Abstract: Ignoring consideration sets in modeling customer purchase decisions may lead to biased estimation of customer preferences, yet consideration sets are difficult to infer in brick-and-mortar contexts. We show that the challenge of estimating consideration set models in brick-and-mortar contexts can partially be overcome with an emerging source of data: “heatmap data” collected using in-store sensors. In contrast with clickstream data in e-commerce settings, which identify and track individuals, heatmap data show customer traffic only in the aggregate. Despite this limitation, we show that heatmap data enable us to recover many of the benefits of individual-level data. In a setting in which consideration set probabilities are non-parametric, heatmap data enable identification of consideration set models that cannot be estimated using aggregate sales data alone. In a setting in which consideration set probabilities are parameterized, we show that heatmap data enable us to estimate a consideration set model that captures the effects of any variable on both consideration and choice decisions, even without imposing exclusion-restriction assumptions on customer behavior. We also demonstrate that heatmap data can lead to decreased finite-sample bias. Finally, we installed heatmap sensors in an apparel retail store and found that heatmap data can result in distinct coefficient estimates, improved predictive accuracy, and better estimated revenues in a product placement decision problem.
Ziaei, Z., Emadi, S. and Mersereau, A. From click to consideration: a temporal analysis of clickstream data. In preparation for submission to Manufacturing & Service Operations Management.
Abstract: While clickstream data has been accessible for decades and used in both academia and industry, its temporal dimension has remained largely unexplored. We harness this information to redefine consideration-set-based models in online contexts, challenging the common assumption that clicking on a product equates to considering that product for purchase. We propose four distinct models, each uniquely incorporating click duration, to offer different perspectives on the relationship between clicks and considerations. By empirically validating these models with a dataset from JD.com, we aim to redefine our understanding of consideration set formation in the context of online customer behavior. Our analysis demonstrates that distinguishing between clicking on a product’s webpage and considering that product for purchase leads to better goodness of fit and the predictive accuracy of models designed to forecast customer purchasing behavior. However, this separation introduces the curse of dimensionality, which substantially increases the computational burden for model estimation. While we have implemented a maximum simulated likelihood estimation that substantially alleviates the computational burden, the remaining computational burden is still notable and may outweigh the economic benefits of the improvement in model performance.
Abstract: Carbon emission has been widely studied in transportation research in general, yet research on carbon emission specifically in hazardous material transportation, remains surprisingly limited. This study presents a multi-objective model for locating transfer points and routing in a multi-modal network of hazardous materials, accounting for multiple different uncertainties while minimizing carbon emission alongside risk and cost of transportation. This study models uncertainties in accident probabilities, emission factors, and costs of establishing transfer points with polyhedral uncertainty sets and utilizes robust counterpart optimization to tackle these uncertainties. The application of the proposed approach is examined in a case study of petroleum product transportation. The results of our analyses confirm the effectiveness and robustness of the developed model. Specifically, it is found that the proposed model can reduce carbon emissions by 873 g and individuals at risk by 0.28 for every additional dollar spent.
Abstract: In the current decade, called the “decade of vaccines,” the supply chain has faced serious challenges. To solve some of these problems in a systematic way, this paper presents a robust network design model for vaccine supply chain management (SCM). At the same time, the proposed model determines strategic and tactical decisions for vaccines based on different priorities of demand as well as considering the perishability of vaccines. To handle uncertain parameters, we developed a robust counterpart model. In order to verify and demonstrate the benefits of the proposed model, we provide a real-world case study and implement both robust and deterministic models. Preliminary results indicated that the proposed robust model performs better than the deterministic one in designing a vaccine network.