Definition. Conjoint is a family of predictive techniques that includes Full Profile Conjoint, Choice Modeling, Maximum DifferenceAnalysis, Adaptive Conjoint, Paired Comparisons and Choice-Based Conjoint among others. Each of these methods is appropriate in different situations. Conjoint forces consumers make tradeoffs between valued alternatives in a way that simulates actual decision-making in the marketplace. It provides objective, quantitative measures called utility scores that represent the true value of features to each consumer.

Uses. Product design and pricing, brand equity, conflict resolution.

Benefits.  Conjoint Analysis has many potential benefits. For example, it identifies the most valuable features and provides a simulation of demand for different combinations of features at different price levels. Dollar-calibrated Conjoint can be used to estimate the ROI of each feature. When brand is included as a feature, the results provide a measure of brand equity for the client and competitor brands. Conjoint utilities can also be used as a basis for segmenting markets based on consumer preferences and priorities.


Definition. Multivariate segmentation uses a number of variables to identify segments or clusters of consumers that are relatively homogeneous and different from consumers in other groups. QSA’s NeuroClustersm is capable of using as many as 300 variables to create these segments. QSA also has a variety of other segmentation techniques in its repertoire, including Latent Class Analysis and Decision Trees (CHAID). If desired, we will create the segmentation using all of these methods and compare their predictive power. In our experience, however, NeuroClustersm, provides higher resolution, greater stability over time and between markets, and can remove the influence of outliers more effectively than other methods. In addition, it identifies archetypal respondents, who are most typical of the segment, for follow-up qualitative research and/or concept testing.

Uses.  Understanding the structure of the market; identifying targets of opportunity; locating respondents via media, geography or the Internet, develop messaging strategies.

Benefits. More effective choice of communication channels, more precise positioning of messages, less waste circulation and more effective “storytelling” in reports and presentations.


Definition. Predictive modeling (aka Structural Equations Modeling or Path Analysis) goes beyond simple correlation and multiple regression analysis to provide a dynamic map of the factors that directly and indirectly influence attitudes and behavior and provides a quantitative measure that reflects the independent power of each factor.

Uses. Setting strategic organizational priorities, setting priorities for product/service improvement, developing message platforms for advertising, public relations and sales. The analysis is especially powerful when applied to targetable segments of consumers.

Benefits. Predictive modeling literally separates the wheat from the chaff, allowing management to focus its attention and resources on the things that matter most to its audiences and stakeholders. When applied to customer satisfaction research, for example, predictive modeling identifies problems that require immediate attention and strengths that it can leverage in customer communication. When applied to message development, it not also identifies that key messages that marketers must effectively communicate to an audience in order to achieve its goals but it also identifies the supporting messages that will make the key messages more effective, in other plans, a platform for communication.