Online Demand Under Limited Consumer Search

Marketing Science (Impact Factor: 2.36). 11/2010; 29:1001-1023. DOI: 10.2139/ssrn.1340267
Source: DBLP


Using aggregate product search data from, we jointly estimate consumer information search and online demand for durable goods. To estimate demand and search primitives, we introduce an optimal sequential search process into a model of choice and treat the observed market-level product search data as aggregations of individual-level optimal search sequences. The model builds on the dynamic programming framework by Weitzman (1979) and combines it with a choice model. At the individual level, the model has several attractive properties including closed-form expressions for the probability distribution of alternative search sets and breaking the curse of dimensionality. Using numerical experiments, we verify the model's ability to identify consumer tastes and search cost from product search data. Empirically, the model is applied to the camcorder online market and is used to answer manufacturer questions about market structure and competition, and to address policy maker issues about the effect of recommendation tools on consumer surplus outcomes. We find that consumer search for camcorders is typically limited to about 10 choice options, and that this affects the estimates of own and cross-elasticities. We also find that the vast majority of the households benefit from the's product recommendations via lower search costs.

Download full-text


Available from: Paulo Albuquerque
  • Source
    • "Clickstream data can be used to quantify search behavior using machine learning techniques [5], mostly focused on purchase records. While purchasing indicates consumers final preferences in the same category, search is also an essential component to measure intentionality towards a specific category. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We propose a robust classifier to predict buying intentions based on user behaviour within a large e-commerce website. In this work we compare traditional machine learning techniques with the most advanced deep learning approaches. We show that both Deep Belief Networks and Stacked Denoising auto-Encoders achieved a substantial improvement by extracting features from high dimensional data during the pre-train phase. They prove also to be more convenient to deal with severe class imbalance.
    Full-text · Article · Nov 2015
  • Source
    • "It is straightforward to extend the utility function in equation (1) to include a "no purchase," or outside good option, that is costless. The no purchase assumption is common in the search literature when a single good is purchased (Chiang, Chib, and Narasimhan, 1999; Kim, Albuqurque, and Bronnenberg 2010; de los Santos, Hortacsu, and Wildenbeest 2012; and Honka 2013). Since our data does not have no purchase choice occasions we use the former interpretation and assume not all of the budget will be spent on search. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We offer a structural model that recognizes consumers have to incur a cost in order to resolve product uncertainty, and that some individuals have a preference for variety that leads to the selection of more than one product. Nearly all purchased products are characterized by some form of product uncertainty that must be resolved prior to a purchase decision. Many consumer demand situations exist in which individuals have a preference for variety that leads the multiple products being chosen in continuous quantities. Clothing, food, books, and music are but four important examples of goods that are regularly purchased many items at a time. We model the optimal consideration set formation and subsequent product selection decision when consumers have different preferences for variety. Our model is not subject to the curse of dimensionality and can accommodate situations in which a large number of alternatives are available, and the actual products searched is unobserved. We validate our model using a Monte Carlo experiment and apply it to ice cream purchases in which 19 options are available. Our results show that it is important to recognize consumers are limited in their ability to process information and do not consider all products available before making a purchase. The results of the Monte Carlo experiment suggest ignoring search costs and/or preference for variety will lead to biased parameter estimates, including underestimating price sensitivity. In particular, applying a single product purchase model, such as the logit or probit, to situations in which multiple products are sometimes purchased will lead to erroneous managerial conclusions. We also find a direct relationship between consumers' preference for variety and the size of the consideration set, as well as the number of products purchased. What's more, the number of products in the consideration set expands before the number of product purchased increases. Finally, our results also suggest that the variance of prices in the market is a reasonable approximation of consumers' individual product search costs.
    Full-text · Article · Aug 2013 · SSRN Electronic Journal
  • Source
    • "Yao and Mela 2011), Brynjolfsson et al. (2010) have quantified such search costs as quite substantial in online settings when users are exposed to multiple offers on a computer screen, as in a shopbot setting. The reduction in search costs associated with the internet affected prices, price dispersion, product quality, online demand, market structure, unemployment, and many other areas of economic life (see, Lynch and Ariely 2000, Autor 2001, Ellison and Ellison 2009, Kim et al. 2010, etc.). "
    [Show abstract] [Hide abstract]
    ABSTRACT: We explore how internet browsing behavior varies between mobile phones and personal computers. Smaller screen sizes on mobile phones increase the cost to the user of browsing for information. In addition, a wider range of offline locations for mobile internet usage suggests that local activities are particularly important. Using data on user behavior at a (Twitter-like) microblogging service, we exploit exogenous variation in the ranking mechanism of posts to identify the ranking effects. We show (1) Ranking effects are higher on mobile phones suggesting higher cognitive load: Links that appear at the top of the screen are especially likely to be clicked on mobile phones and (2) The benefit of browsing for geographically close matches is higher on mobile phones: Stores located in close proximity to a user’s home are much more likely to be clicked on mobile phones. Thus, the mobile internet is somewhat less “internet-like”: search costs are higher and distance matters more. We speculate on how these changes may affect the future direction of internet commerce.
    Full-text · Article · Jun 2012 · SSRN Electronic Journal
Show more