Call On Structural Equation Modeling

Many of the research questions intended to give you a retailer’s edge are tough to answer accurately because the factors interact in complex ways. For instance, suppose you want to know what motivates people to recommend your store to others.
     A statistical tool for teasing apart questions like this so they can be answered accurately is called Structural Equation Modeling (SEM). When working with a consultant who is using SEM or if you or your consultant are turning SEM findings into real-world action steps, remember that you know your retailing operations better than your consultant or a researcher does.

• Go into the project with a theory based on your experience as a retailer. The strength of SEM is it can give accurate answers to retailing questions so complex it seems like you’re trying to nail Jell-O to the wall. But SEM works only to confirm theories you already have. For our example, what do you believe to be the different types of people who might recommend your store to others and what motivates each of the different types? When researchers at University of Mannheim in Germany and University of Texas-Austin looked at this question using SEM, they started with the theory that there are two types of people who would recommend a retailer to others. The first type is an opinion leader who comes across as an expert in limited product categories. The second type has been called a market maven. This is the person who recommends stores on the basis of good prices, quality service, or high prestige across product categories. Now, does that theory make sense to you?
• Conclude as you started, by reminding yourself you know your business operations well. The researchers say that opinion leaders are more likely to recommend your store if you give them in-depth information about their specialty product categories, while market mavens are more likely to recommend you if you give them high-variety store visits. When research conclusions don’t make sense to you, start out by asking yourself if you might have misunderstood what the consultant said or you might have been blind to factors the consultant discovered. If you decide you did understand correctly and you weren’t blind, then consider that what the consultant or researcher is telling you could be nonsense.

Click below for more:
Look At Mean, Median, Mode, and Range
Skim the Data to Spot Leading Trends
Use Cluster Analysis on Customer Data
Keep Customers Happy About Data Collection