|
|
|
|
|
|
|||||||||
|
Statistical modeling techniques can be used to effectively predict customer behavior, segment markets, and identify important relationships. InfoTrends has experience using statistical modeling techniques for several different applications. Conjoint Analysis The respondent chooses, ranks, or scores different product attributes such as features, brand, price, or any other factor associated with the product. This analysis is most valuable when developing or testing new products to predict what features the product should contain and promote. Discrete Choice Analysis This modeling technique is similar to conjoint in that it predicts customer preferences for product features. It allows the respondent to make choices, however, not just indicate preferences, for a particular product. Discrete choice also more closely mirrors the competitive environment. Multivariate Techniques Multivariate analysis is a family of techniques that allows you to interpret data from an overall perspective, rather than on a question-by-question basis. It enables you to simultaneously examine a large number of questions with one picture or perceptual mapping. Multi-Dimensional Scaling Respondents rate similarities and differences of competitive products, brands, or companies. By using this information, along with purchase behavior, the analysis can uncover where the customer would like to see a product positioned, as well as determine the relationship between you and your closest competitors. Discriminant Analysis Respondents rate their willingness to purchase certain products, based on descriptions provided. Consumer profiles can be generated, and using factor or cluster analysis, customer segment profiles can also be determined. These profiles help predict purchase behavior, as well as consumer willingness to transition to new technology or switch brands or products. This methodology is very useful in predicting brand loyalty. |
