Knowledge that Transforms

Despite the prevalence of single average ratings on online rating platforms, they mask nuanced insights in online ratings. This paper contributes to the burgeoning literature on this topic by illustrating the importance of critical, yet overlooked, information obscured by single average ratings: trend information in online ratings. Online ratings are not static; they change over time as product/service quality changes and consumer expectations and preferences shift. The trend information embedded in online ratings enables consumers to go beyond past ratings to extrapolate expected ratings in the future. I focus on this aspect of online ratings and investigate (a) the effects of displaying trends in online ratings on consumer purchase decisions, (b) the mechanism mediating these trend effects, and (c) the situational contingency factors moderating these trend effects. Drawing on psychological momentum (PM) theory and four randomized controlled experiments (total n=2,015), this paper shows that the presence and explicit presentation of trends in online ratings create a PM experience that sways consumer decisions. It further demonstrates that the potency of these trend effects depends on the trend’s norm alignment, valence, surface forms of online ratings, and trend continuity. I will discuss the contributions of these findings to the research and practice of online ratings.

Declining circulation and advertising revenue have led many newspapers to cut costs by reducing in-house staff and relying more on wire content. In-house production provides unique content, while wire (non-exclusive) content offers broad coverage at lower cost. Using data from a regional U.S. newspaper, we examine how the mix of in-house and wire content affects subscription decisions. To address reader-side selection bias, we use local precipitation as an excluded variable to indirectly randomize readers’ exposure to content mix. We find that a 1% increase in the share of in-house articles raises daily subscriptions by 0.024% (≈9% of baseline) and by another 0.009% (≈4% of baseline) under a paywall. Against this overall trend, wire Business and Sports articles also increase conversion, suggesting opportunities for selective sourcing. Effects are weaker for local readers and for visitors from social media. Finally, readers from aggregators respond to in-house content and paywalls much like direct visitors, indicating that differentiation and paywalls remain effective even as readers turn to aggregators to discover news.

‘Notice and Choice’ has been a mainstay of policies designed to safeguard consumer privacy. This paper investigates distortions in consumer behavior when faced with notice and choice which may limit the ability of consumers to safeguard their privacy. We use data derived from a field experiment at MIT that distributed a new product, Bitcoin, to all undergraduates. There are two primary findings. First, small navigation costs have a tangible effect on how privacy-protective consumers’ choices are, often in sharp contrast with individual stated preferences about privacy. Second, the introduction of irrelevant, but reassuring, information about privacy protection makes consumers less likely to avoid surveillance, regardless of their stated preferences toward privacy.

Machine learning has long held intuitive appeal in aiding theory building, owing to its capacity to automatically uncover intricate patterns from vast amounts of observed data. Yet, recent literature in organizational research has acknowledged a potential hurdle in utilizing machine learning for theory building: a lack of consistency in the machine-learning output. This means that the same machine learning algorithm could learn predictively accurate yet theoretically contradicting patterns when applied to different samples from the same population, or even when run multiple times on the same sample. This article aims to address the fundamental question of whether, when, and how predictive accuracy implies theoretical pertinence in the context of using machine learning to support inductive theory building. Specifically, we offer theoretical arguments to establish the importance of ensuring that a machine learning algorithm’s assumptions regarding input data are properly aligned with the phenomena being studied. Building on these arguments, we develop a 2×2 framework that outlines the conditions under which four distinct types of machine learning algorithms may be better suited to facilitate inductive theory building in behavioral and organizational research.

The issue of smartphone addiction has been extensively studied, yet there is a lack of exploration into effective solutions to mitigate the impulsivity and temptation induced by these addictive technologies. Despite the proliferation of precommitment-based blocking apps introduced by major smartphone manufacturers, little is known about their effectiveness in curbing compulsive mobile indulgence—particularly among users with varying capacities for self-regulation. Drawing on a rational habit-formation framework, this study investigates how precommitment measures reduce susceptibility to mobile temptation over time, with particular attention to app characteristics, precommitment modes, and individual differences. To empirically validate our framework, we employ a structural model using a multiple discrete-continuous extreme value (MDCEV) approach, which endogenizes individuals’ choices of blocker modes (rigid vs. flexible) and simultaneously examines their app usage behaviors. Our unique dataset—capturing consumer app consumption with and without blocking features active—reveals that consistent use of precommitment apps not only fosters positive habit formation but also significantly reduces compulsive app usage across both hedonic and utilitarian categories. Furthermore, we find that flexible precommitment strategies outperform their rigid counterparts in reducing utilitarian app usage. Individual factors, such as gender and age, are found to play a moderating role in the efficacy of these precommitment strategies. Counterfactual simulations reveal that users who build stronger habit stock through repeated blocker use exhibit significantly lower app usage even after the tool is removed. Scaling the analysis to the population level further shows that expanding access to precommitment devices meaningfully reduces overall smartphone engagement and enhances digital well-being. Based on these empirical findings, we derived implications that can guide policymakers and managers in address excessive mobile dependence in public and workplace environments.

The trade-off between recommending specific versus diverse information to users has long been a challenging issue in recommendation systems. In this study, we probe into a novel problem—mixed-grained recommendation (MGR)—to address this challenge. MGR involves determining the optimal recommendation granularity that aligns with users’ needs for item exploitation and exploration. To this end, we propose a novel deep chain-of-preference learning strategy to infer a user’s choice across mixed-grained categories and items, based on category-aware demand-perception alignment in a top-down manner. Specifically, we design a chain-of-preference-empowered deep learning method (CoPDL) that can infer a user’s (1) dynamic and interrelated mixed-grained demands along a multi-granularity item tree, (2) self-adapted perception along the item tree, and (3) choice regarding mixed-grained nodes in the item tree by virtue of top-down category-aware inference. Empirical evaluation results demonstrate the superior performance of CoPDL over state-of-the-art deep learning alternatives for fine-grained, coarse-grained, and mixed-grained recommendations. Further explanatory investigations render insights into how CoPDL fulfills MGR in effectively balancing the trade-off between recommendation specificity and diversity.

Recently, a controversial new privacy policy on mobile app platforms, which requires app developers to display privacy labels and explicitly request data-tracking permissions from users, has sparked a heated discussion among practitioners in digital advertising. In this paper, we build a game-theoretic model to examine how this new policy impacts the key stakeholders of a mobile app platform (i.e., app developers, the platform, and consumers). The model captures how the new policy prompts developers to expand their strategies beyond pricing by introducing data-tracking levels as an additional competitive dimension, which in turn affects the intensity of price competition. We find that while implementing a policy restricting developers’ tracking of consumers may increase the platform’s advertising revenue, it can also put the platform at a disadvantage. This is because it may incentivize developers to lower their prices and engage in more intense price competition, consequently reducing the platform’s commission revenue. Regarding app developers, the new policy may prove beneficial despite its restriction to tracking only consumers who have opted in. Another noteworthy finding is that while platforms have asserted that the new policy is aimed at safeguarding consumers’ privacy, it does not always serve consumers’ best interests. Our paper provides meaningful implications for all key stakeholders of mobile app platforms regarding the implementation of the new privacy policy.

Online healthcare consultation in virtual health is an emerging industry marked by innovation and fierce competition. Accurate and early prediction of healthcare consultation success can help online platforms proactively address patient concerns and improve retention rates. However, this prediction task is inherently challenging due to several factors: patients’ needs often remain unclear until they explicitly articulate them, and their questions may evolve throughout the consultation process. Additionally, the task involves processing multimodal input information, including consultation dialogues and the complex network of various stakeholders in a patient’s healthcare journey. To address these issues, we propose the Dynamic Knowledge Network and Multimodal Data Fusion framework with a dynamic knowledge graph and multimodal data fusion, which enhances the predictive power of online healthcare consultations. Our work has important implications for new business models where specific and detailed online communication processes are stored in the IT database, and at the same time, latent information with predictive power is embedded in the network formed by stakeholders’ digital traces. It can be extended to diverse industries and domains, where the virtual or hybrid model (e.g., integration of online and offline services) is emerging as a prevailing trend.

Engaging in side projects outside of regular employment has become a growing trend among knowledge workers, particularly information technology (IT) professionals. Side projects offer valuable opportunities for employees to learn new skills and foster creativity. However, the legal ownership of side projects remains uncertain, raising questions about how this affects employee innovation at their primary jobs, which we refer to as “employee innovation at work.” In this study, we leverage an exogenous change in the legal arrangement of side project ownership—the Alcatel v. Brown case—to investigate how firms’ enhanced control over side projects influences employees’ innovation performance at work. We find that in states where firms gained greater contractual authority to claim ownership of employees’ side projects, the number of IT patents owned by firms decreased. However, paradoxically, the quality of these patents, as measured by the number of forward citations, improved following the legal change. Further analyses of the underlying mechanisms suggest that these contrasting findings likely stem from shifts in both employee innovation behaviors and firms’ innovation strategies, post-Alcatel v. Brown. Our findings contribute to the information systems literature by highlighting the nuanced effects of side project ownership on IT innovation.

Recent research has shown that enterprise information technology (IT) can drive strategic growth through mergers and acquisitions (M&As). An implicit assumption underlying this research is that firms own their IT infrastructure. Challenging this assumption, however, the emerging trend of cloud sourcing suggests that IT may be owned by third-party vendors. Since third-party ownership of IT can introduce significant transaction costs and operational inefficiencies, cloud sourcing, unlike in-house enterprise IT, may be considered unlikely to drive M&A growth. However, the unique combination of IT infrastructural and service flexibilities that cloud sourcing provides could help in reducing the risk of integration failure posed by M&As, thereby driving M&A growth. Grounded in the transaction cost economics and resource-based views of the firm, respectively, these arguments illustrate the conflicting theoretical viewpoints offered by prior literature. This study seeks to improve our theoretical understanding of the relationship between cloud sourcing and M&A growth by addressing the theoretical conflict. Analyzing a dataset of cloud sourcing deals and M&As comprising 4,075 observations from 673 firms, our research finds that highly standardized cloud services, i.e., SaaS (software as a service), public, and globalized clouds, have a positive impact on M&As, particularly in information industries. Overall, these results indicate that it is only under conditions of enhanced flexibility afforded by a standardized platform and informationally rich operating environments that cloud sourcing positively affects M&A growth. Support for the theoretical propositions is further established through interviews with industry experts and mechanism tests, which reveal that cloud sourcing has a positive impact, specifically on M&As requiring intensive integration, and is associated with reduced disclosure of M&A risks in annual reports. Finally, consistent with our view that cloud sourcing smooths post-M&A integration, this research also finds that firms with cloud sourcing have relatively stronger post-M&A performance.