E-commerce development is reshaping production, logistics and trade systems, yet its environmental consequences remain difficult to evaluate because domestic and cross-border e-commerce operate through different market–logistics regimes and may overlap within the same city. Using the National E-commerce Demonstration City programme and the Cross-border E-commerce Comprehensive Pilot Zone programme as policy-induced institutional indicators of domestic and cross-border e-commerce development, this study examines whether their co-development generates additional carbon effects. Based on a balanced panel of 279 Chinese prefecture-level cities from 2009 to 2019, we combine two-way fixed-effects DID models with a spatial DID framework. Both NEDC-enabled domestic e-commerce development and CBEC-enabled cross-border e-commerce development are found to be associated with local carbon reduction, and their overlap generates an additional local carbon-reduction effect after the separate effects of the two regimes are accounted for. This additional local gain is accompanied by a positive indirect effect on neighboring-city emissions, leaving the regional total effect statistically insignificant. E-commerce co-development therefore produces local carbon-reduction synergy while raising the risk of spatial emission displacement. Technological innovation and economic agglomeration are associated with lower emissions, whereas industrial structure upgrading and entrepreneurial activity are associated with higher short-run emissions; technological innovation attenuates these carbon-increasing associations. Stronger local gains are found in inland and high-digitalization cities. The findings suggest that e-commerce development policies should move beyond platform expansion and local performance targets, and incorporate regional logistics coordination, low-carbon warehousing and supply-chain carbon accounting into the governance of e-commerce and digital trade.
This paper investigates the dynamic effects of attribute information concentration (AIC) in brand posts throughout the new product launch cycle and explores the moderating effects of release period, brand strength, and market competition. The research findings reveal that: (1) AIC is positively associated with the effectiveness of information diffusion, consistent with a consumer-perceived value explanation, with a stronger effect observed before the product release than after, (2) brand strength negatively moderates, while market competition positively moderates, the relationship between AIC and information diffusion effectiveness, and both moderating effects are particularly pronounced before the release. These insights offer a valuable theoretical framework and practical suggestions for companies to dynamically manage their brand post strategies during new product launches.
E- commerce retailers increasingly deploy AI-powered personalized recommendations to drive conversion, yet whether recommendations translate into purchase intention depends on consumer trust and privacy attitudes. Drawing on the S-O-R framework, this study develops and tests a dual trust framework in online retail, distinguishing cognitive trust in recommendations from institutional trust in the retailer. Using survey data from 1044 Chinese online shoppers and structural modeling, we find that AI-powered personalized recommendations increase purchase intention primarily through cognitive trust, with institutional trust providing a complementary path. Importantly, privacy concerns and privacy self-efficacy were found to exert differential moderating influences on each pathway—privacy concerns amplify cognitive trust effects while weakening institutional trust impacts, whereas self-efficacy operates the opposite way. These findings refine trust theory in AI-enabled retailing by distinguishing algorithmic performance and platform governance as two trust routes. Practically, retailers should jointly enhance recommendation transparency and quality to build cognitive trust, and signal institutional assurances and benevolence to bolster institutional trust, tailoring recommendation strategies to heterogeneous privacy attitudes.
This study develops differential game models to optimize livestreaming traffic through personalized video recommendations while addressing viewer privacy concerns. We analyze three game scenarios: Nash non-cooperative, centralized decision-making, and cost-subsidy games. Our findings reveal that platform operators’ strategies and traffic growth coefficients significantly affect viewer privacy behaviors. Increased privacy investment does not always enhance information sharing; beyond a certain threshold, viewers may become less willing to disclose personal data. Cost subsidization by platform operators can encourage private data sharing, leading to increased livestreaming traffic and revenue, thereby achieving Pareto improvement. However, revenue distribution may influence the dynamics between operators and streamers, potentially giving rise to free-riding behavior. This study offers insights into balancing personalized services, privacy protection, and traffic growth for stakeholders on livestreaming platforms.
This study examines how the digital economy (DE) influences tourism productivity in 278 Chinese cities from 2000 to 2022. By applying advanced efficiency assessment methods, we measure tourism total factor productivity (TTFP) and the level of DE development. Using a spatial analysis approach, the study finds that DE has a nonlinear impact on tourism productivity, showing a peak followed by a decline in direct effects and a gradual strengthening of indirect effects. Regional differences are evident: DE significantly enhances tourism productivity in eastern cities with better digital infrastructure, while its impact remains weaker in central and western regions due to slower digital adoption. Moreover, the enhancement of labor productivity (LP) and the reduction of energy consumption (EC) emerged as key pathways through which DE promotes TTFP. These findings provide valuable insights into reducing regional digital disparities, fostering tourism resource flows, and boosting tourism sector efficiency.
With the proliferation of omnichannel retailing, consumers increasingly evaluate products across online and offline channels before purchasing. Leading brands have responded by integrating live streaming e-commerce with offline experiences. However, prior research offers limited guidance on synchronizing these channels in response to such integrated evaluation patterns. To address this gap, this study analyzes 163,256 shopper reviews from TikTok Shop across European markets. Leveraging latent dirichlet allocation for feature word extraction and social network analysis for relation mapping, this study identified 12 key decision factors and quantify their interactions through association rules. The findings highlight the significant interplay between online and offline marketing elements in shaping consumer purchase decisions, particularly the correlations between price and product experience, live streamer and product experience, and price and after-sales service. Furthermore, this study identified four cliques of feature variables, revealing distinct patterns of cross-channel factor interactions. The feature variables within the same clique demonstrate positive interactions, revealing systematic patterns in how online and offline decision factors are mentioned together in consumers’ evaluation narratives. The findings of this study provide actionable insights for marketing managers and practitioners to enhance synergy between live streaming platforms and physical stores.
This study investigates the primary sources of user dissatisfaction in Türkiye’s leading e-commerce applications by analyzing user-generated content through advanced topic modeling techniques. As mobile commerce becomes increasingly integrated into daily life, user reviews offer valuable insights into both functional performance and perceived service quality. Yet, the vast and unstructured nature of these reviews poses challenges for conventional analysis methods. To address this, the BERTopic algorithm which combines contextual embeddings with class-based TF-IDF was applied to 18,685 Turkish-language user reviews collected from the Google Play Store using the google-play-scraper Python library. The analysis revealed recurring concerns around delivery delays, ineffective customer service, and technical issues related to payment and order tracking. The recurrence of these themes across different platforms reflects common patterns in key areas that influence user experience within Türkiye’s digital commerce ecosystem. By offering a structured and semantically coherent thematic mapping of user sentiment, this study provides actionable insights for improving service quality and contributes methodologically to natural language processing research on morphologically rich languages like Turkish.
As e-commerce platforms increasingly adopt augmented reality (AR), understanding how immersive technologies influence consumer wellbeing has become essential. Drawing on the stimulus-organism-response (S-O-R) model and cognitive appraisal theory, this article develops and empirically tests a framework that examines how AR features shape affective satisfaction, buyer empowerment, emotional wellbeing, and purchase intention. A mixed-methods design was employed, involving qualitative interviews and a cross-national survey of AR users. Findings indicate that personalized interactivity and usability realism strengthen immersive experiences and perceived value, which subsequently drive emotional outcomes and behavioral intentions. Moderating effects of perceived artificial intelligence (AI) intrusiveness and AR-related anxiety suggest that psychological discomfort can diminish satisfaction’s impact on purchase intention. Overall, the study integrates cognitive and emotional psychological pathways to extend digital marketing theory and offers actionable insights for designing e-commerce experiences that support consumer wellbeing and ethical engagement, understood as the protection of consumer autonomy, privacy, and psychological comfort during immersive shopping, in e-commerce.
As online shopping becomes increasingly integral to consumers’ lives, driven by the convenience of digital devices and evolving customer expectations, understanding and optimizing mobile application design has emerged as a critical challenge. Designing mobile applications that effectively meet user expectations requires a robust understanding of user preferences and priorities. This study proposes an integrated methodology combining the Kano model, Bayesian Best–Worst Method (BBWM), and Conjoint analysis to address the complexities of mobile shopping application design. By classifying features through the Kano model and employing conjoint analysis for structured evaluation, the approach reduces design alternatives and prioritizes customer-centric features. In this study, a systematic four-stage methodology was employed: (1) refining 55 mobile app features to 33 using machine learning, (2) categorizing features with the Kano model, (3) weighting features through BBWM, and (4) validating satisfaction levels via focus group testing and conjoint analysis. Results provide actionable insights for optimizing mobile shopping platforms and adapting to evolving consumer demands, ensuring long-term success in the competitive e-commerce landscape. This research fills a critical gap in experimental design for e-commerce applications and offers practical solutions for creating platforms that align with customer expectations and business objectives.
Reputation systems are essential for creating trust and reducing information asymmetries in online markets. However, they are vulnerable to biases that distort information, erode trust and potentially cause market failure. This paper surveys the empirical literature on five key sources of bias: (i) strategic actions of sellers such as fake reviews and rebate-for-review programs, (ii) reciprocity, (iii) social influence bias, (iv) selection bias and (v) noise. It analyzes the biases through a unified framework decomposing errors into three distinct signatures: systematic mean shifts, variance inflation and serial correlation. A central finding is that platform design involves trade-offs: interventions targeting one error signature often exacerbate other distortions. By illustrating how biases interact, this paper provides a diagnostic roadmap for platform operators, regulators and policymakers to match interventions to the specific error signatures while compensating for secondary effects, ultimately enabling reputation systems to maintain trust, reduce information asymmetries and enable efficient online markets.
Digitalization is transforming the way products and services are acquired, giving e-commerce an increasingly relevant role. This research aims to assess the current state of e-commerce among automobile manufacturers and analyze whether there is a convergence between the digital strategies of traditional and emerging manufacturers. To this end, a qualitative methodology based on a case study method has been employed, comparing the digital sales platforms of two major manufacturers in the industry through a cross-sectional web content analysis conducted at a single point in time. The results show that e-commerce has a significant presence, although it is still in a developmental stage. In particular, traditional manufacturers face greater difficulties in digitalizing the sales process due to their dependence on automotive dealerships and the complexity of their offerings. Emerging manufacturers, on the other hand, provide a smaller range of digital services, but in a more integrated manner. Furthermore, it is observed that the strategic value of e-commerce lies not only in the online sale of vehicles, but also in its ability to serve as a gateway to more servitized, personalized business models linked to the after-sales relationship.
The in-depth application of artificial intelligence technology has driven e-commerce platforms to transform toward intelligent service-oriented operations. Most existing platform evaluation systems focus on conventional dimensions and lack a quantitative assessment of intelligent services. Traditional evaluation methods also have difficulty addressing the fuzziness of intelligent indicators and conducting scientific quantitative evaluations. In this paper, an evaluation system for the intelligence level of e-commerce platforms is established, which consists of 4 primary indicators, namely, intelligent customer service, personalized recommendation, intelligent search, and intelligent promotion, along with 15 secondary indicators. An improved matter-element extension model adopting AHP–extension combined weighting is applied. By defining the classical domain and joint domain of indicators and calculating correlation degrees and evaluation grades, an objective quantitative evaluation of the platform intelligence level is performed. Taking JD.com as the empirical research subject, this study conducts an analysis based on evaluation criteria from seven experts and 504 valid user questionnaires. The results indicate that JD.com reaches an overall favorable intelligence level, with balanced development across the four major dimensions. Nevertheless, the platform still has deficiencies, such as insufficient reliability of intelligent customer service and low accuracy of personalized recommendations. This study fills the gap caused by the absence of the intelligence dimension in e-commerce platform evaluation and broadens the application scope of the matter-element extension model in digital business. It can provide a theoretical reference and practical support for the intelligent upgrading of e-commerce platforms and the formulation of industrial intelligent evaluation standards.
Recommender systems underpin many online services, yet improving user acceptance and engagement remains a critical challenge beyond optimizing accuracy. This study introduces a novel method for generating personalized persuasive explanations in recommender systems by integrating users’ demographics, Twitter-derived Big Five personality traits, and social network data. Addressing the limitations of one-size-fits-all explanations and static personality mappings, the proposed approach leverages multi-label machine learning models to infer users’ dominant persuasive strategies based on the Cialdini principles of influence. A rich dataset was collected from Persian-speaking Twitter users, including a unique corpus of tweets annotated with personality traits and a second dataset incorporating persuasion effectiveness ratings. Personality traits were identified through multi-label classification, using text representations generated by ParsBERT and FastText. Based on inferred personality traits, demographic information, and social network data, multi-label classifiers were used to predict users’ preferred persuasive strategies, which were then used to generate personalized persuasive explanations. In experimental evaluations on datasets of 1936 participants for personality detection and 415 participants for persuasive strategy prediction, our method achieved up to 0.88 and 0.95 weighted F1 score for personality trait detection and persuasive strategy detection, respectively, and outperformed the examined rule-based baselines in this evaluation. In a controlled user study with 132 participants, participants assigned to our personalized persuasive explanation condition reported substantially higher willingness to follow the recommendations than those exposed to the two rule-based personalization approaches, with increases of 68.6
This study uses Chinese A-share listed companies from 2012 to 2023 as the research sample. Focusing on the cross-sectoral digitalization gap between banks and enterprises, namely the Bank-enterprise Digital Gap, it examines the impact of this gap on corporate ESG performance and further investigates the moderating role of the business environment. The results show that the convergence of the Bank-enterprise Digital Gap significantly improves corporate ESG performance. In other words, the closer the alignment between the digitalization levels of banks and enterprises, the more conducive it is to firms’ adoption and implementation of ESG practices. Mechanism tests indicate that the convergence of the Bank-enterprise Digital Gap enhances corporate ESG performance primarily by reducing firms’ information search costs and risk premium costs. Heterogeneity analysis further reveals that this positive effect is more pronounced among non-state-owned enterprises, small-scale enterprises, and firms characterized by CEO duality. Furthermore, an improved business environment strengthens the positive effect of the convergence of the Bank-enterprise Digital Gap on corporate ESG performance. In particular, the innovation environment, financial services, and human resources environment exert significant positive moderating effects. This study extends the literature on the determinants of corporate ESG performance from the perspective of cross-sectoral digital matching between banks and enterprises, and reveals the mechanism through which digital collaboration empowers corporate sustainable development. The findings suggest that, while promoting the digital transformation of both the banking sector and enterprises, policymakers should also attach importance to narrowing the Bank-enterprise Digital Gap and continuously improving the business environment, thereby providing external support for enterprises to enhance ESG performance and contribute to the high-quality development of China’s economy.
With the rising popularity of online video channels, influencers increasingly make their vlogs “shoppable” by adding a “shopping cart” to simplify the customer journey. However, when presented inappropriately, shopping carts may disrupt consumers’ experience and incur their reactance. This study examines the underlying mechanism of shopping cart executions on consumer acceptance of sponsored vlogs, drawing on the theoretical lens of the persuasion knowledge model (including both conceptual and attitudinal persuasion knowledge) and narrative transportation theory. In the scenario of narrative vlogs with shopping carts, a 3 (placement timing compared to product appearance: prior to/concurrent with/after product appearance) × 2 (placement prominence: low/high) between-factor experiment was employed for data analysis. The results show that shopping carts that appear after the product appearance result in the most narrative engagement, followed by the shopping carts prior to the product appearance, and then the shopping carts placed concurrent with the product appearance. Shopping carts placed earlier than product appearance result in more ad recognition (conceptual persuasion knowledge) and fewer inferences of manipulative intent (IMI; attitudinal persuasion knowledge). In addition, shopping carts with high prominence will result in less narrative engagement, but more ad recognition and less IMI than those with low prominence. This study offers insights for platforms and advertisers on optimizing the design of shopping carts in vlogs for consumer responses. This study advances the theoretical research on shopping cart placement, updates the information system (IS) research on information presentation, and identifies the role of persuasion knowledge and narrative engagement in the narrative vlog context.
Finfluencers, or financial influencers who share investment advice online, are reshaping how consumers access, evaluate, and act on financial information. This study examines how their credibility signals influence investment intentions through the mediating role of trust and the moderating effects of regulatory focus and financial literacy. Two experiments were conducted: Study 1 tested the direct and mediated effects of credibility on investment intention, while Study 2 examined boundary conditions. Results show that strong credibility cues enhance trust and investment intention, with promotion-focused consumers responding more strongly than prevention-focused ones. Financial literacy moderates these effects: Low-literacy consumers are more influenced by credibility cues, whereas high-literacy individuals rely on personal judgment. By integrating Signaling Theory with psychological and cognitive mechanisms, this study extends consumer behavior research to financial contexts, highlighting how perceived credibility, trust, and individual differences shape consumers’ financial decision-making in social media environments.
In the highly competitive business ecosystem, brand crises function as a heavy explosive device, severely damaging company goodwill. Additionally, consumers’ reference quality effects can further influence profitability by impacting goodwill. This paper constructs a dynamic binary supply chain model comprising a manufacturer and an electronic retail platform. It analyzes the differences in optimal decision-making, profit levels, and sales mode selection under both non-crisis and crisis scenarios for flagship store and self-operated store models. Research indicates that platform marketing efficiency influences the impact of consumer reference quality effects on corporate goodwill, with post-crisis goodwill potentially surpassing pre-crisis levels. Post-crisis decision-making and resource allocation are closely linked to crisis damage rates. Additionally, sales mode selection under crisis conditions shifts due to factors such as commission rates, platform marketing efficiency, and crisis damage rates. The robustness of core findings is confirmed by considering variations in consumer price sensitivity and hazard myopia.
In digitally mediated services, ride‑hailing platforms increasingly adopt corporate social responsibility initiatives to enhance legitimacy and strengthen consumer relationships. Understanding how consumers evaluate these initiatives is particularly important in digital service contexts, where perceptions of authenticity strongly influence behavioural outcomes. This study examines how four antecedents of corporate social responsibility evaluations—corporate social responsibility fit, perceived impact, reparative motives, and self‑serving motives—shape consumers’ perceptions of corporate social responsibility authenticity. It further analyses how perceived authenticity affects customer well‑being and loyalty in ride‑hailing services. Data were collected through an online survey of 326 users. Results indicate that corporate social responsibility fit, perceived impact, and reparative motives positively enhance perceived authenticity, whereas self‑serving motives negatively affect it. Perceived authenticity positively influences customer well‑being and loyalty and mediates the relationship between well‑being and loyalty. These findings highlight authenticity as a key mechanism linking corporate social responsibility initiatives to consumer outcomes.
This study examined the associations between perceptions of different justice dimensions, negative emotions, and negative engagement intentions. As competition intensifies, companies often encourage their customers to engage in co-creation behavior to enhance customer engagement. However, the benefits of this co-creation in the context of negative customer emotions remain unclear. This study recruited 570 viewers of streaming competition-based reality shows as participants; in these shows, viewers can participate in co-creation behavior. The empirical results indicate that distributive and interpersonal justice are negatively associated with the negative emotion of dislike, whereas no such associations are found for procedural or informational justice. Furthermore, dislike is positively associated with hate. Dislike is positively associated with discontinuous watching and non-purchase intention, while hate is positively associated with retaliation. Our findings highlight the dark side of co-creation behavior; when customers experience the negative emotions of dislike and hate, co-creation behavior amplifies their negative engagement.
In the rapidly expanding live-streaming e-commerce environment, firms struggle with inefficient streamer allocation because the dynamic, time-varying effects of different streamer types remain unresolved. This study investigates how the sales-boosting effects of corporate versus external streamers evolve over time and across product types. Utilizing a Generalized Additive Model (GAM) on a robust 53-week longitudinal dataset tracking 944 top-selling products (47,733 observations) on Douyin, we capture nonlinear temporal dynamics. The empirical results reveal highly significant time-varying trajectories: the sales impact of corporate streamers exhibits a sustainable, gradually strengthening trend, whereas external streamers’ impact significantly decays after an initial spike. Furthermore, product type dynamically moderates these trajectories. The sustainable advantage of corporate streamers is significantly amplified for search goods, while the rapid decay of external streamers’ effectiveness is pronounced for experience goods. To unpack the psychological mechanism driving these macro-level dynamics, a controlled behavioral experiment (N = 200) demonstrates that corporate streamers foster cumulative purchase intentions via cognitive trust, whereas external streamers rely on transient affective trust. Ultimately, this research shifts the streamers marketing paradigm from a static to a dynamic perspective, providing firms with evidence-based guidance for tailoring streamer selection to product types and campaign stages to achieve sustainable sales growth.
