This study investigates why iBuyers—firms that offer instant home purchases using big-data-driven pricing models—struggle with profitability due to adverse selection. To explore this mechanism, I develop a model in which homesellers privately know both their hassle costs of traditional selling and the unobserved quality of their home—features that are difficult for iBuyers to observe or encode in pricing—and use this information when deciding whether to sell to an iBuyer. Using a rich dataset of housing transactions and listings, I estimate the joint distribution of these private information components and evaluate two counterfactual strategies to improve iBuyer profitability. First, I simulate a redesigned contract that lowers the upfront payment and introduces conditional revenue sharing. This contract discourages lower-quality homes and increases expected profit through a cream-skimming mechanism. Second, I augment the iBuyer’s pricing algorithm by incorporating a novel one-dimensional projection of unstructured listing text, constructed using a large language model (LLM), which may partially reflect latent attributes related to unobserved quality. Incorporating this variable into pricing improves profitability—particularly when the contract alone provides weaker cream-skimming. Together, these findings underscore a broader lesson for markets that rely heavily on algorithmic pricing: contract design that internalizes information asymmetries, combined with the use of unstructured data, can help account for private information and mitigate unintended selection effects.

This paper expands the understanding of positioning decisions by investigating the timing of product offerings as an additional dimension of differentiation. Despite its significance, little research has examined the timing aspect, primarily due to the assumption that offering products at all times is optimal when fixed costs for offering one more day are negligible. However, when offering costs vary across periods and are correlated with other product characteristics, the timing of offering becomes crucial. It serves as a distinct product attribute that distinguishes firms from competitors and can be strategically utilized for repositioning purposes in response to market changes. This study explores the relationship between opening days, prices, and repositioning strategies in the US coffee shop industry during the COVID-19 pandemic, a period marked by increased daily opening costs. By analyzing structural demand and supply and considering optimal pricing and Nash opening conditions, I find that higher daily opening costs result in reduced daily entry and consumer surplus compared to scenarios that do not account for inter-market dependence. This research contributes to empirical IO literature by integrating insights from multi-product firms and endogenous product type entry, shedding light on the nuanced trade-offs firms face when choosing opening days and optimizing their market positioning.

We offer a theory of anticompetitive tying in two-sided markets when below-cost or negative pricing is possible. With the coexistence of two consumer groups (one regarding tying and tied goods as complementary and the other as independent), a tying-good monopolist may face difficulties in extracting rent under separate sales and wish to use tying to directly capture the large advertising revenue created in the complementary segment. We uncover two distinct mechanisms by which tying raises monopoly profits but reduces social welfare. Our theory of tying can be applied to real-world antitrust law enforcement, such as the Google Android case.