Managers have to carefully analyze the market conditions and find out how sensitive their products are with respect to price and other marketing-mix variables. Only after fully understanding these sensitivities, can managers at P&G make the optimal strategic decisions regarding, (a) what is the optimum price, distribution and sizing combination to get maximum value, and (b) how to gain market share from their competitors without cannibalizing their own brands. The task of measuring these sensitivities and converting them into optimal decisions is critical to every firm.

We use advanced modeling approaches to answer the above questions for P&G Asia-Pacific. We measured the price and distribution elasticities for all their detergents in one of the most populated countries in the world by developing a system of equations based sales response models and up-tiering models (to evaluate the draw of sales from the lower tier competitiors). In order to incorporate the cross-sectional differences in the overall response coefficients and the heterogeneity in response for different SKU and States, we use a three-step random coefficient regression (RCR) approach to estimate the models. In the process of model development and estimation, we also dervive the properties of the weighted RCR estimators in the case of a system of equations which is a key contribution in this paper. Our models allow for price parameters to vary over time. These models are built to see which of P&G brand’s marketing-mix variables and competition factors have significant effects on the sales volume of P&G brands. Based on our model results, we develop Sales volume and value simulators for Tide and Ariel. These simulators enable the marketing managers at P& G to develop the best pricing/distribution/sizing strategy instantly, find out which competitive brands and SKUs they are actually competing with, and develop proactive marketing strategies to manage revenue/value growth. As a result, P&G gained over 39 million dollars in value growth over a one year period because of the implementation of the recommendations from our modeling approach.