Yi-Chun Akchen, Dmitry Mitrofanov
arXiv 8 Feb 2023 · Econometrics
arXiv:2302.04354 · PDF · DOI · OpenAlex · Extracted main text
Consideration sets play a crucial role in discrete choice modeling, where customers often form consideration sets in the first stage and then use a second-stage choice mechanism to select the product with the highest utility. While many recent studies aim to improve choice models by incorporating more sophisticated second-stage choice mechanisms, this paper takes a step back and goes into the opposite extreme. We simplify the second-stage choice mechanism to its most basic form and instead focus on modeling customer choice by emphasizing the role and power of the first-stage consideration set formation. To this end, we study a model that is parameterized solely by a distribution over consideration sets with a bounded rationality interpretation. Intriguingly, we show that this model is characterized by the axiom of symmetric demand cannibalization, enabling complete statistical identification. The latter finding highlights the critical role of consideration sets in the identifiability of two-stage choice models. We also examine the model's implications for assortment planning, proving that the optimal assortment is revenue-ordered within each partition block created by consideration sets. Despite this compelling structure, we establish that the assortment problem under this model is NP-hard even to approximate, highlighting how consideration sets contribute to nontractability, even under the simplest uniform second-stage choice mechanism. Finally, using real-world data, we show that the model achieves prediction performance comparable to other advanced choice models. Given the simplicity of the model's second-stage phase, this result showcases the enormous power of first-stage consideration set formation in capturing customers' decision-making processes.
appendix boundary found by appendix_titled_section at “Supplementary Proofs for Section~\ref{sec:model}” · 39% of the source is main text. Read the extracted text to check this.
The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.
| Reference | Intensity | Mentions | Sections | Main text | |
|---|---|---|---|---|---|
| 1 | Block, Henry David, Jacob Marschak, et al (1959) Random orderings and stochastic theories of response | 1.000 | 7 | 3 | 100% |
| 2 | Farias, V. F., S. Jagabathula, D. Shah (2013) A nonparametric approach to modeling choice with limited data | 0.950 | 7 | 5 | 86% |
| 3 | Jagabathula, Srikanth, Dmitry Mitrofanov, Gustavo Vulcano (2024) Demand estimation under uncertain consideration sets self | 0.941 | 12 | 4 | 83% |
| 4 | Hauser, John R, Birger Wernerfelt (1990) An evaluation cost model of consideration sets | 0.941 | 6 | 4 | 83% |
| 5 | Hauser, John R (2014) Consideration-set heuristics | 0.941 | 6 | 4 | 83% |
| 6 | Train, Kenneth E (2009) Discrete choice methods with simulation\/ | 0.928 | 5 | 4 | 80% |
| 7 | Chen, Yi-Chun, Velibor V Misić (2022) Decision forest: A nonparametric approach to modeling irrational choice | 0.894 | 7 | 5 | 71% |
| 8 | Manzini, Paola, Marco Mariotti (2014) Stochastic choice and consideration sets | 0.811 | 4 | 2 | 100% |
| 9 | Bronnenberg, Bart J, Michael W Kruger, Carl F Mela (2008) Database paper - the IRI marketing data set | 0.737 | 3 | 3 | 67% |
| 10 | Aouad, Ali, Vivek Farias, Retsef Levi (2021) Assortment optimization under consider-then-choose choice models | 0.737 | 3 | 2 | 100% |
Showing the top 10 of 82 scored citations.