How to Sell the Right Product to the Right Customer at the Right Price

Paat Rusmevichientong
Management Science & Engineering
Stanford University
November 2002

In a competitive market, price remains an important factor that influences consumers' purchasing decisions. It also determines a company's profitability. However, finding the right price for a product is hard. Different consumers want different products and they have different budgets. If we price the product too high, the consumers cannot afford it. If we price it too low, then we do not make a lot of money. My research develops a methodology that determines the optimal pricing strategy that maximizes a company's total sales.

My research is based on a project with General Motors (GM), the largest automobile manufacturer in the United States. GM has set up a website that helps consumers conduct research on their purchase of new vehicles. When a consumer visits the web site, she specifies the criteria for her vehicles and her budget. For instance, the consumer might specify that she wants a 4-door sedan under $17,000 that requires minimal maintenance. Based on the information provided by the consumer, the web site generates a list of recommended vehicles that match her criteria.

Using the data collected from the web site, I developed an algorithm for determining the pricing strategy that maximizes GM's total revenue, given the consumers' preferences and budget constraints. The analysis allows GM to assess the effectiveness of its current pricing policy, and ensures that the prices of GM vehicles are in line with consumers' willingness-to-pay. Preliminary experimental results indicate that the pricing policy determined by my algorithm improves the existing policy.

My approach to product pricing exploits the availability of data on consumers' preferences. Previously, a company determines the price of its products by conducting a focus group with a few hundred individuals. These participants fill out surveys indicating their preferences and incomes. Based on the survey data, the company extrapolates the demand for its product on a national level. The small sample size often prevents accurate estimation of the true demand.

Through the Internet, companies can now elicit preference information from a large number of consumers. The availability of such data enables us to obtain an accurate estimation of the demand, obviating the need for any extrapolation. My research exploits this opportunity by developing an algorithm that computes the optimal pricing strategy directly from the data on consumers' preferences.

In addition to the direct use of preference data, my approach also exploits possible substitutions between products. Substitutions occur when multiple products meet the consumer's needs. For instance, if the consumer is looking for a 4-door sedan, she might be indifferent between a Chevy Prizm and a Saturn. The substitutions between products provide us with flexibility in determining the price for each product, resulting in a better performance.

Currently, I am trying to incorporate some of the ideas from my research into the day-to-day pricing decisions at GM. I am also exploring applications my work to other industries.