Uncovering Motivations & Key Drivers
Episode 07
Market Segmentation: Episode 07 Transcript
Hi, my name is Jerry Thomas and I’m President/CEO of Decision Analyst. Beth [Horn] in the previous video was talking about various segmentation techniques but one thing that’s really unique about segmentation studies is that we have a large database, we have typically large samples, we have long questionnaires, so we end up with this big database that’s perfect in structure where we have a hundred variables, or maybe 200 variables or maybe even more. And rarely in marketing research do we have such a wonderful database to work with.
Given that you have that database there are all kinds of interesting and exciting things you can do. One, of course, is plain old vanilla, cross tabulation, and you ought to take that database, that segmentation database, and look at it by age, and income, and gender, and ethnicity, and geography. There’s much that you can learn from just simple basic cross tabulation analysis.
Then, I would recommend, next is [factor analyses] of the batteries of attitudes, or the batteries of behaviors. However the questionnaire is structured, you’ll have these batteries of attitudinal statements. Do factor analyses on those statements, and this will help you understand the structure of the ideas, which ideas tend to correlate with each other and go together.
And then you can also do key driver analysis and in key driver analyses we’re taking a variable, one variable, and it could be maybe, how much peanut butter you use in a given month, or how often you go to a convenience store, but that’s the variable we want to explain and we can use the other variables in this perfect database we have, and run regression analyses, or correlation or discriminant analyses, or other techniques to help us understand why one variable is greater or lesser, or goes up or goes down. And then you can run hundreds of these types of analyses with this wonderful segmentation database that you have, and it’s seldom done, but there’s a lot of extremely valuable learning that can be achieved through key driver analyses.
I’d like to leave you with a few caveats. When you’re doing this type of analysis, typically you measure how good the prediction, or the explanation of the dependent variable is with something called the ‘coefficient of determination,’ often abbreviated as ‘R square.’ If that ‘R square’ is too high, like it’s above 60 or 70%, you probably have, [in your independent variables a] variable that’s highly correlated with your dependent variable.
So be very cautious if you have an R square that’s too high, likewise, if that R square is really low don’t pay any attention to it, because you’re going to have, based purely on random chance, you’re going to have some correlations of some variables to other variables. So you want reasonably high R square but not too high. And then keep in mind for all of these multivariate techniques that I’ve been talking about that they’re based on correlation or correspondence or association, they do, they indicate, they suggest, causation but they don’t prove causation. So always be careful in how you interpret those results.
Our next video will be on understanding and profiling market segments.
Presenter
Jerry W. Thomas
Chief Executive Officer
Jerry is deeply involved in the firm’s development of new research methods and techniques and in the design of new software systems. He plays a key role in the development of Decision Analyst’s proprietary research services and related mathematical models.
Jerry describes himself as a student of marketing strategy, new product development, mathematical modeling, business survival, and economic growth. In his spare time, he likes to work on his farm in East Texas where he grows grapes, apples, pears, pecans, plums, and peaches; a forest of native trees, grasses, and insects; and wild plants of many types.
He graduated from the University of Texas at Arlington, earned his MBA at the University of Texas at Austin, and studied graduate economics at SMU.
