Advanced Research Methods
Modeling
A model of anything is a representation; in the case of marketing research, a model is a representation of the way people behave in "the real world". ARM decomposes consumer behavior into the set of interconnected elements responsible for explaining why people act as they do. Then, our analyses specify how those elements are related to each other quantitatively, with the goal of being able to prioritize actions you might take and predict their impact on your business.
Among the processes we examine via modeling, we include:
- Advertising Effectiveness
- Brand Equity
- Consumer Decision Choice
- Marketing Mix
- Price Elasticity
- Volumetric Decomposition
- Customer & Employee Loyalty
Multivariate Analytic Techniques
The researchers of ICR's ARM team are skilled in a wide variety of methods, enabling us to select the right tool for your situation. But beyond that, our strong suit comes in our extensive experience with these techniques, which enables us to move beyond data to information and interpretation in light of your business goals. Some of the analyses we are prepared to apply to address your objectives:
- Factor analytic and other data reduction methods
- Regression modeling
- CHAID/CART approaches
- Cluster analysis
- Trade-off analyses (such as DCM)
- Discriminant analysis
- Thurstone scaling on paired comparisons
Research Design
As we listen to your business needs, we are actively constructing the best solution for you. We do not, in fact cannot, separate analysis from design - they are inter-determinate. ARM's extensive analytic experience across numerous industries has given us the ability to design approaches that will return to you the most insight in the most efficient manner possible.
One strategy we employ for this is to integrate disparate analyses to converge on a single, powerful outcome. Two examples:
- We may combine a choice modeling study with a perceptual mapping study to tell you not only which features will be most attractive in a new product or service, but how best to position that offering vis-a-vis your competition.
- We may link pricing models across separate stakeholder audiences (e.g., physicians, patients, managed care; insurance companies, employers and employees) to find an optimum price point for your return.
Similarly, we like to pull as much information from a single data collection exercise as possible. From a single rating grid plus a background question or two, for example, we can produce four interrelated outcomes:
- key driver analyses,
- head-to-head SWOT analyses,
- quadrant mapping,
- perceptual mapping,
to give you rich input on your competitive environment.
Our experience tells us two things: (1) many research problems are similar (2) no two research problems are the same. ARM's researchers recognize what to bring to bear on a business issue, and then custom design a solution for your specific needs.
Sampling Design and Weighting
In a perfect research world, complex weighting would be unnecessary; each respondent would represent the exact proportion of their fellows in the population. Real world pragmatics necessitate otherwise: time and budget limitations may not yield perfectly proportional sampling.
ARM provides weighting solutions for population projection and balancing on a daily basis, which is to emphasize our extensive familiarity with these questions. Employing interative proportional fitting, we work with your target figures, and as many demographic characteristics as necessary, to ensure high efficiency. And we advise on the appropriate sampling schemes at the time of study proposal to help work toward that goal of efficiency. We recognize that weighting is a necessity, but also that relying on weighting should never substitute for careful sample design.
