Qiang Liu 劉強

MKTG Models · Healthcare · AI & Digital Economy

This research stream examines AI, digital markets, and platform-mediated environments, with a focus on how algorithms and digital platforms shape information, strategic interactions, consumer choice, and market outcomes.

Digital platforms in repeat-purchase categories face a tension between facilitating inertial purchases from the same seller and encouraging variety seeking across sellers. Inertia toward a seller may increase contemporaneous conversion but may elevate the risk of platform disintermediation. Variety seeking can strengthen platform stickiness but may come at the expense of foregoing the conversion benefits of familiar options. It is unclear which tendency better serves the platform and how these tendencies unfold across stages of a purchase episode. Using clickstream data on consumers’ path to purchase from a mobile food-ordering platform, we show the coexistence of these tendencies within a single purchase episode: variety seeking during consideration-set formation and inertia at final choice. We identify this pattern with a two-stage structural model that accommodates heterogeneous consumer preferences and correlated unobservables across stages. Counterfactual analyses indicate that nudging the ranking in the impression set to encourage exploration increases revenue by 4.56%. Moreover, contextualized ranking interventions based on consumers’ category experience, inter-session elapsed time, and platform-wide promotions further amplify these gains, yielding revenue increases of 5.59% to 6.12%. In contrast, a conventional one-stage model detects inertia but misses the consideration-stage variety seeking tendency, biasing policy toward reinforcing inertia and thereby forgoing substantial revenue. These findings contribute to theories of state dependence by conceptualizing inertia and variety seeking as stage-contingent behaviors rather than fixed consumer types and have direct implications for platform ranking design.

Crowdfunding has emerged as a market-based solution to give frontline complex public service employees the opportunity to acquire resources by advertising project proposals for donor patrons on crowdfunding platforms. However, whether crowdfunded resources can improve offline service outcomes, and if so, how and when, remains murky. Focusing on the context of public education crowdfunding and applying theories from crowdfunding and services marketing literature, the authors conceptualize that the combination of two factors—namely, teachers’ request for resources meant to satisfy unmet heterogeneous (i.e., diverse and evolving) intellectual needs of students and donors’ screening and approval (i.e., crowd screening) of promising projects—helps improve student academic achievement. Collating novel panel data from DonorsChoose and California Department of Education, the authors show that (1) crowdfunded resources positively affect student academic achievement, (2) student academic achievement improves with the increase in the heterogeneity of intellectual needs that crowdfunded resources likely satisfy, (3) crowd screening of project proposals plays a critical role in the offline effectiveness of crowdfunded resources, and (4) crowd screening effectiveness depends on the type of project.

We study, using a stylized analytical model, an entrepreneur’s design of the optimal product menu for crowdfunding together with the optimal choice of the funding target. We find that in contrast to prior literature on crowdfunding, downward quality distortion of the low‐quality product in the entrepreneur’s product menu under crowdfunding can be higher, lesser or even non‐existent than that under traditional selling. We also show that crowdfunding as an alternative to traditional selling is only attractive in the presence of a non‐zero fixed cost. Our findings provide important implications for entrepreneurs in optimizing their crowdfunding project design.