เปรียบเทียบวิธี

	Hierarchical Bayes Choice Model	Consideration-Set Model
สาขาวิชา	การตลาด	การตลาด
ตระกูล	Regression model	Regression model
ปีกำเนิด	2005	1991
ผู้ริเริ่ม	Peter E. Rossi, Greg M. Allenby & Robert McCulloch	John H. Roberts & James M. Lattin
ประเภท	Hierarchical Bayesian random-coefficients discrete-choice model	Two-stage discrete-choice model with latent consideration
แหล่งต้นตำรับ	Rossi, P. E., Allenby, G. M., & McCulloch, R. (2005). Bayesian Statistics and Marketing. John Wiley & Sons. ISBN: 9780470863671	Roberts, J. H., & Lattin, J. M. (1991). Development and Testing of a Model of Consideration Set Composition. Journal of Marketing Research, 28(4), 429-440.
ชื่อเรียกอื่น	HB Choice Model, Bayesian Random-Coefficients Logit, Hierarchical Bayesian Conjoint, Individual-Level Partworth Model	Consideration Set Composition Model, Consider-Then-Choose Model, Two-Stage Choice Model, Evoked Set Model
ที่เกี่ยวข้อง	3	3
สรุป	Hierarchical Bayes (HB) choice models estimate a separate set of preference weights — partworths — for every individual respondent, while borrowing strength across respondents through a shared population distribution. The model has two levels: at the lower level each person's choices follow a logit driven by their own coefficients, and at the upper level those individual coefficients are treated as draws from a common multivariate distribution whose mean and covariance are themselves estimated. Inference is Bayesian and proceeds by Markov chain Monte Carlo — typically Gibbs sampling with Metropolis steps — which yields a full posterior for each respondent's partworths rather than a single point estimate. The approach, codified by Rossi, Allenby, and McCulloch, solved a long-standing problem in choice modeling: how to recover genuine individual-level heterogeneity from the sparse data each person provides. Sparse individual estimates are stabilized by shrinkage toward the population mean, giving reliable person-level coefficients usable for segmentation, targeting, and realistic market simulation. HB is now the default estimator for conjoint and scanner-based choice analysis.	Consideration-set models formalize the empirical fact that consumers do not evaluate every available brand but choose from a small subset they actively consider. Choice is decomposed into two stages: first a brand is screened into the consideration (or evoked) set, then it competes for selection only against the other considered brands. John Roberts and James Lattin's 1991 model gave this idea a rigorous, estimable form by treating consideration as the outcome of a benefit-cost calculus — a brand is added to the set when the expected incremental benefit of including it exceeds a cost of consideration. The conditional second stage is typically a logit over the considered brands, so the unconditional choice probability is a weighted sum over possible consideration sets. Modeling the first stage matters because ignoring it biases estimated brand effects and substitution patterns: a brand can lose because it is never considered, not because it loses head-to-head. The framework underlies modern thinking about awareness, screening, and the upper funnel in brand competition.