We treat each client's research needs as unique, relying on our vast experience and the brilliance of our team of researchers. Within the constraints of possible limited data and research budgets we develop an analytics plan for addressing our client's research questions, rather than haphazardly applying fancy analytics to impress others. From our research design, through our data collection and analysis, we always apply the most appropriate and rigorous research methods. Our approach to research is customized, rather than syndicated or canned.
Among some of the analytics we often apply are:
General Multivariate Analysis
For ad hoc analyses we have wide-ranging experience with Principal Components and Factor Analysis, Discriminant Analysis, ANOVA/MANOVA, Logistic Regression and a range of other multivariate methods. Correspondence Analysis and Biplots are two popular techniques we provide for image research.
Consumer Segmentation
Not all consumers are the same. Some buy the same brand for the same reasons as other consumers. Others may buy the same brand, but for different reasons. Yet others may buy different brands for same reasons…or different brands for different reasons. Confusing? Yes!
We cannot treat all consumers as the same because they are not. They differ in who they are, how they behave in the marketplace as well as what they say is important to them when making consumption choices. Modern techniques such as Latent Class Clustering are ideal for identifying the consumer segments on which you should focus your resources.
Key Driver Analysis
What are the factors that are the strongest drivers of satisfaction with your service or purchase intent for your brand? Structural Equation Modeling (SEM), Path Analysis, Mixture Modeling, Bayesian Regression and MaxDiff are some of the advanced methods we offer to help untangle the web of customer perceptions and experiences and how they relate to success in the marketplace.
Econometric Modeling
When data are collected over time the usual regression methods can yield inaccurate results. This is because they are designed for cross-sectional data. If you scramble the order of the cases in the data file it will make no difference in the results! So, for example, to study the effectiveness of different advertising campaigns (Advertising Response Modeling) or model other data that are correlated across time specialized time-series methods such as ARMAX, VAR or Arellano-Bond models are recommended.
Specialized Data Analytics
It is often necessary to profile pre-determined groups of consumers or examine how some kinds of consumption behavior are related to other kinds of consumption behavior, as well as with demographics. We use advanced methods that are tailored to the particular requirements of a project instead of “black box” software.
The above are just some of the wide range of creative analytics we can apply to uncover market dynamics.
Posts
