Marketing literature and practitioners are in agreement that it is essential for brands in competitive markets to identify segments that should be targeted and to build rational product strategies that target these segments. It is essential because most markets include consumers with heterogeneous preferences and precise segmentation and targeting creates product differentiation, which prevents direct competition and allow the market to reach an optimal profit optimization equilibrium.

In this consumer markets' era, defined by practitioners as the big data era, consumers' individual transactions and actions, which reveal their preferences, became highly available to marketers. This allows marketers to greatly improve their targeting and to optimize their profits through that.

This dissertation contains three essays that examine optimal products strategies with consideration of individual distributions. Through the models that are built and estimated, individual preferences are identified. Following that, individuals are aggregated into clustered segments, and clear optimization strategy is designed.

All the essays build and discuss structural models and estimation strategies. Each estimation uses unique datasets that were selected and organized carefully for the purpose of robust identification of the varied effects that are examined and analyzed.

Each essay identifies and considers the individual distributions in the analyses. Altogether, the essays provide a deeper understanding of how to consider individual distributions in varied settings and marketing needs that marketers face frequently.

Chapter 1 examines the theory of trying, forgetting, and sales in empirical settings. This is an important model as there are many markets where consumers need to try products for realizing their fit, however after trying, some consumers may forget the fit over time through learning processes of competitive information and other processes.

The theory shows that the trying and forgetting model predicts that sales will occur periodically according to the magnitudes of the effects as the sales are used by the brands as product-fit reminders to the targeted consumers.

For the empirical examination of this theory, a model that includes trying and forgetting effects within the standard demand side model is built and estimated. The model allows consumers to have heterogeneous tastes and includes treatment for possible endogeneity.

Using the demand estimation and including an individual level distribution estimation, the population is divided into segments. Consumers are divided by their utilities for products as it is optimal for firms to target with the regular price the segments that favor the and when they launch sales, they may target more segments as the trying experience may affect their utility and make them be included in the main segment that this firm targets.

This segmentation of the data makes it possible to find the equilibrium in which each firm optimizes profits and the market does not enter a situation of direct competition and a Bertrand game as the firms focus on the segments that favor them and launch temporal sales to introduce or remind consumers of the products fit. This allows the identification prices strategies that optimize profits.

Chapter 1 also builds a novel dynamic game supply side model together with simulation strategy and technique for that. This is a major contribution as it finds the equilibrium of a multi agent, segments, states, and periods dynamic game for these common settings where firms need to design a long-term, per period, pricing menu as they cannot change their product pricing often.

The results of the estimation and simulation show that the trying and forgetting effects are highly significant on the demand side, but are not used well by some brands through their introduction period and afterward, which greatly and negatively influence their market share and long-term profits.

Chapter 2 examines a method of finding individual level preference for attributes across products and the importance that it can have on policy makers, marketers, and consumers.

It specifically discusses the case of reducing overweight in the population through finding the willingness to pay for the fat attribute of products among consumers that consistently buy fattier products at varied categories and introducing these consumers to products that are healthier for them through promotions on those products.

This is an important question as overweight is was recognized as a global epidemic and thus researchers and policy makers are consistently looking for solutions with no consistent finding yet as neither macro taxes of attributes such as sugar or fat nor or macro subsidies of healthier products were feasible, effective, or efficient.

It shows that the standard model does not allow targeted and effective promotions to these consumers as there is a gap in willingness to pay for fat through the population compared to the targeted group. However, using the estimation of the individual level distributions, this part shows that it is possible to convert this segment of consumers to choose healthier products through small magnitude promotional pricing.

Chapter 3 examines a case that is opposite to the previous chapters. While in the previous chapters the segments were revealed through the estimation and individual distribution estimation methods. The data in this chapter saliently reveals that 20\% of consumers increased their per unit spend in a durable goods category at the first months of the US sub-prime recession of 2008. This hints that a large portion of the consumers became price loving at the beginning of one of the most difficult periods of the US economy.

This is clearly the opposite to the expectation, thus chapter examines the data carefully and suggest varied models. Finally, it shows that in this case, a well specified demand model can identify the reasons for the initial confusion coming from the data.

Altogether, the essays examine frequent market settings that were not examined before and provide models together with estimation strategies and methods, which allow better optimization of product strategies through the consideration of individual level distributions and through segmenting the population accordingly.

Sellers of almost any product or service rarely keep their prices constant through time, and frequently offer price discounts or sales. This paper investigates an explanation of sales as a way for uninformed consumers to be willing to experience the product, and learn about its fit, and where informed consumers may forget about (or change) their preferences. We investigate the role of the rate of forgetting on the timing between sales, and of the rate of learning and menu costs on the length of a sale. We also investigate the effect of a seller carrying multiple products on the pattern of sales. Using price series from supermarket categories, and given the assumed simplified preference structure, we obtain empirical estimates of the rates of learning and forgetting, and of the other model parameters.

My thesis consists of three essays in the field of operations management and marketing.

In the first essay, I study the problem of consumer search for information on multiple products. When a consumer considers purchasing a product in a product category, the consumer can gather information sequentially on several products. At each moment the consumer can choose which product to gather more information on, and whether to stop gathering information and purchase one of the products, or to exit market with no purchase. Given costly information gathering, consumers end up not gathering complete information on all the products, and need to make decisions under imperfect information. I solve for the optimal search, switch, and purchase or exit behavior in such a setting, which is characterized by an optimal consideration set and a purchase threshold structure. It is shown that a product is only considered for search or purchase if it has a sufficiently high expected utility. Given multiple products in the consumer's consideration set, the consumer only stops searching for information and purchases a product if the difference between the expected utilities of the top two products is greater than some threshold. Comparative statics show that negative information correlation among products widens the purchase threshold, and so does an increase in the number of the choices. Under my rational consumer model, I show that choice overload can occur when consumers search or evaluate multiple alternatives before making a purchase decision. I also find that it is optimal for sellers of multiple products to facilitate information search for low-valuation consumers, while obfuscate information for those with high valuations.

In the second essay, I conduct an empirical study of peer effects of iPhone adoptions on social networks. I use a unique data set from a provincial capital city in China, in a span of over four years starting from iPhone's first introduction to mainland China. I construct a social network using six month's call transactions between iPhone adopters and all other users on a carrier's network. Strength of social ties is measured by duration of calls. Based on the network structure, I test whether an individual's adoption decision is influenced by his friends' adoptions. A fixed-effect model shows that, on average, a friend's adoption increases one's adoption probability in next month by 0.89%, and the marginal effect decreases in the size of his current neighboring adopters. To further control for potential time-varying correlated unobservables, I instrument adoptions of one's friends by their birthdays, based on the fact that consumers are more likely to adopt iPhones on birthdays. The IV estimation shows a slightly smaller peer effect at 0.75%. I also investigate how network structures modulate the magnitude of peer influence. My results show that peer effect is stronger when the influencer has more friends or has a stronger relationship with the influence.

In the third essay, I study the problem of coordination of operations and marketing decisions for new product introductions. In the industry with radical technology push or rapidly changing customer preference, it is firms' common wisdom to introduce high-end product first, and follow by low-end product line extensions. A key decision in this "down-market stretch" strategy is the introduction time. High inventory cost is pervasive in such industries, but its impact has long been ignored during the presale planning stage. This essay takes a first step towards filling this gap. I propose an integrated inventory (supply) and diffusion (demand) framework, and analyze how inventory cost influences the introduction timing of product line extensions, considering substitution effect among successive generations. I show that under low inventory cost or frequent replenishment ordering policy, the optimal introduction time indeed follows the well-known "Now" or "Never" rule. However, sequential introduction becomes optimal as the inventory holding gets more substantial or the product life cycle gets shorter. The optimal introduction timing can increase or decrease with the inventory cost depending on the marketplace setting, requiring a careful analysis.

The recent spread of COVID-19 across the U.S. led to concerted efforts by states to ``flatten the curve" through the adoption of stay-at-home mandates that encourage individuals to reduce travel and maintain social distance. Combining data on changes in travel activity with COVID-19 health outcomes and state policy adoption timing, we characterize nationwide changes in mobility patterns and isolate the portion attributable to statewide mandates. We find evidence of dramatic nationwide declines in mobility prior to adoption of any statewide mandates. Once states adopt a mandate, we estimate further mandate-induced declines between 2.1 and 7.0 percentage points across methods that account for states' differences in travel behavior prior to policy adoption. In addition, we investigate the effects of stay-at-home mandates on changes in COVID-19 health outcomes while controlling for pre-trends and observed pre-treatment mobility patterns. We estimate mandate-induced declines between 0.13 and 0.17 in deaths (5.6 to 6.0 in hospitalizations) per 100 thousand across methods. Across 43 adopting states, this represents 23,366-30,144 fewer deaths (and roughly one million averted hospitalizations) for the months of March and April - which indicates that death rates could have been 42-54% higher had states not adopted statewide policies. We further find evidence that changes in mobility patterns prior to adoption of statewide policies also played a role in reducing COVID-19 mortality and morbidity. Adding in averted deaths due to pre-mandate social distancing behavior, we estimate a total of 48-71,000 averted deaths from COVID-19 for the two-month period. Given that the actual COVID-19 death toll for March and April was 55,922, our estimates suggest that deaths would have been 1.86-2.27 times what they were absent any stay-at-home mandates during this period. These estimates represent a lower bound on the health impacts of stay-at-home policies, as they do not account for spillovers or undercounting of COVID-19 mortality. Our findings indicate that early behavior changes and later statewide policies reduced death rates and helped attenuate the negative consequences of COVID-19. Further, our findings of substantial reductions in mobility prior to state-level policies convey important policy implications for re-opening.

Take Away Link https://are.berkeley.edu/sites/are.berkeley.edu/files/PolicyTakeAway_Web.pdf