Determining Consumer Preference (Part I)

Anyone launching a new product or service ultimately faces three key questions: (1) what product or service should I launch, (2) who are my target customers, and (3) how much will these customers be willing to pay? Thankfully, there are several methods one can use to obtain answers to these questions.

One of the most widely used techniques, Conjoint Analysis, was developed in the early 1970s by Mark Green, a professor at the University of Pennsylvania’s Wharton School of Business. Market researchers routinely use Conjoint Analysis to determine how people value different features that make up an individual product or service. The technique is so flexible that it has been applied to products ranging from condominium construction to services in health care.

How does Conjoint Analysis Work?
Conjoint Analysis rests on three notions; first that any product or service can be broken down into a set of attributes; second, that consumers can provide accurate feedback about the importance and desirability of each attribute relative to each other; and third, that this consumer feedback can be used to predict acceptance of said product or service in the marketplace.

Determining Product/Service Attributes
Attribute selection is the most important component of Conjoint Analysis. Omitting attributes that consumers care the most about or conversely, including ones that are not important will substantially and detrimentally alter the analysis’s conclusions. The onus of correct attribute selection therefore rests firmly with the individual conducting the analysis. In order to mitigate the influence of personal bias in the creation of an attribute list, many marketing managers rely heavily upon the cross-functional perspectives of their colleagues.

Obtaining Consumer Feedback
After finalizing a list of attributes that define the hypothetical product or service, the next step is to obtain consumer feedback. The simplest approach to obtaining feedback is to ask consumers to rank the list of attributes by preference. Sometimes, however, a simple rank ordering isn’t straightforward. What if, for example, the attribute list is twenty-five or even fifty items long? To address the problem of a lengthy attribute list, some product managers carefully select a subset of attributes from the list. Others ask consumers to bucket the list into three categories: (1) attributes they like very much, (2) those they like moderately, (3) and those they don’t like at all, and then ask the consumers to rank order the items in each bucket. Another approach is to ask the consumer to assign each attribute a probability of purchase, i.e. “how likely are you to purchase this product or service if this attribute was present?” Finally, one can also ask consumers to rate each attribute on a 1-10 scale.

Evaluating Consumer Feedback
Evaluating consumer feedback data depends largely on how the data was collected. In the previous paragraph, we discussed three different data collection methods: (1) rank ordering of attributes, (2) rating of attributes on a 1-10 scale, and (3) assigning attributes a probability of purchase. If the data was rank ordered, the standard approach is to use a monotone analysis of variance (MONANOVA). If the attributes were rated on a 1-10 scale, then a regression analysis is used. If the attributes were assigned probabilities, then one typically uses the logit model.

MONANOVA, Logit, Regression? Do you have an example?
I know, it can be awfully confusing. Don’t worry though, Part II of this series on Conjoint Analysis, will cover the three different techniques on evaluating consumer feedback in depth. And in Part III, we will review the analysis’s fundamental concepts and close with an example applied to our very own SiphsMail service. So stay tuned or subscribe to our RSS feed to have the posts automatically delivered to your inbox. Have a great week!

References
[1] Conjoint Analysis (in Marketing), Wikipedia, http://en.wikipedia.org/wiki/Conjoint_analysis_(in_marketing)
[2] Conjoint Analysis, A Managers Guide: Harvard Business School Case # 9-590-059, Professor Robert J. Dolan

June 9, 2008 |

Determining Consumer Preference (Part I)

Comments

Subscribe

Recent Posts