Marketing Science
CMI offers a full range of analytical services, from simple cross-tabulations to
structural equation modeling. Having this expertise in-house ensures that the research
objectives remain the focus throughout the entire process, from design through analysis
and interpretation.
Total Unduplicated Reach and Frequency (TURF)
Originating in media campaign analysis, TURF is a useful tool in determining the
optimal mix of products in a line extension. It takes into account “reach,” which
represents the number of consumers interested in at least one product in the line,
and “frequency,” which represents the average number of purchases per household
(or business) for a product line, based on specific combinations of products in
the line.
Case Study: Brand and Category Extensions
Back to Top
Conjoint Analysis/Discrete Choice Modeling
“Trade off” approaches are extremely useful in understanding consumer decision-making.
CMI conducts conjoint analysis and discrete choice modeling to assist in new product/service
design, price sensitivity estimation, and market segmentation. Depending on client
needs and the business objectives, analysis can be conducted at the aggregate level
or the individual level (individual level analysis leads to more accurate prediction
of market share).
Case Study:
New Packaging Development
Back to Top
Structural Equation Modeling (SEM)
Structural Equation Modeling (SEM) is a powerful analytical tool that depicts the
causal relationships among a set of performance measures and latent traits underlying
these measures. SEM allows for modeling both direct and indirect relationships simultaneously
to determine the true dynamics of business processes. It deals with multicollinearity
(high correlation among performance indicators) and estimates measurement error
more accurately than traditional regression analysis. Within SEM, a wide variety
of approaches are available for handling different types of data and data distribution,
such as missing data, categorical data, and non-normal data.
Case Study: Structural Equation Modeling
Back to Top
Latent Class Analysis/Modeling (LCA)
Latent Class Analysis/Modeling (LCA) is a statistical method used to categorize
individuals into distinctive groups or classes based on their responses to a set
of indicators. It does not rely on such traditional assumptions as linearity, normal
distribution, and homogeneity. In addition, the categorical indicators do not have
to be standardized on the same scale (i.e. the indicators can be a mix of dichotomous,
trichotomous, and polytomous variables). Therefore, LCA is a more powerful and flexible
analytic technique than factor analysis, cluster analysis, and regression analysis.
The results from LCA add great insight into targeted business or marketing strategies.
Case Study: Latent Class Analysis/Modeling
Back to Top
QTabsTM
QTabsTM, CMI’s online crosstab solution, was developed by CMI to meet
the needs of corporate decision makers who want their questions answered now, without
waiting for someone else to fit additional analysis into their schedule. With QTabsTM,
users can access survey data via the Internet at their convenience, creating customized
crosstabs “on the fly”. QTabsTM is flexible in design, and
allows for significance testing, recoding of variables, and filtering.
View QTabsTM
demonstration
Back to Top