12  Online Environments

Marketing now plays out in environments that are computer-mediated: search engines, social platforms, marketplaces, review sites, mobile apps, livestreams, and the augmented and artificially intelligent interfaces layered on top of them. These environments are not merely new channels through which old marketing flows. They change what consumers do, what firms can observe, and what researchers can measure. A consumer’s path to purchase is now studded with digital touchpoints that leave traces—clicks, queries, posts, ratings, dwell times—and those traces constitute a data stream of unprecedented granularity. The central premise of this chapter is that online environments are simultaneously a substantive domain (consumer and firm behavior changes online) and a methodological opportunity (the residue of that behavior is observable at scale).

The chapter is organized around the life cycle of digital marketing knowledge. We begin with how researchers collect data from online environments—application programming interfaces (APIs) and web scraping—and the validity threats and ethical constraints that attend them, because every empirical claim that follows depends on data gathered this way. We then trace the intellectual history of digital, social media, and mobile (DSMM) marketing, which supplies the organizing constructs for the rest of the chapter. The body proceeds channel by channel and mechanism by mechanism: email; social media and its functional anatomy; the firm-versus-user-generated content distinction; images as a data type; mobile and smartphones; online reviews and word of mouth; the e-marketplace and clickstream; search engines; livestreaming, augmented reality, and artificial intelligence as emerging interfaces; customer engagement; and the wisdom of crowds. By the end the reader should be able to situate a digital marketing phenomenon within the literature, identify the construct and the identification strategy a credible study of it requires, and reason about the data-collection and ethical constraints that bound what can be learned.

A note on scope and connective tissue. Online environments are where many constructs developed elsewhere in this book are now measured and contested. Brand image and brand-related online activity (@sec-branding) are increasingly inferred from images and posts; word-of-mouth and diffusion (@sec-innovation) are the engine of social platforms; advertising response (@sec-advertising) is estimated from clickstream and field experiments; and the marketing–finance link (@sec-marketing-finance) is tested by mapping social media metrics to firm value. This chapter is therefore best read as the empirical and environmental counterpart to those construct-focused chapters.

12.1 Collecting Data from Online Environments

Before any substantive claim can be made, data must be gathered—and online data are gathered in ways that introduce their own validity threats. Boegershausen et al. (2022) provide the methodological foundation: APIs and web scraping let researchers study genuinely new phenomena, raise the ecological validity of marketing research by observing behavior in situ rather than in the laboratory, advance methods for handling a new class of data, and improve measurement of constructs that were previously available only through self-report. The promise is real, but it is conditional on the data being collected and documented carefully.

12.1.1 Two Collection Modes: APIs and Scraping

An API is a sanctioned interface through which a platform returns structured data in response to programmatic requests; web scraping extracts data from rendered web pages that were designed for human consumption. APIs are cleaner and contractually safer but expose only what the platform chooses to expose; scraping can reach anything visible but is brittle, legally fraught, and ethically constrained. A firm rule follows directly: researchers should not create fake accounts to scrape data, because doing so violates platform terms and the norms of research ethics. The choice between modes is not merely technical—it determines the population the data represent and the inferences they will support.

12.1.2 Threats to Internal Validity

Online data are generated by systems that are themselves changing, and two threats to internal validity are endemic to them.

The first is algorithmic interference. The data a platform shows are filtered by sorting and recommendation systems, so what a researcher observes is the joint product of consumer behavior and an algorithm optimizing for the platform’s own objectives. Xu, Zhang, and Zhou (2019) demonstrate that ranking and recommendation algorithms confound observational inferences: an apparent effect of position or popularity may reflect the algorithm’s choices rather than consumer preference. Because the algorithm is typically unobserved and time-varying, it acts as a lurking variable that no amount of cross-sectional control can fully remove.

The second is inter-temporal instability. Data sources change—platforms alter their interfaces, fields appear and disappear, definitions drift—so a panel assembled over months may not measure the same construct at its endpoints. The practical consequence is that longitudinal online data require explicit documentation of when and how each variable was collected, lest a measurement artifact be mistaken for a substantive trend.

A discipline for reproducible collection

Following Gebru et al. (2021), researchers should treat data collection as a documented process, not an undocumented act. Three practices make online data research-grade: (i) a logbook that records important events during collection (outages, interface changes, sampling decisions); (ii) formal documentation of the collection process itself, in the spirit of “datasheets for datasets” (Gebru et al. 2021), so a reader can reconstruct what was gathered and why; and (iii) long-term archival of the data in a citable repository (Zenodo, Dataverse and the Harvard Dataverse, DataCite, re3data), released under a Creative Commons license where permissible so others can reuse it. These practices are the online analogue of a pre-registered protocol: they fix the chain of custody between the world and the dataset.

The substantive and technical challenges of computer-mediated research can be organized into three stages—data gathering (with its consumer-privacy and legal/ethical constraints on scraping), data analysis (feature extraction and dimensionality reduction on high-dimensional text, image, and behavioral data), and interpretation—and these stages frame the substantive challenges that recur below: sentiment-based and location-sensitive targeting, the rise of computers acting as consumers, and pervasive privacy risk (Yadav and Pavlou 2014).

12.2 The Evolution of Digital, Social Media, and Mobile Marketing

The modern field did not arrive all at once. Lamberton and Stephen (2016) review two decades of academic work on DSMM marketing (2000–2015) and recover a structure that organizes the entire chapter: across the period, digital and social media served three recurring functions—a facilitator of individual expression (self-expression and communication), a decision-support tool, and a market-intelligence source—and the relative emphasis among these shifted across three eras as internet penetration grew.

In the first era (2000–2005), online communities enabled self-expression while decision aids reshaped choice. Häubl and Trifts (2000) show that interactive decision aids let consumers reach higher-quality alternatives at lower search cost and make better choices; Brynjolfsson, Hu, and Smith (2003) find that online retailers face stronger price competition than offline retailers, yet the internet boom produced neither a ruinous price war nor pervasive choice overload. In the second era (2005–2010, as penetration crossed 50%), online word of mouth (WOM) became central and the decision-support and market-intelligence functions merged: customers acquired through WOM exhibit greater long-term value than conventionally acquired customers (Villanueva, Yoo, and Hanssens 2008; Trusov, Bucklin, and Pauwels 2009a), and a parallel methodological literature learned to identify influential users and the structure of diffusion in social networks (Katona and Sarvary 2008, 2010; Trusov, Bodapati, and Bucklin 2010). In the third era (2011–2015, with penetration near 80%), consumers became markedly more expressive and user-generated content (UGC) matured into a marketing instrument; at the individual level, content consumption and production trade off against each other, so heavier consumers produce less and vice versa (Ghose and Han 2011). Two features recur across eras: papers increasingly carry methodological keywords (clickstream, online chat), signaling that new data types demanded new methods, and academic attention tracked practitioner attention closely.

Kannan and Li (2017) supply a complementary digital marketing framework synthesizing the field; Sridhar and Fang (2019) extend it to digital, data-rich, and developing (“D3”) markets; and Verhoef and Bijmolt (2019) examine the digital business models that these environments enable. Yadav and Pavlou (2014) review marketing in computer-mediated environments specifically, and Grewal, Gupta, and Hamilton (2021) broaden the lens to multimedia and modality, mapping how platforms (Twitter, Facebook, Instagram) and sensory modalities (visual, auditory, haptic, olfactory, gustatory) jointly shape consumer response.

Several focused studies sharpen the picture of how content creation itself affects experience and value. Tonietto and Barasch (2020) show that generating content during an experience increases immersion and accelerates perceived time, which in turn raises enjoyment; consumers prompted by incentives or social norms obtain the same experiential benefits as those who create content organically. At the firm level, Lacka et al. (2021) estimate the price impact of firm-generated tweets among S&P 500 information-technology firms, defining price impact as the impact on the variance of stock price: tweets carrying only valence or only subject matter (consumer or competitor orientation) produce merely temporary price impacts, whereas tweets carrying both attributes produce permanent impacts, with negative-valence tweets about competitors generating the largest permanent effect. Firm responses can also backfire: Golmohammadi et al. (2021) show that responding to social media complaints increases the public exposure of those complaints, and a high-publicization response strategy (detailed, multi-exchange replies) can decrease perceived quality and firm value, blunt the firm’s own posts, and increase the volume of future complaints—effects stronger for firms targeted by retail investors. Figure 12.1 traces how the emphasis among these functions has shifted across eras of rising internet penetration.

flowchart TB
  subgraph F["Three recurring functions"]
    A["Facilitator of<br/>individual expression"]
    B["Decision-support<br/>tool"]
    C["Market-intelligence<br/>source"]
  end
  E1["Era 1 (2000–2005)<br/>online communities;<br/>decision aids"] --> F
  E2["Era 2 (2005–2010, ~50%)<br/>online WOM;<br/>support + intelligence merge"] --> F
  E3["Era 3 (2011–2015, ~80%)<br/>expressive consumers;<br/>UGC as a tool"] --> F

Figure 12.1: Three functions of digital and social media marketing and their shifting emphasis across eras of rising internet penetration, after Lamberton and Stephen (2016).

12.3 Email

Email is the oldest digital direct-marketing channel and remains a clean testbed for personalization because messages can be randomized and outcomes (opens, clicks, purchases, unsubscribes) precisely logged. The evidence favors personalization, but the mechanism is subtle. Sahni, Wheeler, and Chintagunta (2018) find through field experiments that adding consumer-specific information—most simply, the recipient’s name—raises opening and sales and reduces the probability of unsubscribing, so personalization improves the contact rather than merely exploiting it. The effect generalizes beyond commerce: Munz, Jung, and Alter (2020) show that similarity between a donor and a donation recipient—a shared name, or even a shared first letter of the surname—encourages donors to open appeals and give more, evidence that the implicit-egotism mechanism behind name-based personalization operates in prosocial as well as commercial settings.

12.4 Social Media

A working definition fixes the object of study. Appel et al. (2020) define social media as “a collection of software-based digital technologies—usually presented as apps and websites—that provide users with digital environments in which they can send and receive digital content or information over some type of online social network” (p. 80), and emphasize that it is best understood as a “technology-centric—but not entirely technological—ecosystem in which a diverse and complex set of behaviors, interactions, and exchanges involving various kinds of interconnected actors (individuals and firms, organizations, and institutions) can occur” (p. 80).

Social media is a technology-centric—but not entirely technological—ecosystem in which a diverse and complex set of behaviors, interactions, and exchanges involving various kinds of interconnected actors can occur.

— Appel et al. (2020), p. 80

The present state of social media is described along two axes: the platforms (major and minor, established and emerging) that supply the underlying technology and business model, and the use cases that describe how people and organizations use them. Appel et al. (2020) categorize use into communication between known others (family, friends), communication among strangers who share interests, and access to or contribution of digital content (news, gossip, reviews). Their forward look structures the field’s open questions across three horizons: in the immediate future, an omni-social presence in which sharing is embedded in every digital tool, the rise of (micro and virtual) influencers, and intensifying privacy concerns; in the near future, social media as a tool against loneliness, as integrated customer care, and as a political instrument; and in the far future, increased sensory richness through augmented and virtual reality, complete online–offline convergence, and content generated by non-humans (AI and social bots), which complicates attribution and erodes trust as bought likes and shares proliferate.

12.4.1 The Functional Anatomy of Social Media

Kietzmann et al. (2011) decompose social media into seven functional building blocks—identity (self-disclosure), conversation, sharing, presence, relationships, reputation (of people and of content), and groups—and argue that firms should manage each block deliberately. Their managerial counterpart is the four Cs: cognize (recognize and understand one’s social media landscape), congruity (develop strategies congruent with each platform’s functionality), curate (act as curator of interactions and content), and chase (track the conversation and information flow). The contribution is to replace a one-size-fits- all “be on social media” posture with a functional diagnosis: different blocks are salient on different platforms, and strategy should match the blocks that matter. Figure 12.2 lays out the seven blocks firms must diagnose before acting.

mindmap
  root((Social media))
    Identity
    Conversation
    Sharing
    Presence
    Relationships
    Reputation
    Groups

Figure 12.2: The seven functional building blocks of social media (Kietzmann et al. 2011), each a distinct lever firms must diagnose before acting.

12.4.2 Platform Differentiation and Motivation

Why do consumers choose one platform over another, and what does that imply for where firms should invest? Pelletier et al. (2020) ground the answer in uses and gratifications theory, which holds that the benefits a person extracts from a medium depend on the purpose for which they consume it (Severin, Tankard, et al. 1997). Mapping four motivations—social, informational, entertainment, and convenience—onto platform affordances, they find that consumers seeking information gravitate to Twitter (real-time text), that social purposes draw them to Twitter and Instagram, and that Instagram dominates for entertainment and for co-creating with brands, because users seeking hedonic experience prefer visual, low-text content. Their most consequential finding is a corrective to practitioner intuition: Facebook—the largest network and the one most favored by marketers—shows the lowest usage and co-creation intentions, so “one size does not fit all,” and consumers who want to co-create migrate to platforms with the broadest array of communication mechanisms. Because consumers co-create brand stories on these platforms, social media can build or destroy brand equity depending on the fit between platform affordances and the firm’s objective.

12.4.3 Listening, Sentiment, and the Metrics Problem

Firms increasingly listen to social media to gauge sentiment, but naive listening is biased. Schweidel and Moe (2014) jointly model the sentiment expressed in posts and the choice of venue in which to post, showing that both the content of a post and the sentiment toward the brand affect where it is written; because venues differ systematically in who posts and how, common monitoring approaches that pool across venues yield misleading brand-sentiment metrics. The payoff of correcting for this selection is predictive: a model-based measure of brand sentiment serves as a leading indicator of external performance. The broader literature on social media metrics and dashboards (Peters et al. 2013) makes the same point at the strategic level—that metrics must be theory-driven to be decision-relevant.

12.4.4 Owned versus Earned Social Media

A central strategic distinction separates owned social media (OSM), brand-controlled content, from earned social media (ESM), exposure generated by voluntary user mentions and recommendations the firm neither produces nor controls. Colicev et al. (2018) trace both to consumer mindset metrics and ultimately to shareholder value using a persistence-modeling framework (Dekimpe and Hanssens 1995) applied to brands that follow a corporate-branding strategy and are listed on a US exchange. Their findings are nonlinear and mechanism-specific: the volume of ESM engagement raises brand awareness and purchase intent but not customer satisfaction, whereas the valence (positive and negative) of ESM most strongly moves satisfaction; OSM lifts brand awareness and satisfaction but not purchase intent in general, raising purchase intent only for high-involvement utilitarian brands and reputable brands—suggesting that socially responsible practice lends OSM legitimacy. Purchase intent and satisfaction in turn raise shareholder value, completing a mediated path from social media activity to firm value that earlier work had posited but not traced explicitly. The construction of ESM measures—cumulative fan following, engagement volume (Facebook “people talking about this,” retweets, video views), and a composite volume–valence metric scored by a naïve Bayes classifier—illustrates the measurement craft the chapter’s data-collection section anticipated, with mindset metrics drawn from survey panels and brand-level heterogeneity coded in the spirit of Bart, Stephen, and Sarvary (2014).

12.4.5 Brand Buzz in the Echoverse

Hewett et al. (2016) situate social media within the echoverse—“the entire communication environment in which a brand/firm operates, with actors contributing and being influenced by each other’s actions” (p. 1)—whose actors (firms, news media, online WOM, consumer sentiment, business outcomes) form a feedback system. Estimating a vector autoregression with Granger-causality tests on longitudinal US financial-services data, they recover an asymmetric structure with several managerially sharp implications. Firm communications influence traditional media, consumer sentiment, and business outcomes; advertising affects firm outcomes directly but does not move traditional media, online WOM, or sentiment. Negative news propagates through a self-reinforcing spiral, but timely tweets and press releases can arrest it. Online WOM affects consumer sentiment and business outcomes, but not the reverse, so firms rationally attend more to WOM than to survey sentiment. A cautionary case study shows that an extremely positive tone can backfire in a negative firestorm, whereas a consistent, higher-volume stream of moderate messages performs better.

12.4.6 Social Media in B2B and Across Sectors

Social media marketing is not confined to consumer brands. Salo (2017) reviews business-to-business (B2B) social media use; Buratti, Parola, and Satta (2018) show it offers an accessible, low-cost route to competitive advantage even in conservative sectors such as tanker shipping and ocean carriage; and Kärkkäinen, Jussila, and Väisänen (2010) document that B2B firms use social media less than B2C firms but that the front-end of new-product development and the launch phase offer the highest social media innovation potential for them. On the response side, Herhausen et al. (2022) show that active listening and empathy in a firm’s reply evoke gratitude in high-arousal customers even when the underlying failure is not yet recovered.

12.4.7 Path to Purchase, Substitution, and Word-of-Mouth Value

Social media shapes the consumer journey along the classic know–feel–do pathway. For low-involvement fast-moving consumer goods, Srinivasan, Rutz, and Pauwels (2015) find that earned media drives sales and that, on top of traditional drivers (distribution, price), both earned and owned social media explain part of the path to purchase—though higher consumer activity on earned and owned media can also provoke disengagement (unlikes). The temporal relationship between social media and shopping is itself nonmonotonic: Y. Zhang et al. (2017) find that social media usage is positively correlated with shopping activity in the long run but that, immediately after a social media session, online shopping activity is lower, consistent with attention being temporarily diverted. Finally, the monetary value of WOM can be quantified: Trusov, Bucklin, and Pauwels (2009b) show that WOM referrals have stronger and longer carryover effects and higher response elasticities than traditional marketing, and that their value can be computed from the advertising revenue of the impressions they generate— a figure that serves as an upper bound on the financial incentive a firm should offer to encourage WOM.

12.4.8 Engagement, Influence, and the Frontier of Social Media Research

Recent work pushes into engagement measurement, influencer economics, content strategy, and disclosure. Yang, Zhang, and Zhang (2025) introduce a Product Engagement (PE) score that estimates pixel-level engagement in influencer video ads using a deep 3D convolutional network and an object-detection algorithm, then validate it on TikTok ads against Taobao sales: the score predicts sales lift and, crucially, distinguishes engagement driven by the product from engagement driven by the influencer, which matters for aligning incentives and pricing content. N. Li, Haviv, and Lovett (2024) exploit YouTube’s “adpocalypse” supply shock as a natural experiment and find that influencer videos act as substitutes for video-game purchases but as complements for game usage, so firms should align monetization (in-game purchases versus subscriptions) with whichever effect dominates. Huang and Lin (2024) show that displaying the first “like” on an ad sharply raises both likes and clicks, but additional likes mainly raise further likes rather than clicks; more likes help the clickthrough of lesser-known but not well-known brands, evidence of dual informational and normative social influence.

On content strategy, Lysyakov et al. (2024) develop a similarity measure of rival firms’ social media content and find that retailers who compete closely offline diverge in their Twitter content, and that divergent firms gain higher engagement and faster follower growth by exploiting the platform’s advanced functionalities more effectively. S. Wang and Greenwood (2025) leverage the phased rollout of post-length-limit extensions on a major Chinese microblogging platform and find that longer limits raise daily posting—driven mainly by already-active users—and raise likes per post, improving content quality but further concentrating content generation among a vocal minority, making the silent majority less visible. On disclosure, Ershov, He, and Seiler (2024) analyze over 100 million posts with a novel method for detecting undisclosed sponsorship and estimate that 96% of sponsored posts are undisclosed (82% under conservative assumptions), a rate that barely declines despite enforcement and that consumers frequently fail to recognize—an inconvenient fact that any study of influencer marketing must confront.

12.5 Who Generated the Content? Firm- versus User-Generated Content

A recurring fault line in digital marketing is who produced a piece of content, because source governs how it is processed and trusted. Firm-generated content (FGC) is controlled and polished but discounted for self-interest; user-generated content (UGC) is credible but uncontrolled. Colicev, Kumar, and O’Connor (2019) formalize the comparison through information processing and source credibility, mapping each type onto the marketing funnel: UGC correlates strongly with awareness and satisfaction, whereas FGC—especially its vividness—is more effective for consideration and purchase intent, with UGC valence dominating UGC volume at these stages. Brands with higher corporate reputation exhibit stronger relationships between FGC dimensions and the funnel, echoing the legitimacy logic of the OSM/ESM results above. The foundational observation that negative WOM behaves differently from positive WOM—both observationally and experimentally—predates this framework (Sen and Lerman 2007) and motivates the valence-centered measurement it uses. Figure 12.3 summarizes how the two content sources map onto the funnel stages.

flowchart LR
  UGC["User-generated content<br/>(credible, uncontrolled)"] --> AW[Awareness]
  UGC --> SAT[Satisfaction]
  FGC["Firm-generated content<br/>(controlled, discounted)"] --> CON[Consideration]
  FGC --> PI[Purchase intent]

Figure 12.3: Firm-generated and user-generated content map differently onto marketing-funnel stages (Colicev, Kumar, and O’Connor 2019): UGC dominates awareness and satisfaction; FGC vividness aids consideration and intent.

A growing literature studies the hybrid and the responsive. Beck et al. (n.d.) examine firm-generated user content (FGUC)—brand-crafted content that consumers are encouraged to share—and find it increases sharing by lowering effort, but asymmetrically: satisfied consumers share it while dissatisfied consumers resist it because it misaligns with their experience, so FGUC amplifies positive WOM and suppresses negative WOM and should be deployed only when experiences are likely to be positive. On the firm’s reply, Saljoughian et al. (n.d.) build on dialogic-listening theory—empathetic understanding, unconditional positive regard, mutual understanding, presentness, and genuineness, plus shifting to private channels and signing replies with an agent’s name—and, analyzing roughly a million tweets across 206,000 threads at four US banks over a decade, show that specific response elements significantly improve subsequent user sentiment, with effectiveness contingent on the user’s prior sentiment. The managerial lesson is to match response tone and structure to initial sentiment rather than applying a uniform playbook.

12.6 Images as a Data Type

Images are a high-dimensional, increasingly tractable data source for marketing, and a body of work now extracts marketing-relevant features from them with computer vision (connecting to brand image, @sec-branding). The competition for attention in such data-rich environments is fierce (Wedel and Kannan 2016), and even the design of feature ads can be optimized: Pieters, Wedel, and Zhang (2007) show that optimal allocation across brand, text, pictorial, price, and promotion elements is achievable without increased cost, and propose two entropy-based measures of clutter that characterize the salience of feature ads through attention-engagement theory.

12.6.1 Engineering a Score from Pixels

A representative pipeline scores the visual potential of an image from interpretable features. Structural complexity—visual complexity due to edges and detail (Pieters, Wedel, and Batra 2010)—is measured by Canny edge detection (Forsythe, Sheehy, and Sawey 2003), because edge information correlates highly with perceived complexity. Color complexity (Reinecke et al. 2013) is captured by a color-variety algorithm (Bing Zhou, Shuang Xu, and Xin-xin Yang 2015) that, for an image with \(m\) distinct colors, \(n_i\) pixels of color \(i\), and \(N\) pixels in total, computes the entropy-like quantity \[ C \;=\; \sum_{i=0}^{m} n_i \,\log\!\left(\frac{n_i}{N}\right), \tag{12.1}\] which rises with the number and balance of colors present. Image permanence—how well an image persists in memory after viewing—is scored by the AMNet memorability index (Khosla et al. 2012, 2015; Fajtl et al. 2018), validated to near-human consistency (Leyva and Sanchez 2021). The pipeline is completed by the number of faces (via a vision API), the aspect ratio (height over width), the average HSV values across pixels, and themes recovered by latent Dirichlet allocation over image labels. Each feature is interpretable and cheap to compute at scale, which is what makes image-based scoring practical.

12.6.2 What Visual Features Do to Engagement

The mapping from features to engagement is contingent on platform and category. Y. Li and Xie (2019) find a robust positive mere-presence effect of imagery on engagement and that high-quality, professionally shot pictures consistently raise engagement on both Twitter and Instagram, but that colorfulness matters only in some categories and that human faces and image–text fit raise engagement on Twitter but not on Instagram—so the value of a visual feature is platform-specific. For visual UGC, Jalali and Papatla (2016) operationalize color composition through hue, chroma, and brightness and find that click rates are higher for photos with more green and less red and cyan, and with higher chroma of red and blue. In the marketplace, professional imagery has measurable demand consequences: S. Zhang et al. (2022) use deep learning on 7,423 Airbnb properties over 16 months and find that verified, professionally photographed listings achieve 8.98% higher occupancy than host-photographed ones, with twelve composition-, color-, and figure-ground attributes significantly related to demand. Appearance also drives labor-market matching: Troncoso and Luo (2022) use computer vision and choice models on more than two million applications to 63,014 Freelancer.com jobs and show that freelancers who look the part are more likely to be hired regardless of actual performance, demographics, or attractiveness, with the effect strongest where reputation systems fail to differentiate candidates and only partially counteracted by platform recommendations.

12.7 Mobile and Smartphones

Mobile devices are not merely small screens; they are continuously present, deeply personal, and capable of new touchpoints in the customer journey. Rangone and Renga (2006) offer an early framework for mobile advertising. The strategic question of whether mobile investments create firm value is addressed by Boyd, Kannan, and Slotegraaf (2019), who use stock-market returns to show that announcing a branded mobile app raises firm value, and that the features emphasized in the app’s design—peer-to-peer interaction about the brand, personal brand–customer interaction, and the purchase phase—significantly shape how much value is created. Monetizing apps is harder than launching them: Cao, Chintagunta, and Li (2023) run a large-scale randomized field experiment on a non-advertising app and find that a “soft landing” (continued limited free access after monetization) and exclusive secondary offerings to subscribers each reduce willingness to subscribe in isolation, but interact positively when combined, because the soft landing reframes monetization as fair rather than aggressive—an interaction that generalizes in a follow-up experiment.

Smartphones also carry emotional weight. Melumad and Pham (2020) show that consumers derive not only functional but emotional benefits from their phones: under stress, consumers turn to their smartphones for greater stress relief, and this pacifying effect is stronger for one’s own phone than for another’s—evidence that the device functions as a personal attachment object, not merely a tool.

12.8 Online Reviews and Word of Mouth

Online reviews are the most studied online data type in marketing because they are abundant, structured (a rating plus text), and directly tied to sales. The literature can be organized around four questions: who writes reviews and how expertise shapes them; how text and emotion in reviews are processed; how reviews move sales across products and time; and how the integrity of the review ecosystem can fail.

12.8.1 Expertise, Emotion, and Helpfulness

Expertise systematically changes both the production and the effect of reviews. As people gain expertise they become more responsive to negative feedback—novices respond to positive feedback while experts seek and respond to negative feedback—a pattern robust across domains (Finkelstein and Fishbach 2012). Expertise also flattens evaluations: Nguyen et al. (2020) find that greater expertise yields less extreme evaluations and smaller effects on a brand-valence metric that affects page rank and consumer evaluation, so an excellent experience may earn a lower rating from an expert than from a novice, and observationally experts are more likely to post negative reviews. Emotional language is a separate lever: Folse et al. (2016) define negatively valenced emotional expression (intense language, all caps, exclamation points, emoticons) and show that such reviews are seen as more helpful and damage product attitudes when written by experts, but provoke a negative self-reflection and leave product attitudes unchanged when written by novices, with language complexity moderating the effect of expertise on trustworthiness.

The form of the text matters for measurement. Büschken and Allenby (2016) propose a sentence-based topic model for consumer reviews—rather than the usual word-based analysis—and show on hotel and restaurant data (Expedia, we8there) that sentence-based topics are more coherent and distinguished and improve inference and prediction of consumer ratings. Trust in reviews also depends on the type of purchase: Dai, Chan, and Mogilner (2019) show that reviews are less trusted for experiential than for material purchases, because ratings of experiences are less representative of objective quality; relatedly, X. (Shane). Wang et al. (2021) use machine learning and NLP to extract and monitor product attributes from review text, building embedded representations that characterize attributes by their surrounding words and cluster them into multi-level benefit abstractions, closing the gap between concrete engineered attributes and abstract meta-attributes that surveys capture inconsistently and slowly.

12.8.2 How Reviews Move Sales

The reviewer’s rank, the timing of reviews, and brand type all shape sales effects. Counterintuitively, Yazdani, Gopinath, and Carson (2018) find—using instrumental variables on 182 music albums and Amazon review data—that bottom-ranked reviewers move sales more than top-ranked reviewers, whose influence concentrates in new releases and in products with high review variability, with the disparity attributable to content and reviewer identity and robust across categories and metrics. The first review is especially fateful: Park, Shin, and Xie (2021) show that valence and volume are not independent and that a positive first review creates a long-term advantage in future WOM valence and volume while a negative first review imposes a long-term disadvantage, entrenched by information-availability bias that makes early negativity hard for firms to correct. Reviews of one product generation also spill onto another: L. Li, Gopinath, and Carson (2022) find that the valence of a previous generation’s reviews raises current-generation sales (amplified when current-generation reviews are uncertain or favorable) while current-generation valence depresses the previous generation’s sales.

Brand type moderates these effects. Hoskins et al. (2021), using US beer data, show that niche brands are more affected by online WOM because consumers hold weaker prior brand imagery, that reviewers favor local niche brands, that the online community’s sentiment outweighs professional critics’ for the average reviewer, and that prior reviews gain traction when the reviewer’s expertise is high or shares geography with the next reviewer. Identity-relevant brands are partially insulated: Ordabayeva, Cavanaugh, and Dahl (2022) show that negative reviews from socially distant (but not socially close) individuals are less harmful to identity-relevant brands, because a threat to an identity-relevant brand prompts consumers to strengthen their bond with it—an effect absent for positive reviews or irrelevant brands. Not all rating systems move sales: Nishijima, Rodrigues, and Souza (2021) find that Rotten Tomatoes’ “fresh”/“rotten” categorization has no effect on box-office performance.

12.8.3 Herding, Polarity, and Sequencing

Reviews are not independent draws. Sunder, Kim, and Yorkston (2019) document substantial herding in online ratings driven by reference groups (the crowd and friends): as raters gain experience the crowd’s influence wanes while friends’ influence grows, divergent reference-group opinions produce varied herding, and a firm’s diverse product portfolio both raises perceived quality and dampens social influence on ratings. Schoenmueller, Netzer, and Stahl (2020) identify polarity self-selection—the people who post are disproportionately the very satisfied and the very dissatisfied—which reduces the informativeness of the average rating. Sequencing effects appear even between rating and tipping: Chen et al. (2022) show that when customers rate before tipping they tip less, because they feel they have already rewarded the provider through the rating, an effect stronger when they tip from their own pocket, have flexible categorization, or believe the provider benefits from ratings—and mitigated by highlighting the consistency motive between rating and tipping. Q&A sections answered by other consumers and sellers improve product–consumer fit and reduce mismatch-driven negative reviews (Banerjee, Dellarocas, and Zervas 2021).

12.8.4 When the Review Ecosystem Fails

The credibility of reviews can be undermined from within. He, Hollenbeck, and Proserpio (2022) document a sizable market in which fake reviews are bought and sold (often through private Facebook groups): fake reviews raise ratings and sales in the short run, most large firms abstain, marketplaces struggle to police them promptly, and when firms stop buying fake reviews their ratings fall—driven by a rising share of one-star reviews— especially for young, low-quality products. The integrity question extends to whether review activity reflects genuine support: Sharma, Frake, and Watson (2025) exploit George Floyd’s murder as an exogenous shock and, using a difference-in-differences design on Yelp reviews and SafeGraph foot traffic, distinguish symbolic support (review frequency and sentiment) from substantive support (revenue and visits) for Black-owned businesses, finding a significant rise in symbolic support—especially from White and Democrat-leaning consumers—but no substantial rise in substantive support, plus more “not recommended” reviews suggestive of performative engagement.

12.8.5 Slacktivism

The Black-Lives-Matter results connect to slacktivism: supporting a political or social cause through social media or online petitions with minimal commitment. Silence, however, is not costless. Qin et al. (2024) study corporate silence on visible social issues (Blackout Tuesday) and find that staying silent significantly reduces consumer engagement—slower follower growth and fewer likes on Instagram—with peer activism amplifying the backlash to silence and narrow-market-niche firms facing less backlash than broad-market firms. The managerial implication is that firms must weigh issue salience, peer actions, and market positioning when deciding between silence and activism on contentious issues.

12.9 The e-Marketplace and Clickstream

The e-marketplace is where browsing and buying are jointly observed, and clickstream data—the sequence of pages a user views—make the path to purchase visible. Early descriptive work established the terrain: Dholakia and Rego (1998) document the marketing information on commercial home pages and the factors that drive hit rates, drawing on the information-content paradigm to assess informativeness, while Burke (2002) surveys 2,120 consumers across 128 facets of shopping and finds general satisfaction with convenience, quality, selection, and value but dissatisfaction with service, product information, and speed, concluding that new interface technologies help when customized to consumer segments and product categories. Shoppers are heterogeneous: Rohm and Swaminathan (2004) segment them by shopping motive (convenience, store orientation, information use, variety seeking) into convenience shoppers, variety seekers, balanced buyers, and store-oriented shoppers.

12.9.1 Modeling Browsing Paths

Clickstream models turn navigation into prediction. Montgomery et al. (2004) propose a dynamic multinomial probit model of web browsing on data from an online bookseller and show that a memory component is essential: the dynamic model predicts paths far better than static multinomial probit or first-order Markov models, because a user’s path reveals their goal. Practically, purchasers can be predicted with over 40% accuracy after only six page views, against a 7% base rate—evidence that path data personalize web design and product suggestions. The browsing literature complements this with individual-level models: Bucklin and Sismeiro (2003) model browsing from clickstream and show that the propensity to browse depends on visit depth and repeat visits and that aggregate site metrics cannot capture individual behavior. Context matters too: Mallapragada, Chandukala, and Liu (2016) use a multivariate mixed-effects Type II Tobit model on 773,262 sessions yielding 9,664 transactions across 43 categories and 385 sites, finding that a site’s product-variety breadth raises visit length and basket value but lowers page views, and that communication functionality raises basket value especially for hedonic products.

12.9.2 Real-Time Adaptation and Cross-Selling

The frontier is acting on observed intent in real time. Ding, Li, and Chatterjee (2015) develop a dynamic model that learns a shopper’s unobserved real-time intent from cart actions and instantly adapts the web page, lowering cart abandonment by 32.4% and raising conversion by 6.9% when adaptation begins after the first page view, with the most efficient intervention point after three page views—balancing learning against timeliness. Sequential structure also organizes cross-selling. S. Li, Sun, and Wilcox (2005) use a structural multivariate probit model on bank data to show that customers buy products in a natural order along a financial-maturity continuum, with women and older customers more influenced by overall satisfaction and more-educated or male-led households progressing faster. S. Li, Sun, and Montgomery (2011) extend this with a stochastic dynamic programming framework that delivers the right product to the right customer at the right time through the right channel: cross-selling solicitations serve mainly an educational role (83%) rather than advertising (15%) or immediate promotion (2%), and the optimized policy raises immediate response 56%, long-term response 149%, and long-term profitability 177%.

12.9.3 Auctions, Signals, and Social Influence In-Store

Online auctions revived the classic lemons problem, and platforms responded with quality signals. S. Li, Srinivasan, and Sun (2009) build a taxonomy of internet-auction quality and credibility markers grounded in signaling and auction theory, adapt the Park–Bradlow model to eBay, and confirm that credible signals shape bidding and participation and differ in effect across buyer segments—the first empirical assessment of how extensive auction institutional tools address adverse selection. Social influence operates in physical retail too, observable through video tracking: X. Zhang et al. (2014) use a hierarchical-Bayes bivariate model on video-tracking data to show that direct social interactions (with salespeople or other shoppers) lengthen visits, deepen product interaction, and raise purchase probability, that larger groups are swayed more by group discussion than by strangers, and that moderate crowd density encourages interaction while excessive density degrades the experience. X. Zhang, Li, and Burke (2018) extend this with a three-stage dynamic linear model showing that group conversations shape where shoppers browse, whether they buy, and how much they spend, moderated by group size and cohesion. Herding appears in digital goods as well: Ding and Li (2019) use a simultaneous-equations herding model in a hierarchical-Bayes framework on a Chinese literature platform and find rational herding in both consumption and purchase—stronger for purchase—reduced by product features but amplified by producer reputation and competition.

12.9.4 Peer-to-Peer Platforms

Platform business models reshape how consumers reason about purchases. Costello and Reczek (2020) show that peer-to-peer (P2P) platforms (Uber, Lyft, Airbnb) mediating flows between providers and consumers can shift buyers into an “empathy lens”— thinking about how a purchase affects the provider—rather than the “exchange lens” of money-for-service with a firm, raising purchase-related outcomes when the platform foregrounds providers, a pattern confirmed across field and lab studies and absent in traditional business models.

12.10 Search Engines

Search is where demand is expressed as a query, and the structure of the search-results page shapes behavior. Attention concentrates at the top: Guan and Cutrell (2007) document the “golden triangle,” in which users attend to the top three organic results and spill over to the fourth (often paid). Search behavior is shallow and slow-moving: E. J. Johnson et al. (2004), using panel data on more than 10,000 households across books, CDs, and air travel, find that households visit only about 1.2 book, 1.3 CD, and 1.8 travel sites per active month, that search follows a logarithmic process with limited breadth, and that more-active shoppers search more sites. Search data are also predictive of demand before it materializes: Kulkarni, Kannan, and Moe (2012) show that pre-launch search volume forecasts opening-week motion-picture sales, with advertising data improving the forecast—turning queries into a leading indicator for launch planning.

The design of search aids is a fertile field-experimental domain. Lei, Chen, and Sen (2024) show that firms leveraging an external autocomplete API gain measurably: removing access cut search-suggestion clickthrough rate (CTR) by 4.6%, and an instrumental-variables analysis establishes that a 10% change in suggestion CTR causes a 1.85% change in top-slot results’ CTR, with the negative effect of losing external data halving over the long run as firms accumulate internal data—a strategic trade-off between external data as growth instrument and as constraint on internal capability. Query recommender systems likewise expand demand: Zheng et al. (2023) find that access to a query recommender on an Asian food-delivery app raises 30-day purchase volume by 1–2% and broadens consumption diversity, working in complement with autocomplete by expanding consideration sets while autocomplete refines queries. Finally, sponsored search can cannibalize organic traffic: Moshary (2025) block sponsored ads for 3% of a platform’s visitors and find that sponsored search reduces total transactions yet remains profitable, because ad revenue more than offsets lost commission—so what harms total market efficiency can still be a net financial positive for the platform.

12.11 Emerging Interfaces: Livestreaming, Augmented Reality, and AI

Three interface technologies are reshaping how consumers encounter products, each introducing a distinct psychological mechanism.

Livestreaming fuses entertainment, social presence, and commerce in real time. Lin, Yao, and Chen (2021) show that happy streamers make audiences happier and more generous with tips, an effect amplified for broadcasters who are more experienced, receive more tips, and are more popular—evidence that emotional contagion is a monetizable mechanism. Emotional display does not always help in screen-mediated selling, however: Bharadwaj et al. (2021) extract salesperson emotions from livestream video and find that in one-to-many screen-mediated communication salespeople should not display emotions, a contingency that cautions against importing face-to-face selling norms wholesale.

Augmented reality (AR) overlays digital content on the physical world. Yim, Chu, and Sauer (2017) show that, relative to web-based product presentation, AR delivers superior communication benefits—novelty, immersion, enjoyment, utility—leading to better medium attitudes and purchase intention, with immersion mediating the path from interactivity and vividness to usefulness and enjoyment in the AR condition but not on the web. The construct of vividness, “the ability of a technology to produce a sensorially rich mediated environment” (Steuer 1992, 80), and the two schools of interactivity (a technological view emphasizing speed, mapping, and range; and a subjective-perception view tied to motivation) are the theoretical scaffolding here, and their overlap is a measurement challenge the literature continues to negotiate.

Artificial intelligence introduces agents that consumers reason about differently from humans. Garvey, Kim, and Duhachek (2021) show that consumers respond better to an AI agent than a human when an offer is worse than expected, and better to a human when an offer is better than expected, because AI agents are perceived to have weaker intentions; firms can anthropomorphize AI to claim credit for good offers and deflect blame for bad ones. The visual perception of AI-rendered and animated content also has its own biases: Jia, Kim, and Ge (2020) document an inverse size–speed association in digital ads—faster animation makes an object seem smaller, learned from animate agents such as animals— which firms can mitigate by reducing similarity to those domains, educating consumers, or highlighting size information, and which can even reverse when a positive size–speed association in the base domain is made accessible. A broader typology of digital agents is offered by Miao et al. (2021), who define avatars as “digital entities with anthropomorphic appearance, controlled by a human or software, that are able to interact” (p. 5) and organize them on form realism and behavioral realism. Even the platform on which content is consumed conditions preferences: KOUKOVA, KANNAN, and RATCHFORD (2008) find experimentally that content-consumption preferences (e-book versus print, and so on) vary across platforms.

12.12 Customer Engagement

Engagement reframes the customer as more than a buyer. Kumar et al. (2010) decompose a customer’s engagement value into four components: customer lifetime value (purchase behavior), customer referral value (incentivized referrals), customer influencer value (organic influence on other customers), and customer knowledge value (the value of the customer’s feedback to the firm). Harmeling et al. (2016) visualize the mechanism by which engagement translates into firm performance, completing the conceptual bridge from behavior to value. Engagement is heterogeneous and observable in clickstream: Bucklin and Sismeiro (2003) show that browsing propensity varies with site-visit depth and repeat visits, and Reimer, Rutz, and Pauwels (2014) show that segments respond differently to marketing instruments—heavy users of digital music are less price-sensitive and more responsive to TV advertising than the consumer-packaged-goods pattern would predict, while light users are price-sensitive and prone to opt out of targeted communication. The lesson is that engagement-based segmentation must be estimated, not assumed.

12.13 The Wisdom of Crowds

Online environments enable production and evaluation by distributed crowds—open source, crowdsourcing, and crowdfunding—whose success turns on network structure and social dynamics rather than on a single firm’s effort.

In open source, network embeddedness is double-edged. Grewal, Lilien, and Mallapragada (2006) apply latent-class regression to affiliation networks of open source projects and find that embeddedness profoundly shapes technical and commercial outcomes, but heterogeneously: in some regimes it promotes success and in others it harms it, with project age and visibility offering additional explanation. Mallapragada, Grewal, and Lilien (2012) model time-to-release with a split-hazard model on 817 SourceForge projects and show that a founder’s central position in the developer-user community can speed release by up to 31% and that highly engaged end users cut release time by 11%, confirming the dual-community (developer/end-user) structure of open source. In crowdsourcing, Mahr, Rindfleisch, and J Slotegraaf (2015) apply dual-processing theory and find that creative (spontaneous, intuitive) and deliberate (systematic, analytical) problem-solving styles each succeed under different conditions—creative under high contextual familiarity and short time investment, deliberate under low familiarity and long investment—while combining both styles lowers success.

In crowdfunding, social affiliation can backfire. Herd, Mallapragada, and Narayan (2022) find that while more backers raise funding, affiliation among backers lowers funding amounts and success through “vicarious moral licensing”—potential backers feel less obligated when affiliated others have already funded—an effect attenuated by creator engagement (descriptions, updates) and backer engagement (shares) and robust across instruments, models, and measures. Finally, the emergence of crowd structure resists simple explanation: S. L. Johnson, Faraj, and Kudaravalli (2014) use agent-based modeling on 28 online communities to show that power-law distributions of user popularity arise from no single mechanism (neither preferential attachment nor any one social factor alone) but require a mixture of social mechanisms—a caution against monocausal models of online influence.

12.14 Social Listening and Verification Tools

The practical infrastructure for studying social media consists of social listening platforms (SLPs)—commercial tools such as Brandwatch, Meltwater, Sprinklr, Talkwalker, and others that aggregate, classify, and visualize social media at scale— and a parallel ecosystem of fake-follower and influencer-audit tools that estimate the authenticity of an account’s audience. These tools are not neutral instruments: they embed sampling, language-coverage, and classification choices that shape the sentiment and reach metrics they report, which is precisely why the listening literature insists on model-based, selection-corrected measurement (Schweidel and Moe 2014) rather than dashboard counts taken at face value. For the researcher, an SLP is a data source whose provenance must be documented with the same rigor demanded of any scraped or API-collected dataset (Chapter 12).

12.15 Worked Example: Sentiment as a Leading Indicator

To make the listening argument concrete, the following reproducible example illustrates why venue- or selection-aware sentiment can lead an external performance series while a naïve average lags or misleads. We simulate latent brand health, generate posts whose probability of appearing depends on sentiment (polarity self-selection), and compare a naïve mean rating against a selection-corrected measure as predictors of next-period sales. The example is pedagogical, not an estimator for production use, but it reproduces the qualitative pattern documented by Schweidel and Moe (2014) and Schoenmueller, Netzer, and Stahl (2020).

set.seed(42)

n_weeks <- 60
# Latent brand health follows a smooth AR(1)-like path.
health <- as.numeric(filter(rnorm(n_weeks, 0, 0.3), 0.8, method = "recursive"))
health <- scale(health)[, 1]

# Next-week sales depend on this week's latent health plus noise.
sales <- 100 + 15 * dplyr::lag(health, default = 0) + rnorm(n_weeks, 0, 5)

# Each week, many latent opinions are drawn around health, but the probability
# of POSTING rises with the extremity of the opinion (polarity self-selection).
naive_sentiment      <- numeric(n_weeks)
corrected_sentiment  <- numeric(n_weeks)
for (t in seq_len(n_weeks)) {
  opinions  <- rnorm(400, mean = health[t], sd = 1)
  post_prob <- plogis(2 * abs(opinions) - 1)          # extreme -> more likely to post
  posted    <- opinions[runif(400) < post_prob]
  naive_sentiment[t] <- mean(posted)                  # biased by self-selection
  # Selection correction: reweight posted opinions by inverse posting propensity.
  w <- 1 / plogis(2 * abs(posted) - 1)
  corrected_sentiment[t] <- weighted.mean(posted, w)
}

# Which sentiment series better predicts NEXT week's sales?
df <- data.frame(
  sales_next        = dplyr::lead(sales),
  naive             = naive_sentiment,
  corrected         = corrected_sentiment
)
df <- df[complete.cases(df), ]

cor_naive     <- cor(df$naive,     df$sales_next)
cor_corrected <- cor(df$corrected, df$sales_next)

cat("Corr(naive sentiment, next-week sales):     ", round(cor_naive, 3), "\n")
#> Corr(naive sentiment, next-week sales):      0.952
cat("Corr(corrected sentiment, next-week sales): ", round(cor_corrected, 3), "\n")
#> Corr(corrected sentiment, next-week sales):  0.955

The selection-corrected series tracks latent health—and therefore next-period sales—more closely than the naïve mean, because it undoes the over-representation of extreme opinions. The general lesson generalizes far beyond this toy: online metrics are generated by a selection process (who posts, what the algorithm shows, which venue is chosen), and credible inference requires modeling that process rather than averaging its output.

12.16 Key Takeaways

• Online environments are simultaneously a substantive domain and a measurement opportunity; every empirical claim inherits the validity threats of how its data were collected. Algorithmic interference (Xu, Zhang, and Zhou 2019) and inter-temporal instability are endemic, and documentation in the spirit of Gebru et al. (2021) is the price of reproducibility.
• The field’s organizing functions—expression, decision support, and market intelligence—were stable across two decades even as platforms changed (Lamberton and Stephen 2016); new data types (clickstream, text, images) drove new methods.
• Source governs content effects: UGC dominates awareness and satisfaction while FGC vividness aids consideration and intent (Colicev, Kumar, and O’Connor 2019), and the owned/earned distinction (Colicev et al. 2018) traces social media activity to firm value through a mediated mindset path.
• Reviews are the most tractable online data type, but their integrity is fragile— herding (Sunder, Kim, and Yorkston 2019), polarity self-selection (Schoenmueller, Netzer, and Stahl 2020), and fake reviews (He, Hollenbeck, and Proserpio 2022) all bias the naïve average, so sentiment must be measured, not merely counted (Schweidel and Moe 2014).
• Clickstream and field experiments make the path to purchase observable and actionable in real time (Montgomery et al. 2004; Ding, Li, and Chatterjee 2015), and search design is now a field-experimental science with measurable, sometimes counterintuitive, welfare consequences (Moshary 2025; Lei, Chen, and Sen 2024).
• Emerging interfaces—livestreaming, AR, and AI—each introduce a distinct mechanism (emotional contagion, immersion, perceived intention) rather than merely a new screen, and crowd-based production succeeds through network structure and social dynamics, not single-firm effort (S. L. Johnson, Faraj, and Kudaravalli 2014; Herd, Mallapragada, and Narayan 2022).

Appel, Gil, Lauren Grewal, Rhonda Hadi, and Andrew T Stephen. 2020. “The Future of Social Media in Marketing.” Journal of the Academy of Marketing Science 48 (1): 79–95.

Banerjee, Shrabastee, Chrysanthos Dellarocas, and Georgios Zervas. 2021. “Interacting User-Generated Content Technologies: How Questions and Answers Affect Consumer Reviews.” Journal of Marketing Research 58 (4): 742–61. https://doi.org/10.1177/00222437211020274.

Bart, Yakov, Andrew T. Stephen, and Miklos Sarvary. 2014. “Which Products Are Best Suited to Mobile Advertising? A Field Study of Mobile Display Advertising Effects on Consumer Attitudes and Intentions.” Journal of Marketing Research 51 (3): 270–85. https://doi.org/10.1509/jmr.13.0503.

Beck, Jonathan M, Clay M Voorhees, Paul W Fombelle, and Katherine N Lemon. n.d. “EXPRESS: Putting Words in Their Mouths: Firm-Generated User Content and Consumer Sharing Behavior.” Journal of Marketing, 00222429251331483.

Bharadwaj, Neeraj, Michel Ballings, Prasad A. Naik, Miller Moore, and Mustafa Murat Arat. 2021. “A New Livestream Retail Analytics Framework to Assess the Sales Impact of Emotional Displays.” Journal of Marketing, September, 002224292110130. https://doi.org/10.1177/00222429211013042.

Bing Zhou, Shuang Xu, and Xin-xin Yang. 2015. “Computing the Color Complexity of Images.” 2015 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD), August. https://doi.org/10.1109/fskd.2015.7382237.

Boegershausen, Johannes, Hannes Datta, Abhishek Borah, and Andrew T. Stephen. 2022. “EXPRESS: Fields of Gold: Scraping Web Data for Marketing Insights.” Journal of Marketing, May, 002224292211007. https://doi.org/10.1177/00222429221100750.

Boyd, D Eric, PK Kannan, and Rebecca J Slotegraaf. 2019. “Branded Apps and Their Impact on Firm Value: A Design Perspective.” Journal of Marketing Research 56 (1): 76–88.

Brynjolfsson, Erik, Yu (Jeffrey) Hu, and Michael D. Smith. 2003. “Consumer Surplus in the Digital Economy: Estimating the Value of Increased Product Variety at Online Booksellers.” Management Science 49 (11): 1580–96. https://doi.org/10.1287/mnsc.49.11.1580.20580.

Bucklin, Randolph E., and Catarina Sismeiro. 2003. “A Model of Web Site Browsing Behavior Estimated on Clickstream Data.” Journal of Marketing Research 40 (3): 249–67. https://doi.org/10.1509/jmkr.40.3.249.19241.

Buratti, Nicoletta, Francesco Parola, and Giovanni Satta. 2018. “Insights on the Adoption of Social Media Marketing in B2B Services.” The TQM Journal 30 (5): 490–529. https://doi.org/10.1108/tqm-11-2017-0136.

Burke, Raymond R. 2002. “Technology and the Customer Interface: What Consumers Want in the Physical and Virtual Store.” Journal of the Academy of Marketing Science 30 (4): 411–32.

Büschken, Joachim, and Greg M Allenby. 2016. “Sentence-Based Text Analysis for Customer Reviews.” Marketing Science 35 (6): 953–75.

Cao, Jingcun, Pradeep Chintagunta, and Shibo Li. 2023. “From Free to Paid: Monetizing a Non-Advertising-Based App.” Journal of Marketing Research, 00222437221131562.

Chen, Jinjie, Alison Jing Xu, Maria A. Rodas, and Xuefeng Liu. 2022. “EXPRESS: Order Matters: Rating Service Professionals Reduces Tipping Amount.” Journal of Marketing, April, 002224292210986. https://doi.org/10.1177/00222429221098698.

Colicev, Anatoli, Ashish Kumar, and Peter O’Connor. 2019. “Modeling the Relationship Between Firm and User Generated Content and the Stages of the Marketing Funnel.” International Journal of Research in Marketing 36 (1): 100–116. https://doi.org/10.1016/j.ijresmar.2018.09.005.

Colicev, Anatoli, Ashwin Malshe, Koen Pauwels, and Peter O’Connor. 2018. “Improving Consumer Mindset Metrics and Shareholder Value Through Social Media: The Different Roles of Owned and Earned Media.” Journal of Marketing 82 (1): 37–56. https://doi.org/10.1509/jm.16.0055.

Costello, John P., and Rebecca Walker Reczek. 2020. “Providers Versus Platforms: Marketing Communications in the Sharing Economy.” Journal of Marketing 84 (6): 22–38. https://doi.org/10.1177/0022242920925038.

Dai, Hengchen, Cindy Chan, and Cassie Mogilner. 2019. “People Rely Less on Consumer Reviews for Experiential Than Material Purchases.” Edited by Darren W Dahl, Margaret C Campbell, and Cait Lamberton. Journal of Consumer Research 46 (6): 1052–75. https://doi.org/10.1093/jcr/ucz042.

Dekimpe, Marnik G., and Dominique M. Hanssens. 1995. “The Persistence of Marketing Effects on Sales.” Marketing Science 14 (1): 1–21. https://doi.org/10.1287/mksc.14.1.1.

Dholakia, Utpal M, and Lopo L Rego. 1998. “What Makes Commercial Web Pages Popular? An Empirical Investigation of Web Page Effectiveness.” European Journal of Marketing 32 (7/8): 724–36.

Ding, Amy Wenxuan, and Shibo Li. 2019. “Herding in the Consumption and Purchase of Digital Goods and Moderators of the Herding Bias.” Journal of the Academy of Marketing Science 47: 460–78.

Ding, Amy Wenxuan, Shibo Li, and Patrali Chatterjee. 2015. “Learning User Real-Time Intent for Optimal Dynamic Web Page Transformation.” Information Systems Research 26 (2): 339–59.

Ershov, Daniel, Yanting He, and Stephan Seiler. 2024. “How Much Influencer Marketing Is Undisclosed? Evidence from Twitter.” Evidence from Twitter (April 22, 2024).

Fajtl, Jiri, Vasileios Argyriou, Dorothy Monekosso, and Paolo Remagnino. 2018. “AMNet: Memorability Estimation with Attention.” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, June. https://doi.org/10.1109/cvpr.2018.00666.

Finkelstein, Stacey R., and Ayelet Fishbach. 2012. “Tell Me What I Did Wrong: Experts Seek and Respond to Negative Feedback.” Journal of Consumer Research 39 (1): 22–38. https://doi.org/10.1086/661934.

Folse, Judith Anne Garretson, McDowell Porter III, Mousumi Bose Godbole, and Kristy E. Reynolds. 2016. “The Effects of Negatively Valenced Emotional Expressions in Online Reviews on the Reviewer, the Review, and the Product.” Psychology and Marketing 33 (9): 747–60. https://doi.org/10.1002/mar.20914.

Forsythe, Alex, Noel Sheehy, and Martin Sawey. 2003. “Measuring Icon Complexity: An Automated Analysis.” Behavior Research Methods, Instruments and Computers 35 (2): 334–42. https://doi.org/10.3758/bf03202562.

Garvey, Aaron M., TaeWoo Kim, and Adam Duhachek. 2021. “EXPRESS: Bad News? Send an AI. Good News? Send a Human.” Journal of Marketing, December, 002224292110669. https://doi.org/10.1177/00222429211066972.

Gebru, Timnit, Jamie Morgenstern, Briana Vecchione, Jennifer Wortman Vaughan, Hanna Wallach, Hal Daumé III, and Kate Crawford. 2021. “Datasheets for Datasets.” Communications of the ACM 64 (12): 86–92. https://doi.org/10.1145/3458723.

Ghose, Anindya, and Sang Pil Han. 2011. “An Empirical Analysis of User Content Generation and Usage Behavior on the Mobile Internet.” Management Science 57 (9): 1671–91. https://doi.org/10.1287/mnsc.1110.1350.

Golmohammadi, Alireza, Taha Havakhor, Dinesh K. Gauri, and Joseph Comprix. 2021. “Complaint Publicization in Social Media.” Journal of Marketing 85 (6): 1–23. https://doi.org/10.1177/00222429211002183.

Grewal, Rajdeep, Sachin Gupta, and Rebecca Hamilton. 2021. “Marketing Insights from Multimedia Data: Text, Image, Audio, and Video.” Journal of Marketing Research 58 (6): 1025–33. https://doi.org/10.1177/00222437211054601.

Grewal, Rajdeep, Gary L Lilien, and Girish Mallapragada. 2006. “Location, Location, Location: How Network Embeddedness Affects Project Success in Open Source Systems.” Management Science 52 (7): 1043–56.

Guan, Zhiwei, and Edward Cutrell. 2007. “An Eye Tracking Study of the Effect of Target Rank on Web Search.” Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, April. https://doi.org/10.1145/1240624.1240691.

Harmeling, Colleen M., Jordan W. Moffett, Mark J. Arnold, and Brad D. Carlson. 2016. “Toward a Theory of Customer Engagement Marketing.” Journal of the Academy of Marketing Science 45 (3): 312–35. https://doi.org/10.1007/s11747-016-0509-2.

Häubl, Gerald, and Valerie Trifts. 2000. “Consumer Decision Making in Online Shopping Environments: The Effects of Interactive Decision Aids.” Marketing Science 19 (1): 4–21. https://doi.org/10.1287/mksc.19.1.4.15178.

He, Sherry, Brett Hollenbeck, and Davide Proserpio. 2022. “The Market for Fake Reviews.” Marketing Science, February. https://doi.org/10.1287/mksc.2022.1353.

Herd, Kelly B, Girish Mallapragada, and Vishal Narayan. 2022. “Do Backer Affiliations Help or Hurt Crowdfunding Success?” Journal of Marketing 86 (5): 117–34.

Herhausen, Dennis, Lauren Grewal, Krista Hill Cummings, Anne L. Roggeveen, Francisco Villarroel Ordenes, and Dhruv Grewal. 2022. “EXPRESS: Complaint Deescalation Strategies on Social Media.” Journal of Marketing, August, 002224292211199. https://doi.org/10.1177/00222429221119977.

Hewett, Kelly, William Rand, Roland T. Rust, and Harald J. van Heerde. 2016. “Brand Buzz in the Echoverse.” Journal of Marketing 80 (3): 1–24. https://doi.org/10.1509/jm.15.0033.

Hoskins, Jake, Shyam Gopinath, J Cameron Verhaal, and Elham Yazdani. 2021. “The Influence of the Online Community, Professional Critics, and Location Similarity on Review Ratings for Niche and Mainstream Brands.” Journal of the Academy of Marketing Science 49: 1065–87.

Huang, Shan, and Song Lin. 2024. “Do More" Likes" Lead to More Clicks? Evidence from a Field Experiment on Social Advertising.” Journal of Marketing.

Jalali, Nima Y., and Purushottam Papatla. 2016. “The Palette That Stands Out: Color Compositions of Online Curated Visual UGC That Attracts Higher Consumer Interaction.” Quantitative Marketing and Economics 14 (4): 353–84. https://doi.org/10.1007/s11129-016-9178-1.

Jia, He (Michael), B. Kyu Kim, and Lin Ge. 2020. “Speed Up, Size Down: How Animated Movement Speed in Product Videos Influences Size Assessment and Product Evaluation.” Journal of Marketing 84 (5): 100–116. https://doi.org/10.1177/0022242920925054.

Johnson, Eric J, Wendy W Moe, Peter S Fader, Steven Bellman, and Gerald L Lohse. 2004. “On the Depth and Dynamics of Online Search Behavior.” Management Science 50 (3): 299–308.

Johnson, Steven L, Samer Faraj, and Srinivas Kudaravalli. 2014. “Emergence of Power Laws in Online Communities.” Mis Quarterly 38 (3): 795–A13.

Kannan, P. K., and Hongshuang “Alice” Li. 2017. “Digital Marketing: A Framework, Review and Research Agenda.” International Journal of Research in Marketing 34 (1): 22–45. https://doi.org/10.1016/j.ijresmar.2016.11.006.

Kärkkäinen, Hannu, Jari Jussila, and Jaani Väisänen. 2010. “Social Media Use and Potential in Business-to-Business Companies’ Innovation.” In Proceedings of the 14th International Academic MindTrek Conference: Envisioning Future Media Environments, 228–36. MindTrek ’10. New York, NY, USA: Association for Computing Machinery. https://doi.org/10.1145/1930488.1930536.

Katona, Zsolt, and Miklos Sarvary. 2008. “Network Formation and the Structure of the Commercial World Wide Web.” Marketing Science 27 (5): 764–78. https://doi.org/10.1287/mksc.1070.0349.

———. 2010. “The Race for Sponsored Links: Bidding Patterns for Search Advertising.” Marketing Science 29 (2): 199–215. https://doi.org/10.1287/mksc.1090.0517.

Khosla, Aditya, Akhil S. Raju, Antonio Torralba, and Aude Oliva. 2015. “Understanding and Predicting Image Memorability at a Large Scale.” 2015 IEEE International Conference on Computer Vision (ICCV), December. https://doi.org/10.1109/iccv.2015.275.

Khosla, Aditya, Jianxiong Xiao, Phillip Isola, Antonio Torralba, and Aude Oliva. 2012. “Image Memorability and Visual Inception.” SIGGRAPH Asia 2012 Technical Briefs on - SA ’12. https://doi.org/10.1145/2407746.2407781.

Kietzmann, Jan H, Kristopher Hermkens, Ian P McCarthy, and Bruno S Silvestre. 2011. “Social Media? Get Serious! Understanding the Functional Building Blocks of Social Media.” Business Horizons 54 (3): 241–51.

KOUKOVA, N, P KANNAN, and B RATCHFORD. 2008. “Product Form Bundling: Implications for Marketing Digital Products.” Journal of Retailing 84 (2): 181–94. https://doi.org/10.1016/j.jretai.2008.04.001.

Kulkarni, Gauri, PK Kannan, and Wendy Moe. 2012. “Using Online Search Data to Forecast New Product Sales.” Decision Support Systems 52 (3): 604–11.

Kumar, V., Lerzan Aksoy, Bas Donkers, Rajkumar Venkatesan, Thorsten Wiesel, and Sebastian Tillmanns. 2010. “Undervalued or Overvalued Customers: Capturing Total Customer Engagement Value.” Journal of Service Research 13 (3): 297–310. https://doi.org/10.1177/1094670510375602.

Lacka, Ewelina, D. Eric Boyd, Gbenga Ibikunle, and P.K. Kannan. 2021. “EXPRESS: Measuring the Real-Time Stock Market Impact of Firm-Generated Content.” Journal of Marketing, August, 002224292110428. https://doi.org/10.1177/00222429211042848.

Lamberton, Cait, and Andrew T. Stephen. 2016. “A Thematic Exploration of Digital, Social Media, and Mobile Marketing: Research Evolution from 2000 to 2015 and an Agenda for Future Inquiry.” Journal of Marketing 80 (6): 146–72. https://doi.org/10.1509/jm.15.0415.

Lei, Xiaoxia, Yixing Chen, and Ananya Sen. 2024. “The Value of External Data Capabilities in Search Markets: Evidence from a Field Experiment.” Available at SSRN 4452804.

Leyva, Roberto, and Victor Sanchez. 2021. “Video Memorability Prediction via Late Fusion of Deep Multi-Modal Features.” In 2021 IEEE International Conference on Image Processing (ICIP), 2488–92. IEEE.

Li, Linyi, Shyam Gopinath, and Stephen J Carson. 2022. “History Matters: The Impact of Online Customer Reviews Across Product Generations.” Management Science 68 (5): 3878–3903.

Li, Nan, Avery Haviv, and Mitchell J Lovett. 2024. “Opposing Influences of YouTube Influencers: Purchase and Usage Effects in the Video Game Industry.” Preprint, Submitted March 8.

Li, Shibo, Kannan Srinivasan, and Baohong Sun. 2009. “Internet Auction Features as Quality Signals.” Journal of Marketing 73 (1): 75–92.

Li, Shibo, Baohong Sun, and Alan L Montgomery. 2011. “Cross-Selling the Right Product to the Right Customer at the Right Time.” Journal of Marketing Research 48 (4): 683–700.

Li, Shibo, Baohong Sun, and Ronald T Wilcox. 2005. “Cross-Selling Sequentially Ordered Products: An Application to Consumer Banking Services.” Journal of Marketing Research 42 (2): 233–39.

Li, Yiyi, and Ying Xie. 2019. “Is a Picture Worth a Thousand Words? An Empirical Study of Image Content and Social Media Engagement.” Journal of Marketing Research 57 (1): 1–19. https://doi.org/10.1177/0022243719881113.

Lin, Yan, Dai Yao, and Xingyu Chen. 2021. “Happiness Begets Money: Emotion and Engagement in Live Streaming.” Journal of Marketing Research 58 (3): 417–38. https://doi.org/10.1177/00222437211002477.

Lysyakov, Mikhail, PK Kannan, Siva Viswanathan, and Kunpeng Zhang. 2024. “Retailer Differentiation in Social Media: An Investigation of Firm-Generated Content on Twitter.” Journal of Marketing, 00222429241298654.

Mahr, Dominik, Aric Rindfleisch, and Rebecca J Slotegraaf. 2015. “Enhancing Crowdsourcing Success: The Role of Creative and Deliberate Problem-Solving Styles.” Customer Needs and Solutions 2: 209–21.

Mallapragada, Girish, Sandeep R Chandukala, and Qing Liu. 2016. “Exploring the Effects of ‘What’(product) and ‘Where’(website) Characteristics on Online Shopping Behavior.” Journal of Marketing 80 (2): 21–38.

Mallapragada, Girish, Rajdeep Grewal, and Gary Lilien. 2012. “User-Generated Open Source Products: Founder’s Social Capital and Time to Product Release.” Marketing Science 31 (3): 474–92.

Melumad, Shiri, and Michel Tuan Pham. 2020. “The Smartphone as a Pacifying Technology.” Edited by Darren W Dahl, Amna Kirmani, and Peter R Darke. Journal of Consumer Research 47 (2): 237–55. https://doi.org/10.1093/jcr/ucaa005.

Miao, Fred, Irina V. Kozlenkova, Haizhong Wang, Tao Xie, and Robert W. Palmatier. 2021. “An Emerging Theory of Avatar Marketing.” Journal of Marketing, May, 002224292199664. https://doi.org/10.1177/0022242921996646.

Montgomery, Alan L, Shibo Li, Kannan Srinivasan, and John C Liechty. 2004. “Modeling Online Browsing and Path Analysis Using Clickstream Data.” Marketing Science 23 (4): 579–95.

Moshary, Sarah. 2025. “Does Sponsored Search Advertising Augment Organic Search? Evidence from an e-Commerce Platform.” Management Science.

Munz, Kurt P., Minah H. Jung, and Adam L. Alter. 2020. “Name Similarity Encourages Generosity: A Field Experiment in Email Personalization.” Marketing Science 39 (6): 1071–91. https://doi.org/10.1287/mksc.2019.1220.

Nguyen, Peter, Xin (Shane) Wang, Xi Li, and June Cotte. 2020. “Reviewing Experts’ Restraint from Extremes and Its Impact on Service Providers.” Edited by J Jeffrey Inman and Andrew T Stephen. Journal of Consumer Research 47 (5): 654–74. https://doi.org/10.1093/jcr/ucaa037.

Nishijima, Marislei, Mauro Rodrigues, and Thaís Luiza Donega Souza. 2021. “Is Rotten Tomatoes Killing the Movie Industry? A Regression Discontinuity Approach.” Applied Economics Letters, April, 1–6. https://doi.org/10.1080/13504851.2021.1918324.

Ordabayeva, Nailya, Lisa A. Cavanaugh, and Darren W. Dahl. 2022. “EXPRESS: The Upside of Negative: Social Distance in Online Reviews of Identity-Relevant Brands.” Journal of Marketing, January, 002224292210747. https://doi.org/10.1177/00222429221074704.

Park, Sungsik, Woochoel Shin, and Jinhong Xie. 2021. “The Fateful First Consumer Review.” Marketing Science 40 (3): 481–507. https://doi.org/10.1287/mksc.2020.1264.

Pelletier, Mark J., Alexandra Krallman, Frank G. Adams, and Tyler Hancock. 2020. “One Size Doesn’t Fit All: A Uses and Gratifications Analysis of Social Media Platforms.” Journal of Research in Interactive Marketing 14 (2): 269–84. https://doi.org/10.1108/jrim-10-2019-0159.

Peters, Kay, Yubo Chen, Andreas M Kaplan, Björn Ognibeni, and Koen Pauwels. 2013. “Social Media Metrics—a Framework and Guidelines for Managing Social Media.” Journal of Interactive Marketing 27 (4): 281–98.

Pieters, Rik, Michel Wedel, and Rajeev Batra. 2010. “The Stopping Power of Advertising: Measures and Effects of Visual Complexity.” Journal of Marketing 74 (5): 48–60. https://doi.org/10.1509/jmkg.74.5.48.

Pieters, Rik, Michel Wedel, and Jie Zhang. 2007. “Optimal Feature Advertising Design Under Competitive Clutter.” Management Science 53 (11): 1815–28. https://doi.org/10.1287/mnsc.1070.0732.

Qin, Marco, Xueming Luo, Todd Schifeling, and Yang Wang. 2024. “When Corporate Silence Is Costly: Negative Consumer Responses to Corporate Silence on Social Issues.” Available at SSRN 5014706.

Rangone, Andrea, and Filippo Renga. 2006. “Mobile Advertising: A Framework for the Appraisal of the Campaigns.” 2006 International Conference on Mobile Business, June. https://doi.org/10.1109/icmb.2006.323477.

Reimer, Kerstin, Oliver J. Rutz, and Koen Pauwels. 2014. “How Online Consumer Segments Differ in Long-Term Marketing Effectiveness.” Journal of Interactive Marketing 28 (4): 271–84. https://doi.org/10.1016/j.intmar.2014.05.002.

Reinecke, Katharina, Tom Yeh, Luke Miratrix, Rahmatri Mardiko, Yuechen Zhao, Jenny Liu, and Krzysztof Z. Gajos. 2013. “Predicting Users’ First Impressions of Website Aesthetics with a Quantification of Perceived Visual Complexity and Colorfulness.” Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, April. https://doi.org/10.1145/2470654.2481281.

Rohm, Andrew J, and Vanitha Swaminathan. 2004. “A Typology of Online Shoppers Based on Shopping Motivations.” Journal of Business Research 57 (7): 748–57. https://doi.org/10.1016/s0148-2963(02)00351-x.

Sahni, Navdeep S., S. Christian Wheeler, and Pradeep Chintagunta. 2018. “Personalization in Email Marketing: The Role of Noninformative Advertising Content.” Marketing Science 37 (2): 236–58. https://doi.org/10.1287/mksc.2017.1066.

Saljoughian, Mohammad “Mike”, Kelly Hewett, Harald J van Heerde, and William “Bill” Rand. n.d. “EXPRESS: How Firms Can Steer Social Media Conversations.” Journal of Marketing Research, 00222437251329816.

Salo, Jari. 2017. “Social Media Research in the Industrial Marketing Field: Review of Literature and Future Research Directions.” Industrial Marketing Management 66 (October): 115–29. https://doi.org/10.1016/j.indmarman.2017.07.013.

Schoenmueller, Verena, Oded Netzer, and Florian Stahl. 2020. “The Polarity of Online Reviews: Prevalence, Drivers and Implications.” Journal of Marketing Research 57 (5): 853–77. https://doi.org/10.1177/0022243720941832.

Schweidel, David A, and Wendy W Moe. 2014. “Listening in on Social Media: A Joint Model of Sentiment and Venue Format Choice.” Journal of Marketing Research 51 (4): 387–402.

Sen, Shahana, and Dawn Lerman. 2007. “Why Are You Telling Me This? An Examination into Negative Consumer Reviews on the Web.” Journal of Interactive Marketing 21 (4): 76–94. https://doi.org/10.1002/dir.20090.

Severin, Werner Joseph, James W Tankard, et al. 1997. Communication Theories: Origins, Methods, and Uses in the Mass Media. Longman New York.

Sharma, Siddharth, Justin Frake, and Jared Watson. 2025. “Symbolic Vs. Substantive Support: The Impact of Black Lives Matter on Black-Owned Businesses.” Marketing Science.

Sridhar, Shrihari, and Eric Fang. 2019. “New Vistas for Marketing Strategy: Digital, Data-Rich, and Developing Market (D3) Environments.” Journal of the Academy of Marketing Science 47 (6): 977–85. https://doi.org/10.1007/s11747-019-00698-y.

Srinivasan, Shuba, Oliver J. Rutz, and Koen Pauwels. 2015. “Paths to and Off Purchase: Quantifying the Impact of Traditional Marketing and Online Consumer Activity.” Journal of the Academy of Marketing Science 44 (4): 440–53. https://doi.org/10.1007/s11747-015-0431-z.

Steuer, Jonathan. 1992. “Defining Virtual Reality: Dimensions Determining Telepresence.” Journal of Communication 42 (4): 73–93. https://doi.org/10.1111/j.1460-2466.1992.tb00812.x.

Sunder, Sarang, Kihyun Hannah Kim, and Eric A Yorkston. 2019. “What Drives Herding Behavior in Online Ratings? The Role of Rater Experience, Product Portfolio, and Diverging Opinions.” Journal of Marketing 83 (6): 93–112.

Tonietto, Gabriela N., and Alixandra Barasch. 2020. “Generating Content Increases Enjoyment by Immersing Consumers and Accelerating Perceived Time.” Journal of Marketing 85 (6): 83–100. https://doi.org/10.1177/0022242920944388.

Troncoso, Isamar, and Lan Luo. 2022. “Look the Part? The Role of Profile Pictures in Online Labor Markets.” Marketing Science.

Trusov, Michael, Anand V. Bodapati, and Randolph E. Bucklin. 2010. “Determining Influential Users in Internet Social Networks.” Journal of Marketing Research 47 (4): 643–58. https://doi.org/10.1509/jmkr.47.4.643.

Trusov, Michael, Randolph E. Bucklin, and Koen Pauwels. 2009a. “Effects of Word-of-Mouth Versus Traditional Marketing: Findings from an Internet Social Networking Site.” Journal of Marketing 73 (5): 90–102. https://doi.org/10.1509/jmkg.73.5.90.

Trusov, Michael, Randolph E Bucklin, and Koen Pauwels. 2009b. “Effects of Word-of-Mouth Versus Traditional Marketing: Findings from an Internet Social Networking Site.” Journal of Marketing 73 (5): 90–102.

Verhoef, Peter C., and Tammo H. A. Bijmolt. 2019. “Marketing Perspectives on Digital Business Models: A Framework and Overview of the Special Issue.” International Journal of Research in Marketing 36 (3): 341–49. https://doi.org/10.1016/j.ijresmar.2019.08.001.

Villanueva, Julian, Shijin Yoo, and Dominique M Hanssens. 2008. “The Impact of Marketing-Induced Versus Word-of-Mouth Customer Acquisition on Customer Equity Growth.” Journal of Marketing Research 45 (1): 48–59. https://doi.org/10.1509/jmkr.45.1.48.

Wang, Shuting, and Brad N Greenwood. 2025. “Too Long, Didn’t Read? The Impact of Length Limits on Microblogging Engagement.” Management Science.

Wang, Xin (Shane), Jiaxiu He, David J. Curry, and Jun Hyun (Joseph) Ryoo. 2021. “Attribute Embedding: Learning Hierarchical Representations of Product Attributes from Consumer Reviews.” Journal of Marketing, November, 002224292110478. https://doi.org/10.1177/00222429211047822.

Wedel, Michel, and P.K. Kannan. 2016. “Marketing Analytics for Data-Rich Environments.” Journal of Marketing 80 (6): 97–121. https://doi.org/10.1509/jm.15.0413.

Xu, Heng, Nan Zhang, and Le Zhou. 2019. “Validity Concerns in Research Using Organic Data.” Journal of Management 46 (7): 1257–74. https://doi.org/10.1177/0149206319862027.

Yadav, Manjit S., and Paul A. Pavlou. 2014. “Marketing in Computer-Mediated Environments: Research Synthesis and New Directions.” Journal of Marketing 78 (1): 20–40. https://doi.org/10.1509/jm.12.0020.

Yang, Jeremy, Juanjuan Zhang, and Yuhan Zhang. 2025. “Engagement That Sells: Influencer Video Advertising on TikTok.” Marketing Science 44 (2): 247–67. https://doi.org/10.1287/mksc.2021.0107.

Yazdani, Elham, Shyam Gopinath, and Steve Carson. 2018. “Preaching to the Choir: The Chasm Between Top-Ranked Reviewers, Mainstream Customers, and Product Sales.” Marketing Science 37 (5): 838–51.

Yim, Mark Yi-Cheon, Shu-Chuan Chu, and Paul L. Sauer. 2017. “Is Augmented Reality Technology an Effective Tool for E-Commerce? An Interactivity and Vividness Perspective.” Journal of Interactive Marketing 39 (August): 89–103. https://doi.org/10.1016/j.intmar.2017.04.001.

Zhang, Shunyuan, Dokyun Lee, Param Vir Singh, and Kannan Srinivasan. 2022. “What Makes a Good Image? Airbnb Demand Analytics Leveraging Interpretable Image Features.” Management Science 68 (8): 5644–66.

Zhang, Xiaoling, Shibo Li, and Raymond R Burke. 2018. “Modeling the Effects of Dynamic Group Influence on Shopper Zone Choice, Purchase Conversion, and Spending.” Journal of the Academy of Marketing Science 46: 1089–1107.

Zhang, Xiaoling, Shibo Li, Raymond R Burke, and Alex Leykin. 2014. “An Examination of Social Influence on Shopper Behavior Using Video Tracking Data.” Journal of Marketing 78 (5): 24–41.

Zhang, Yuchi, Michael Trusov, Andrew T. Stephen, and Zainab Jamal. 2017. “Online Shopping and Social Media: Friends or Foes?” Journal of Marketing 81 (6): 24–41. https://doi.org/10.1509/jm.14.0344.

Zheng, Shuang, Siliang Tong, Hyeokkoo Eric Kwon, Gordon Burtch, and Xianneng Li. 2023. “Recommending What to Search: Sales Volume and Consumption Diversity Effects of a Query Recommender System.” Available at SSRN 4667778.