Pricing is a Science – Not a Guessing Game

Most of us are familiar with the television game show The Price Is Right. It has aired continuously since 1972. The contestants in the show compete to win cash and prizes by guessing the pricing of merchandise. It is not unlike what happens in business. Businesses try to get the price of products and services right to maximize revenue, and when they do, there is a prize – financial success. The major difference is that optimal pricing should not be a guessing game.

All too often, businesses opt for offering their product or service at some cheap price without realizing that the customers’ perception of quality and value is strongly influenced by the price point. There is a price at which the product or service is too low for a consumer to consider buying it. And, of course, it is quite possible to set the price too high. The good news is that used in conjunction with a survey, we have a range of statistical modeling techniques that allow us to estimate the optimal price at which demand is maximized.

Commonly used pricing techniques include:

Gabor-Granger

A commonly used pricing technique, the Gabor-Granger approach requires respondents to indicate how likely they are to buy a product or service at a stated price. A variety of increasingly higher prices are offered until the consumer indicates that they will purchase at that price. The optimum price for each study participant is calculated. The price wheel is a variation of this technique. Unfortunately the problem with the Gabor-Granger approach is that consumers tend to over or under state the price they are willing to pay particularly as pricing data is gathered in the absence of competitive products or other market conditions. When used with Gabor-Granger, price elasticity of demand generates a single number that summarizes price sensitivity. By segmenting the customer base, it is possible to price elasticity for different customer group.

van Westendorp with Revenue Forecast Extension

The van Westendorp approach identifies a range of acceptable prices and an optimal price point based on an analysis of price/value ratings obtained from consumers. The technique is often used with the Revenue Forecast Extension to determine the optimal price by taking purchase intent into account. Unfortunately the van Westerndorp terminology is confusing as one of the key outputs (Optimal Price Point) does not always appear at the point at which revenue is maximized. The Point of Marginal Expansiveness (PME), or that point at which market penetration and revenue fall sharply, is typically the recommended price.

Conjoint Analysis

In Conjoint Analysis study participants are asked to rank various product attributes, and they answer several questions about their interest in the product. The technique may be used to test prices traded off against other product features, providing an understanding of the importance of price relative to product features. This approach typically involves multiple regression and sometimes hierarchical Bayesian analyses. A simulator is then built allowing the client to understand the purchase intent for a large number of product feature combinations (not just those evaluated in the study).

More flexible and agile pricing methodologies include:

Discrete Choice Modeling

In a simplified version of real life, participants in Discrete Choice studies are required to choose between two or more hypothetical products or services at different price points. These studies examine real world choices a customer will make and typically include key competitive brands. The product or service under study and those of the competitors are set at different price points simulating real market conditions. The model also takes into account not making a purchase. The data is subjected to a multinomial logistic regression, and a simulator is created for the brand under study as well as competitive brands. Using the model and the simulator, estimated sales and market can be projected.

Maximum Difference (aka “Maxdiff”)

Maxdiff employs customer trade off rather than rating scales. Study participants identify the best and worst choice from a variety of groupings of three or more products or services at different prices. The order of questions is randomized and price levels are randomly assigned. A discrete choice model and a simulator are developed. Modification of price within the simulator allows the client to quickly identify what happens to demand as price increases or decreases.

Monte Carlo Simulation

The most sophisticated of the pricing approaches, the Monte Carlo Simulation takes in to account the customer’s price value perception, product, variable fixed costs, and market size. At each price point the model shows the penetration of the product or service. Typically the model output is presented in an Excel spreadsheet. The key benefit of this approach is that it allows the client to run multiple “what if” scenarios by changing the parameters in the worksheet.

Pricing should not be a game of chance. It is possible to obtain very accurate price elasticity measures using surveys and robust statistical modeling techniques. As a supplement to other types of strategic research, pricing research is a powerful way to identify the optimal price to maximize revenues.

This article was written by Kirsty Nunez.  Kirsty Nunez is President and Chief Research Strategist at Q2 Insights. Q2 Insights is a market research consulting firm with offices in San Diego and New Orleans. Kirsty can be reached at (760) 230-2950 and kirsty.nunez@q2insights.com. 

This entry was posted in Data Analysis and tagged Tags: Marketing Research, Multivariate Statistics, Pricing Research, Statistical Modeling on March 12, 2012 by Kirsty Nunez