Sponsored search advertising is ascendant--Forrester Research reports expenditures rose 28% in 2007 to $8.1 billion and will continue to rise at a 26% compound annual growth rate [VanBoskirk, S. 2007. U.S. interactive marketing forecast, 2007 to 2012. Forrester Research (October 10)], approaching half the level of television advertising and making sponsored search one of the major advertising trends to affect the marketing landscape. Yet little empirical research exists to explore how the interaction of various agents (searchers, advertisers, and the search engine) in keyword markets affects consumer welfare and firm profits. The dynamic structural model we propose serves as a foundation to explore these outcomes. We fit this model to a proprietary data set provided by an anonymous search engine. These data include consumer search and clicking behavior, advertiser bidding behavior, and search engine information such as keyword pricing and website design. With respect to advertisers, we find evidence of dynamic bidding behavior. Advertiser value for clicks on their links averages about 26 cents. Given the typical $22 retail price of the software products advertised on the considered search engine, this implies a conversion rate (sales per click) of about 1.2%, well within common estimates of 1%-2% [Narcisse, E. 2007. Magid: Casual free to pay conversion rate too low. GameDaily.com (September 20)]. With respect to consumers, we find that frequent clickers place a greater emphasis on the position of the sponsored advertising link. We further find that about 10% of consumers do 90% of the clicks. We then conduct several policy simulations to illustrate the effects of changes in search engine policy. First, we find the search engine obtains revenue gains of 1% by sharing individual-level information with advertisers and enabling them to vary their bids by consumer segment. This also improves advertiser revenue by 6% and consumer welfare by 1.6%. Second, we find that a switch from a first- to second-price auction results in truth telling (advertiser bids rise to advertiser valuations). However, the second-price auction has little impact on search engine profits. Third, consumer search tools lead to a platform revenue increase of 2.9% and an increase of consumer welfare by 3.8%. However, these tools, by reducing advertising exposures, lower advertiser profits by 2.1%.sponsored search advertising, two-sided market, dynamic game, structural models, empirical IO, customization, auctions

We propose a structural model of consumer sequential search under uncertainty about attribute levels of products. Our identification of the search model relies on exclusion restriction variables that separate consumer utility and search cost. Because such exclusion restrictions are often available in online click-stream data, the identification and corresponding estimation strategy is generalizable for many online shopping websites where such data can be easily collected. Furthermore, one important feature of online search technology is that it gives consumers the ability to refine search results using tools such as sorting and filtering based on product attributes. The proposed model can integrate consumers’ decisions of search and refinement. The model is instantiated using consumer click-stream data of online hotel bookings provided by a travel website. The results show that refinement tools have significant effects on consumer behavior and market structure. We find that the refinement tools encourage 33% more searches and enhance the utility of purchased products by 17%. Most websites by default rank search results according to their popularity, quality, or relevance to consumers (e.g., Google). When consumers are unaware of such default ranking rules, they may engage in disproportionately more searches using refinement tools. Consequently, overall consumer surplus may deteriorate when total search cost outweighs the enhanced utility. In contrast, if the website simply informs consumers that the default ranking already reflects product popularity, quality, or relevance, consumers search less and their surplus improves. We also find that refinement tools lead to a less concentrated market structure. This paper was accepted by J. Miguel Villas-Boas, marketing.

No abstract

The authors analyze the impact of a tax on sweetened beverages using a unique dataset of prices, quantities sold, and nutritional information across several thousand taxed and untaxed beverages for a large set of stores in Philadelphia and its surrounding area. The tax is passed through at an average rate of 97%, leading to a 34% price increase. Demand in the taxed area decreases by 46% in response to the tax. Cross-shopping to stores outside of Philadelphia offsets more than half of the reduction in sales in the city and decreases the net reduction in sales of taxed beverages to only 22%. There is no significant substitution to bottled water and modest substitution to untaxed natural juices. The authors show that tax avoidance through cross-shopping severely constrains revenue generation and nutritional improvement, thus making geographic coverage an important policy decision.

Because utility/profits, state transitions, and discount rates are confounded in dynamic models, discount rates are typically fixed for the purpose of identification. The authors propose a strategy of identifying discount rates. The identification rests on imputing the utility/profits using decisions made in a context in which the future is inconsequential, the objective function is concave, and the decision space is continuous. They then use these utilities/profits to identify discount rates in contexts in which dynamics become material. The authors exemplify this strategy using a field study in which cell phone users transitioned from a linear to a three-part-tariff pricing plan. They find that the estimated discount rate corresponds to a weekly discount factor (.90), lower than the value typically assumed in empirical research (.995). When using a standard .995 discount factor, they find that the price coefficient is underestimated by 16%. Moreover, the predicted intertemporal substitution pattern and demand elasticities are biased, leading to a 29% deterioration in model fit and suboptimal pricing recommendations that would lower potential revenue gains by 76%.

We review research on revenue models used by online firms who offer digital goods. Such goods are nonrival, have near zero marginal cost of production and distribution, low marginal cost of consumer search and low transaction costs. Additionally, firms can easily observe and measure consumer behavior. We start by asking what consumers can offer in exchange for digital goods.We suggest that consumers can offer their money, personal information, or time. Firms, in turn, can generate revenue by selling digital content, brokering consumer information, or showing advertising. We discuss the firm's trade-off in choosing between the different revenue streams, such as offering paid content or free content while relying on advertising revenues. We then turn to specific challenges firms face when choosing a revenue model based on either content, information or advertising. Additionally, we discuss nascent revenue models that combine different revenue streams such as crowdfunding (content and information) or blogs (information and advertising). We conclude with a discussion of opportunities for future research including implications for firms' revenues models from the increasing importance of the mobile internet.

No abstract

We combine COVID-19 case data with mobility data to estimate a modified susceptible-infected-recovered (SIR) model in the United States. In contrast to a standard SIR model, we find that the incidence of COVID-19 spread is concave in the number of infectious individuals, as would be expected if people have inter-related social networks. This concave shape has a significant impact on forecasted COVID-19 cases. In particular, our model forecasts that the number of COVID-19 cases would only have an exponential growth for a brief period at the beginning of the contagion event or right after a reopening, but would quickly settle into a prolonged period of time with stable, slightly declining levels of disease spread. This pattern is consistent with observed levels of COVID-19 cases in the US, but inconsistent with standard SIR modeling. We forecast rates of new cases for COVID-19 under different social distancing norms and find that if social distancing is eliminated there will be a massive increase in the cases of COVID-19.