Product feature prioritization guides product managers to identify the most important product features or feature mixes (e.g., consumers’ simultaneous consideration of cleanliness and sleep quality with value for the money in hotel services) out of various features characterizing the product category. In e-commerce, consumers’ online product reviews are a critical source of information detailing their consumption experiences. However, e-commerce consumers usually discuss only a few features in their online reviews. This data structure creates a problem of missing information. We test multiple missing data imputation methods and empirically select the best one for the next step (called feature engineering) to identify important features and feature mixes. In identifying important features, most research has focused on individual features without considering the simultaneous dependencies involving multiple features. Our feature engineering approach considers not only individual features but also a mix of features in product evaluations. Thus, our procedure for feature prioritization allows product managers to identify important feature mixes for existing product improvements and new product development. Our empirical applications involve four product categories (hotels, restaurants, moisturizers, and tablets) examined under various conditions; moreover, they consistently underscore the crucial role of consumers’ value-for-the-money perceptions in selecting and evaluating product quality features.
Salesperson attractiveness (SA) plays a pivotal role in marketing; however, existing research has focused mainly on SA derived from physical cues, overlooking its non-physical aspect. To fill this gap, two studies were conducted in the livestreaming context. Study 1 focused on SA prediction and interpretation. Using a dataset of 1,933 livestreaming videos, this research customized a multimodal machine learning (MMML) model with a marketing metric (product category). The customized MMML model integrates the salesperson’s verbal, vocal, and visual cues, as well as their cross-modal interactions, achieving the highest prediction accuracy of 92.17% in predicting SA. An explainable AI model was subsequently applied to identify how specific verbal, vocal, and visual cues contributed to the prediction of SA. Study 2 explored the sales effect of SA and its boundary condition. The results indicate a U-shaped relationship between SA and sales, with this relationship being more pronounced for experience products than for search products. Overall, this research offers a more comprehensive understanding of SA by integrating both physical and non-physical cues, establishing an empirical framework to predict and interpret SA and providing marketers with novel insights into its effect on livestreaming e-commerce sales.
The exponential growth of digital content has transformed how organizations interpret consumer sentiment and anticipate societal trends. Among these trends, perceived injustice, a subjective experience of unfair treatment, poses significant implications for social dynamics and economic stability. While previous research on digital content emphasizes real-time, impulsive, and moderated formats such as social media reactions, this article explores the overlooked signaling potential of deliberate, less moderated content. A linguistic time series analysis of over 58 million words across 1,620 self-published e-books published over the course of 15 years reveals that expressions of perceived injustice embedded in deliberate content are correlated with consumer confidence and protest participation up to 12 months later. These findings extend current interactive marketing scholarship by integrating linguistics with time series analysis, offering both conceptual insights and practical implications for organizations aiming to anticipate disruptive consumer attitudes and actions.
The success of most social media platforms depends on their ability to attract and retain users. Although more than half of the world’s population uses some form of social media, consumers are increasingly discontinuing their use of these platforms. This research investigates a previously unexamined factor that influences consumers’ decisions to abandon (i.e., deactivate or delete) social media, users’ perceptions of their own inauthenticity on social media. Five studies (including one in the appendix) find that greater self-perceptions of online inauthenticity increase consumers’ willingness to abandon social media. Offering insight into the mechanism, the authors reveal the underlying role of self-threat. Specifically, perceived online self-inauthenticity threatens the self-concept, and consumers can counter this threat by disposing of the inauthentic social media account. That is, social media abandonment may be a compensatory response to the self-threat evoked by one’s own inauthenticity. Altogether, this work makes theoretical contributions to the literatures on social media use, identity, and service termination, while also providing guidance to social media platforms that seek to maintain active users and curtail abandonment.
The growing literature on brand activism highlights its increasing relevance in marketing but offers a offers limited insight into how consumers respond to different types of activism. Addressing this gap, we distinguish between social and political brand activism and examine how consumer engagement varies across these contexts. Our analysis of 1.7 million social media posts from 269 corporate accounts over a decade (2012-2022) reveals that social and political activism evoke divergent consumer responses. Social activism is on average associated with greater consumer engagement. In contrast, the relationship between political activism and engagement is significantly moderated by brand competence, with brands perceived as competent achieving greater engagement when sharing political messages. We supplement our observational study with a controlled experiment which both provides causal evidence consistent with the role of brand competence in shaping engagement with activist content and reveals that self-brand connection as a key psychological mechanism driving engagement with activist content. These findings advance the literature on brand activism by demonstrating that activism is not a homogeneous construct and by deepening understanding of how consumers engage with brands' social and political expressions.
Social media platforms often use tiered status systems to categorize influencers into different tiers based on their follower count. However, few studies have examined the effects of these follower-based status systems on influencers’ content-creation behaviors. This study addresses this gap using data from a leading Chinese social media platform. The authors explore the platform’s influencer status–assignment rule that grants “mid-tier” status to influencers with between 50,000 and 500,000 followers; specifically, the authors estimate the impacts of gaining mid-tier status on influencers’ content creation using a regression discontinuity approach. The results show that obtaining mid-tier status positively impacts the quantity of sponsored but not organic content. However, status change does not affect the quality (measured through user engagement) of sponsored or organic content. Further analysis reveals that the positive effect on sponsored content quantity operates through multichannel networks (MCNs), which selectively allocate sponsorship opportunities based on influencers’ follower-based status.
Social media memes have become a pervasive vehicle for communicating and contesting the idealized expectations that shape social roles. Yet marketers lack a clear framework for understanding how different types of memes align with the ideological positions consumers hold, and what that alignment means for brand messaging. Drawing on a qualitative discourse analysis of Instagram memes, we examine how meme types map onto competing ideological positions, extending existing typologies that classify memes as assertive, directive, or expressive. We introduce two subtypes of expressive memes—expressive-pragmatism and expressive-fantasy— capturing, respectively, emotional authenticity in recognition of dominant norms and symbolic, often humorous resistance to dominant norms. Our findings show that memes function as digital canvases on which consumers negotiate identity and contest norms, and that this ideological patterning offers brands a framework for calibrating messaging tone and positioning, rather than a claim that memes directly cause shifts in consumer belief or behavior. We conclude with implications for marketers and brand strategists seeking to engage authentically with the ideologically diverse communities that populate digital spaces.
Influencer-hosted livestream commerce has surged in popularity. Yet, little empirical research has examined what drives viewers along the conversion funnel in these time-compressed sessions. Drawing on the cognitive-affective interpersonal involvement framework from parasocial relationship (PSR) theory, this research proposes that PSR strength propels viewers from initial interest to purchase, mediated by two viewer-generated situational stimuli, namely product-related informational density and positive emotional engagement. The research first introduces a novel operationalization of PSR via digital fan badges and validates these badges as behavioral proxies for PSR using both qualitative and quantitative evidence. Observational analysis of over 1,500 influencer-hosted live sessions supports the hypothesized pathway: both product-related informational density and positive emotional engagement significantly mediate the effect of a viewer’s PSR strength on product interest, but only positive emotional engagement significantly influences actual purchase behavior beyond interest. Path analysis indicates that positive emotional engagement is the strongest mediator linking PSR strength to both product interest and purchase. These findings highlight the critical role of PSR-driven positive emotion in shaping the conversion funnel in livestream commerce. This research also offers a promising approach to capturing viewer-influencer PSR at scale using behavioral proxies.
Retargeting, the practice of showing ads to previous website visitors, has been a core value proposition for many digital advertising platforms. However, recent privacy restrictions (e.g., Apple’s iOS 14, the General Data Protection Regulation) have significantly reduced its efficacy, leading many advertisers to cut digital ad spending. Despite the widespread belief that retargeting is essential for online advertising efficacy, empirical evidence of its impact on incremental sales remains limited. This article presents the results of a large-scale field experiment conducted on the Google and Facebook ad networks for a major player in the bed-in-box category. Contrary to prevalent industry assumptions, the experiment reveals large and consistent returns to prospecting ads across both ad networks, while retargeting ads yield significantly lower returns. These findings challenge the perceived necessity of retargeting in digital advertising strategies and provide new insights into the role of prospecting ads in influencing consumer decisions for durable goods.
Influencer livestreaming has become a key marketing strategy for firms to engage with consumers. A recent trend, co-livestreaming, features the influencer and the brand representative jointly presenting and promoting products. However, the effectiveness of co-livestreaming remains unclear. Building on signaling theory, in this study we investigate how co-livestreaming (vs. single-influencer livestreaming) affects livestream performance. Using secondary data analysis and three experiments, we find that co-livestreaming outperforms single-influencer livestreaming, significantly enhancing both transactional (product sales) and relational (new follower count) performance. This effect is serially mediated by perceived brand endorsement and influencer credibility. However, the positive impact of co-livestreaming on performance diminishes as the popularity of the influencer and brand increases. These findings reveal the mechanism and boundary conditions of co-livestreaming effectiveness, providing actionable insights for firms and influencers to adopt more effective livestreaming strategies.
Negative online reviews are persistent, platform-amplified signals that can trigger information-driven reputational crises, often paracrises. Yet online reputation management (ORM) research remains fragmented and underrepresents strategies firms use in practice. Grounded in crisis communication theory and triangulating academic, practitioner, corporate, legal, and regulatory evidence, we develop and validate a hierarchical taxonomy of 17 ORM strategies organized into three macro logics: Micro-engagement, addressing focal reviews; Rebuild, rebalancing the evaluative climate; and Visibility, managing exposure through attention control and algorithmic salience. Citation-network construct tests validate the taxonomy’s scholarly structure: shared group membership nearly triples citation odds after group-level controls in a balanced dyad design (N = 5,115; OR = 2.922, 95% CI [1.889, 4.521]). A cosine-similarity topology identifies the most proximate groups, and a construct-level merge diagnostic rejects collapse on objective, strategy class, mechanism, ethical/legal profile, and observability, supporting the 17-group structure. The taxonomy shows that research concentrates on Micro-engagement and Rebuild, whereas Visibility practices (generic naming, decoy websites, and search-result engineering) remain under-theorized. Because Visibility tactics reshape information accessibility, they can suppress negative content while dampening positive signals and reducing transparency. We provide a foundation for cumulative research and a decision map for managers, platforms, and policymakers navigating platform-mediated review crises.
Privacy regulations, such as the General Data Protection Regulation (Article 17) and California Consumer Privacy Act, give consumers the right to request the deletion of their personal information. While researchers have begun to examine the implications of these laws, the effects of different data request options (to provide, keep, or delete information) on consumers' personal-information-sharing behavior have not been thoroughly studied. Drawing on prior work on autonomy, the authors investigate a new construct, "enactment autonomy." They demonstrate that data requests can alter perceptions of shared autonomy, influencing feelings of vulnerability and subsequent information-sharing behavior. Six experimental studies, including one behavioral study and a within-paper meta-analysis, provide support for the proposed data request effect as well as boundary conditions related to perceived information sensitivity and the presence of a third party.
Data is an indispensable asset in the AI ecosystem. This article investigates consumers' lay understanding of the different types of data that AI systems use to generate recommendations, and how this understanding influences their likelihood of accepting those recommendations. Across one pilot study and four main studies, the authors establish consumers' mental construction of three different data types and experimentally validate two mechanisms that shape recommendation acceptance: (1) perceived individuality threat associated with these data types and (2) their processing acceptability.
Retargeting, the practice of showing ads to previous website visitors, has been a core value proposition for many digital advertising platforms. However, recent privacy restrictions (e.g., those in Apple's iOS 14 and the General Data Protection Regulation) have significantly reduced its efficacy, leading many advertisers to cut digital ad spending. Despite the widespread belief that retargeting is essential for online advertising efficacy, empirical evidence of its impact on incremental sales remains limited. This article presents the results of a large-scale field experiment conducted on the Google and Facebook ad networks for a major player in the bed-in-a-box category. Contrary to prevalent industry assumptions, the experiment reveals large and consistent returns on prospecting ads across both ad networks, while retargeting ads yield significantly lower returns. These findings challenge the perceived necessity of retargeting in digital advertising strategies and provide new insights into the role of prospecting ads in influencing consumer decisions for durable goods.
Increasing competition pushes firms to accelerate their product launches, leading to closely spaced launches within firms. Especially in fast-paced markets, such as digital entertainment, the postlaunch period of one product often overlaps with the prelaunch period of the next. Existing studies examined firm-initiated communications in separate pre- versus postlaunch periods. In contrast, the authors study the effects of a publisher's overlapping product launches on social media engagement and associated spillovers of social media posts across partners promoting the focal product: publishers, product creators, and product accounts. The regressions use daily data covering the pre- and postlaunch periods of video games released by major publishers. The results show that overlaps harm the engagement rates of publishers and product accounts but benefit creators. Partner spillovers are asymmetric: Publisher and creator posts drive engagement to product accounts but not vice versa, and creator posts benefit publishers but not vice versa. Overlap effects on spillovers are usually most prominent when overlap occurs in the focal game's prelaunch period, with limited available product information, rather than postlaunch. Partners should align their social media strategies around overlapping launches to boost engagement for the focal product: Creators should intensify communication prelaunch, whereas publishers should increase communication postlaunch.
Supplier to small business to consumer (S2b2C) has emerged as a novel e-commerce model, where S2b2C platforms sourceproducts from upstream suppliers and distribute them through individual sellers who promote and sell products within theirown social networks. This model differs from traditional e-commerce in two key aspects: It segments the entire consumer mar-ket into distinct communities (i.e., consumer segmentation) and creates additional consumer value through social interactionswith individual sellers (i.e., interaction utility). This article examines the impacts of consumer segmentation and interaction utilityon the market outcomes. Specifically, the authors compare the S2b2C model with traditional e-commerce in terms of prices,demands, profits, consumer surplus, and social welfare and report the followingfindings. First,firms benefit from an asymmetricconsumer-segmentation structure, where the high-end seller targets consumers with sufficiently high valuations, while the low-end seller serves the remaining market segment. Second, conventional wisdom on double marginalization may not hold under theS2b2C model. Despite the introduction of an additional layer in the S2b2C supply chain, both the wholesale price and the retailprice can be lower. Finally, S2b2C can result in win-win, win-lose, lose-win, or lose-lose outcomes forfirms and consumers,depending on the consumer-segmentation structure and the value consumers place on social interactions.
Brands increasingly adopt two strategies in tandem: They use artificially intelligent agents, such as service robots, to more efficiently serve customers, and they engage in corporate social responsibility (CSR) activities to positively shape brand-related perceptions and behaviors. In this research, across four studies (including a pilot study, an online interaction with an AI-powered robot, and three preregistered studies, involving real and fictitious brands and intentional and actual behavior) the authors provide evidence for the notion that these two strategies are not necessarily compatible. The authors reveal the counterintuitive and nuanced finding that brands engaging in high-fit CSR suffer more from the introduction of service robots than brands that engage in low-fit CSR. This outcome is explained by a perceived lack of alignment in consumers' attributions regarding brands' motivations behind the two strategies, especially regarding the brand's profit-maximization intentions. The authors further identify two managerially actionable moderators. First, a communal service robot strategy can mitigate the negative impact of service robots in instances of high-fit CSR. Second, a more communal CSR implementation (i.e., volunteering time compared with donating money) can lead to a negative impact of service robots when a firm pursues a low-fit CSR strategy.
Voice shopping is on the rise. However, with the absence of visual stimuli, consumers face difficulties in processing brand and product information during the shopping process. Drawing on the load theory of attention and the literature on imagery and information processing, the authors conducted two studies to examine how the level of imagery in auditory product descriptions and the number of product attributes shape consumer responses. The results indicate that high-imagery descriptions increase brand stimulation, which in turn improves brand attitude, purchase intention, and actual product purchase. Additionally, the imagery level of an auditory product description positively influences brand recall directly. These results provide new insights into how auditory message design shapes consumer evaluation and decision-making in voice-based shopping contexts.
Technological advances have enabled firms to personalize advertisements, recommendations, products, and services to individual consumers. This article shows that personalization can backfire when it embarrasses consumers. Study 1 demonstrates that consumers are embarrassed and respond less favorably to retailers offering a personalized shopping experience when they are purchasing a stigmatized product (weight-loss medication) but not when they are purchasing a neutral product (headache medicine). Study 2 similarly shows how personalization can backfire in the context of an online music streaming service; consumers respond less favorably to a service that recommends playlists associated with dissociative identities when they believe the recommendations were personalized. Lastly, Study 3 demonstrates this effect using a field experiment in which participants believe they are seeing a personalized or nonpersonalized advertisement evoking a stigmatized condition (poor skin) or positive condition (clear skin). When consumers received personalized product or service associated with a dissociative or stigmatized identity, they feel more embarrassed, which causes them to respond less favorably toward the business compared with if they had received a nonpersonalized product or service.
Every year, several thousands of marketing articles are published in academic journals, often with the aim of disseminating new insights not only to the academic community but also to managerial practice. However, there is wide acknowledgment of gaps in the marketing science value chain, hindering the flow of marketing knowledge to other researchers and managers. The authors posit that interactive, research-driven (IRD) apps that provide a deeper understanding of the usability of the research contribution are a viable solution to improve the diffusion of marketing knowledge. They shed light on the motivations and barriers to develop IRD apps as well as the market potential and impact of IRD apps through a multimethod examination, which includes interviews, secondary data analyses, and experimental studies. The authors find an untapped potential of IRD apps among published articles and complementary evidence of their value to researchers and managers. They further provide a tutorial to guide the development of IRD apps that complement static research papers and close with a forward-looking section about the future of IRD apps.
