| This dissertation is composed of three essays. In the first essay, titled 'Massively Categorical Variables: Revealing the Information in Zipcodes,' we discuss how to use data more effectively in direct marketing campaigns. We introduce the idea of massively categorical variables (variables such as zipcode that take on too many values to be treated in the standard manner) and show how to use them in directly as explanatory variables. We also develop a decision-theoretic model to determine the financial worth of alternative data sources and statistical techniques.; In the second essay, titled 'Effort or Timing: The Effect of Bonuses,' we explore the effects of lump-sum bonuses on salespeople's behavior. In particular, we examine whether bonuses motivate salespeople to work harder or whether they inspire them to manipulate the timing of output. We find that the effects due to effort are much stronger than the effects due to timing games, a result that is consistent with the widespread use of lump-sum bonuses in industry.; In the third essay, titled 'Market-Source Decompositions of Demand,' we discuss how to determine whether marketing investments create new demand by expanding the market or by taking it from rivals. We compare several decompositions that have been suggested in the marketing and economics literatures and develop a new measure based on the elasticity of demand. We then show that the assumed model of demand greatly influences the results of decomposition.; To highlight differences among the assumed models of demand, we empirically estimate the effects of consumer- and physician-directed marketing investments on the demand for a type of pharmaceutical drug. We find that the proportion of demand created by market expansion is higher for consumer-directed marketing investments than it is for physician-directed ones. We show how these proportions change as consumers enter the market and as market shares change among brands. |
