coupled with psychological strain, in which positive experiences, negative perceptions and adverse usage behaviours coexist simultaneously. Differences in tolerance levels and content preferences across user groups further amplify the divergence in research conclusions across various dimensions.

Young and middle-aged adults and those with higher educational qualifications form the core user base of Douyin; this demographic has higher expectations regarding five key areas: information security, information overload, shortcomings in AIGC content, advertising regulations, and standards for the production of situational videos. Existing empirical studies, both domestically and internationally, have predominantly selected university students and residents of a single region as research samples. This significant bias in sample composition directly results in anomalous weightings for various evaluation indicators, leading to systematic deviations in overall satisfaction scores and conclusions regarding content perception. There are fundamental differences in users’ core demands across different usage scenarios; users who favour entertainment browsing, live-stream shopping and AI-generated content, or those seeking immersive narrative experiences, each have distinct priorities. At the same time, platform functionality, the content ecosystem and advertising regulations are constantly undergoing iteration and updates, meaning that the strength of various moderating variables also fluctuates dynamically. The combined effect of these multiple factors makes it difficult to reach consistent empirical conclusions across different studies, which is also a key reason for the divergence of views within the academic community.

The lack of standardisation in evaluation systems and the measurement criteria for variables is the core reason why it has been difficult to reach a consensus in the relevant academic research. Furthermore, with the continuous emergence of new research topics—such as AIGC content, advertising disclosure and situational narratives—the satisfaction evaluation frameworks and indicator classification standards established by different studies have become rather fragmented. With regard to high-frequency core variables such as content quality, advertising intrusion and contextual experience, there are marked differences amongst scholars in terms of variable definitions, the selection of measurement scales and the establishment of evaluation criteria. Most empirical models incorporate only core independent variables, thereby omitting necessary mediating and moderating variables; moreover, traditional classical theoretical models struggle to adapt to the current short-video platform ecosystem, where multiple content formats coexist. Static, single-dimensional analyses are unable to fully identify complex mechanisms of influence, ultimately leading to conflicting and divergent empirical conclusions.

Existing empirical research in China employs a diverse range of methodological frameworks, with methods such as multiple regression, factor analysis, fuzzy comprehensive evaluation and traditional linear SEM being widely applied. As the operational logic and criteria for different econometric models vary, this in itself can lead to a certain degree of bias in the empirical results. Currently, there are still relatively few case studies in academia utilising fsQCA (Configuration-Based Qualitative Categorical Analysis). However, this method can effectively identify configuration effects such as multivariate synergies, substitution relationships and complementary mechanisms, making it better suited to the multi-factor coupling characteristics of short-video user satisfaction—an area that represents a prominent shortcoming in the current research framework. Furthermore, the vast majority of studies continue to rely on static cross-sectional data for their analyses. In an industry environment characterised by the continuous iteration of AIGC technology, constant adjustments to advertising regulatory policies, and ever-changing trends in scenario-based content creation, static data struggles to capture the dynamic patterns of user satisfaction over time. At the same time, some studies lack sufficient standardisation in areas such as sampling procedures, questionnaire design and data processing; the resulting distortion of raw data further amplifies the discrepancies in conclusions between different studies.

The present study has several limitations, which in turn point towards directions for future research.

First, existing literature tends to treat the positive user experience, multidimensional psychological strain and negative user behaviour as separate categories, and few studies have systematically examined how these variables intertwine and dynamically balance one another; research on the user experience of situational short videos also remains disconnected from mainstream satisfaction frameworks. Future research should therefore move beyond such fragmented approaches by integrating these topics into a multivariate coupled analysis model, so as to systematically dissect the logic underlying user satisfaction across different content formats and the interactive pathways among variables, ultimately establishing a comprehensive, multidimensional and integrated research framework for short-form video user satisfaction.

Second, survey research combined with traditional static statistical models remains the prevailing paradigm, whereas configurational analysis methods—such as fsQCA, which can uncover the synergistic, substitutive and complementary effects of multiple factors—have not been fully utilised; most studies merely replicate established frameworks when confronting new content formats such as AIGC and situational short videos. Future research should therefore integrate diverse empirical tools, including SEM, PLS-SEM, multiple regression, fuzzy comprehensive evaluation, grounded theory and fsQCA configuration analysis, so as to retain the precise testing of static effects offered by traditional linear models whilst uncovering complex coupling effects, thereby enhancing both the comprehensiveness and the interpretability of empirical analysis.

Third, existing satisfaction assessment frameworks are constructed largely on explicit, surface-level indicators, leaving implicit, deep-level variables—such as multidimensional psychological strain, risk-averse personality traits, the quality of situational narratives and the originality of AIGC content—insufficiently explored, whilst moderating variables of significant value, such as platform reputation and privacy protection mechanisms, have received limited attention ([29]; [16]; [18]). At the same time, most studies rely on one-off cross-sectional data and coarse demographic segmentation. Future research should therefore expand the variable framework and optimise sampling and segmentation by incorporating multidimensional behavioural indicators—such as content preferences, acceptance of AIGC, sensitivity to advertising and preferences for situational videos—and should adopt long-term, continuous longitudinal tracking to capture the dynamic evolution of user satisfaction as platforms undergo technical updates, content upgrades and iterations of advertising policies ([16]; [18]).

Fourth, given the proliferation of usage scenarios—from pure entertainment browsing to live-streaming e-commerce and scripted content—each corresponding to a distinct user-experience logic, future research should move beyond single-scenario, one-off designs and conduct cross-scenario comparative analyses to identify how the weighting and underlying mechanisms of satisfaction indicators differ across scenarios. Such research could draw on established theoretical frameworks and variable systems from both domestic and international sources whilst adapting them to the distinctive characteristics of the domestic Douyin platform, thereby establishing a research framework for short-form video user satisfaction that is tailored to the local market environment.