Published Papers

Explosive growth in the number of users on various social media platforms has transformed the way firms strategize their marketing activities. To take advantage of the vast size of social networks, firms have now turned their attention to influencer marketing wherein they employ independent influencers to promote their products on social media platforms. Despite the recent growth in influencer marketing, the problem of network seeding (i.e., identification of influencers to optimally post a firm's message or advertisement) neither has been rigorously studied in the academic literature nor has been carefully addressed in practice. We develop a data-driven optimization framework to help a firm successfully conduct (i) short-horizon and (ii) long-horizon influencer marketing campaigns, for which two models are developed, respectively, to maximize the firm's benefit. The models are based on the interactions with marketers, observation of firms' message placements on social media, and model parameters estimated via empirical analysis performed on data from Twitter. Our empirical analysis discovers the effects of collective influence of multiple influencers and finds two important parameters to be included in the models, namely, multiple exposure effect and forgetting effect. For the short-horizon campaign, we develop an optimization model to select influencers and present structural properties for the model. Using these properties, we develop a mathematical programming based polynomial time procedure to provide near-optimal solutions. For the long-horizon problem, we develop an efficient solution procedure to simultaneously select influencers and schedule their message postings over a planning horizon. We demonstrate the superiority of our solution strategies for both short- and long-horizon problems against multiple benchmark methods used in practice. Finally, we present several managerially relevant insights for firms in the influencer marketing context.

Modern communication technologies enable efficient exchange of information but often sacrifice direct human interaction inherent in more traditional forms of communication. This raises the question of whether the lack of personal interaction induces individuals to exploit informational asymmetries. We conducted two experiments with a total of 848 subjects to examine how human versus machine interaction influences cheating for financial gain. We find that individuals cheat about three times more when they interact with a machine rather than a person, regardless of whether the machine is equipped with human features. When interacting with a human, individuals are particularly reluctant to report unlikely and therefore, suspicious outcomes, which is consistent with social image concerns. The second experiment shows that dishonest individuals prefer to interact with a machine when facing an opportunity to cheat. Our results suggest that human presence is key to mitigating dishonest behavior and that self-selection into communication channels can be used to screen for dishonest people.

We study the impacts of social interactions on competing firms' quality differentiation, pricing decisions, and profit performance. Two forms of social interactions are identified and analyzed: (1) market-expansion effect (MEE)-the total market expands as a result of both firms' sales-and (2) value-enhancement effect (VEE)-a consumer gains additional utility of purchasing from one firm based on this firm's previous and/or current sales volume. We consider a two-stage duopoly competition framework, in which both firms select quality levels in the first stage simultaneously and engage in a two-period price competition in the second stage. In the main model, we assume that each firm sets a single price and commits to it across two selling periods. We find that both forms of social interactions tend to lower prices and intensify price competition for given quality levels. However, MEE weakens the product-quality differentiation and is benign to both high-quality and low-quality firms. It also benefits consumers and improves social welfare. By contrast, VEE enlarges the quality differentiation and only benefits the high-quality firm, but is particularly malignant to the low-quality firm. It further reduces the consumers' monetary surplus. Such impact is consistent, regardless of whether the VEE interactions involve previous or current consumers. We further discuss several model extensions, including dynamic pricing, combined social effects, and various cost structures, and verify that the aforementioned impacts of MEE and VEE are qualitatively robust to those extensions. Our results provide important managerial insights for firms in competitive markets and suggest that they need to not only be aware of the consumers' social interactions, but also, more importantly, distinguish the predominant form of the interactions so as to apply proper marketing strategies.

An important challenge for many firms is to identify the life transitions of its customers, such as job searching, expecting a child, or purchasing a home. Inferring such transitions, which are generally unobserved to the firm, can offer the firms opportunities to be more relevant to their customers. In this paper, we demonstrate how a social network platform can leverage its longitudinal user data to identify which of its users are likely to be job seekers. Identifying job seekers is at the heart of the business model of professional social network platforms. Our proposed approach builds on the hidden Markov model (HMM) framework to recover the latent state of job search from noisy signals obtained from social network activity data. Specifically, we use the latent states of the HMM to fuse cross-sectional survey responses to a job-seeking status question with longitudinal user activity data, resulting in a partially HMM. Thus, in some time periods, and for some users, we observe a direct measure of the true job-seeking status. We demonstrate that the proposed model can predict not only which users are likely to be job seeking at any point in time but also what activities on the platform are associated with job search and how long the users have been job seeking. Furthermore, we find that targeting job seekers based on our proposed approach can lead to a 29% increase in profits of a targeting campaign relative to the approach that was used by the social network platform.

Companies often post user-generated reviews online so that potential buyers in different clusters (age, geographic region, occupation, etc.) can learn from existing customers about the quality of an experience good and cluster preferences before purchasing. In this paper, we evaluate two common user-generated review provision policies for selling experience goods to customers in different clusters with heterogeneous preferences. The first policy is called the association-based policy (AP) under which a customer in a cluster can only observe the aggregate review (i.e., average rating) generated by users within the same cluster. The second policy is called the global-based policy (GP) under which each customer is presented with the aggregate review generated by all users across clusters. We find that, in general, the firm benefits from a policy that provides a larger number of "relevant reviews" to customers. When customers are more certain about the product quality and when clusters are more diverse, AP is more profitable than GP because it provides cluster-specific reviews to customers. Otherwise, GP is more profitable as it provides a larger number of less relevant reviews. Moreover, we propose a third provision policy that imparts the union of the information by AP and GP and show that it is more profitable for the firm. Although the third policy always renders a higher consumer welfare than GP, it may generate a lower consumer welfare than AP.

In social trading, less experienced investors (followers) are allowed to copy the trades of experts (traders) in real time after paying a fee. Such a copy-trading mechanism often runs into a transparency-revenue tension. On the one hand, social trading platforms need to release traders' trades as transparently as possible to allow followers to evaluate traders accurately. On the other hand, complete transparency may undercut the platform's revenue because followers can free ride. That is, followers can manually copy the trades of a trader to avoid paying the following fees. This study addresses this critical tension by optimizing the level of transparency through delaying the release of trading information pertaining to the trades executed by traders. We capture the economic impact of the delay using the notions of profit-gap and delayed-profit. We propose a mechanism that elucidates the economic effects of the profit-gap and delayed-profit on followers and, consequently, the amount of money following a trader: protection effect and evaluation effect. Empirical investigations find support for these two effects. We then develop a stochastic control formulation that optimizes platform revenue, where the control is the optimal delay customized at the trader level and calculated as a function of the current amount of money following a trader and the number of views on the trader's profile page. The optimized revenue can be incorporated into an algorithm to provide a systematic way to infuse the platform's goals into the ranking of the traders. A counterfactual study is conducted to demonstrate the performance of the optimal delay policy (versus a constant-delay policy) using data from a leading social trading platform operating in the foreign exchange market.

This paper quantifies the effects of online review platforms on restaurant revenue and consumer welfare. Using a novel data set containing revenues and information from major online review platforms in Texas, I show that online review platforms help consumers learn about restaurant quality more quickly. The effects on learning show up in restaurant revenues. Specifically, doubling the review activity increases the revenue of a high-quality independent restaurant by 5%-19% and decreases that of a low-quality restaurant by a similar amount. These effects vary widely across restaurants' locations. Restaurants around highway exits are affected twice as much as those in nonhighway areas, implying that reviews are more useful to travelers and tourists than locals. The effects also decline as restaurants age, consistent with the diminishing value of information in learning. In contrast, chain restaurants are affected to a much lesser degree than independent restaurants. Building on this evidence, I develop a structural demand model with aggregate social learning. Counterfactual analyses indicate that online review platforms raise consumer welfare much more for tourists than for locals. By encouraging consumers to eat out more often at high-quality independent restaurants, online review platforms increased the total industry revenue by 3.0% over the period from 2011-2015.

We examine how freemium platforms can design social referral programs to encourage growth and engagement without sacrificing revenue. On the one hand, social referral programs generate new referrals from users who would not have paid for the premium features. On the other hand, they also attract new referrals from users who would have paid but prefer to invite others, resulting in more referrals but fewer paying users. We use data from a large-scale randomized field experiment in an online dating platform to assess the effects of adding referrals programs to freemium platforms and changing the referral requirements on users' behavior, namely, on their decisions to invite, pay, and engage with the platform. We find that introducing referral programs in freemium platforms can significantly contribute to increasing the number of referrals at the expense of revenue. Platforms can avoid the loss in revenue by reserving some premium features exclusively for paying users. We also find that increasing referral requirements in social referral programs can work as a double-edged sword. Increasing the referral threshold results in more referrals and higher total revenue. Yet these benefits appear to come at a cost. Users become less engaged, decreasing the value of the platform for all users. We explore two mechanisms that help to explain the differences in users' social engagement. Finally, and contrary to prior findings, we find that the quality of the referrals is not affected by the referral requirements. We discuss the theoretical and practical implications of our research.

In a finitely repeated game with asymmetric information, we experimentally study how individuals adapt the nature of their lies when settings allow for reputation building. Although some lies can be detected ex post by the uninformed party, others remain deniable. We find that traditional market mechanisms, such as reputation, generate strong changes in the way people lie and lead to strategies in which individuals can maintain plausible deniability; people simply hide their lies better by substituting deniable lies for detectable lies. Our results highlight the limitations of reputation to root out fraud when a deniable lie strategy is available.

We experimentally study information transmission by experts motivated by their reputation for being well-informed. In our game of reputational cheap talk, a reporter privately observes information about a state of the world and sends a message to an evaluator; the evaluator uses the message and the realized state of the world to assess the reporter's informativeness. We manipulate the key driver of misreporting incentives: the uncertainty about the phenomenon to forecast. We highlight three findings. First, misreporting information is pervasive even when truthful information transmission can be an equilibrium strategy. Second, consistent with the theory, reporters are more likely to transmit information truthfully when there is more uncertainty on the state. Third, evaluators have difficulty learning reporters' strategies and, contrary to the theory, assessments react more strongly to message accuracy when reporters are more likely to misreport. In a simpler environment with computerized evaluators, reporters learn to best reply to evaluators' behavior and, when the state is highly uncertain and evaluators are credulous, to transmit information truthfully.

Content-sharing social network platforms rely heavily on user-generated content to attract users and advertisers, but they have limited authority over content provision. We develop an intervention that leverages social interactions between users to stimulate content production. We study social nudges, whereby users connected with a content provider on a platform encourage that provider to supply more content. We conducted a randomized field experiment (N=993,676) on a video-sharing social network platform where treatment providers could receive messages from other users encouraging them to produce more, but control providers could not. We find that social nudges not only immediately boosted video supply by 13.21% without changing video quality but also, increased the number of nudges providers sent to others by 15.57%. Such production-boosting and diffusion effects, although declining over time, lasted beyond the day of receiving nudges and were amplified when nudge senders and recipients had stronger ties. We replicate these results in a second experiment. To estimate the overall production boost over the entire network and guide platforms to utilize social nudges, we combine the experimental data with a social network model that captures the diffusion and over-time effects of social nudges. We showcase the importance of considering the network effects when estimating the impact of social nudges and optimizing platform operations regarding social nudges. Our research highlights the value of leveraging co-user influence for platforms and provides guidance for future research to incorporate the diffusion of an intervention into the estimation of its impacts within a social network.

Social media influencers are category enthusiasts who often post product recommendations. Firms sometimes pay influencers to skew their product reviews in favor of the firm. We ask the following research questions. First, what is the optimal level of affiliation (if any) from the firm's perspective? Affiliation introduces positive bias to the influencer's review but also decreases the persuasiveness of the review. Second, because affiliated reviews are often biased in favor of the firm, what is the impact of affiliation on consumer welfare? We find that the affiliation decision depends on the cost of information acquisition, the consumer's prior and awareness, and the disclosure regime. When the consumer's prior belief is low, the firm needs to affiliate less closely or not at all in order to preserve the influencer's persuasiveness, the change in the consumer's belief following the influencer's review. In contrast, when the consumer's prior belief is high, the firm fully affiliates with the influencer to both maximize awareness and prevent a negative review. We also show that the firm's involvement can be Pareto improving if the information acquisition cost is relatively high, and a partial disclosure rule may increase consumer welfare.