Enables testing of product attributes (for example, price, operating cost, performance, emissions, style) against the preferences of individual customers and (by aggregation) all customers to obtain market share.  Typically used to test market response to new products or product developments.  This methodology is widely used in the auto and aircraft industries.

Initial design efforts identify attributes (product characteristics) and their ranges under consideration.  Surveys are conducted of customers (business or consumers). Details of questions can vary, as can the particular survey methodology (computer-based, interview, or mail survey).  The relative level of importance of each attribute is quantified.  Each customer’s attitude toward a change in each attribute (such as their attitude toward price at several different price levels above and below the level of interest) is measured.   From this, each customer’s (each survey respondent’s) “Preference Function” can be calculated. 

A product with specified attributes (for example, price, operating cost, acceleration, emissions levels) is tested against each customer’s preference function.  As long as the sampled customers are assumed to reflect the overall market, their combined (aggregate) response (in favor of a product with one set of attributes or another set) produces a market share forecast.  Variations in product attributes produce variations in market share.

Conjoint analysis enables product analysts and designers to test market reaction to changes in product attributes.  If done well with sufficiently detailed surveys, the results can provide designers and marketing personnel  direct feedback as to the market response (and therefore sales) resulting from specific changes in a product’s design. 

For the analysis to work correctly, the survey needs to assess reactions to all variables of importance to the customer.  For example, to obtain details about product characteristics but not to ask about attitudes toward a new-entrant manufacturer in a conservative market (e.g, aircraft), would produce misleading findings as the manufacturer identify might be more important than individual product attributes.   Also, if later design efforts change some attribute to be outside of the range used in the survey, then the survey results are not likely useful.  (That is, if attitudes were assessed in a survey for prices varying from US$15,000 to $30,000, but later there was interest in testing a product with price US$35,000, the analysis would not be considered accurate without redoing the survey.