Customer satisfaction surveys yield a lot of useful information. For instance:
In addition to all this valuable information, management needs to know the factors that drive customer loyalty. Only then can they prioritize resources and initiatives in products/services development and people development.
Derived importance models employ statistical techniques such as multiple regression to determine the impact of each factor on an outcome such as customer loyalty. The relationship can be represented by equations of the type depicted in Exhibit 6.8. In this hypothetical illustration the model reveals that two factors are important in driving customer loyalty. The coefficients in the equations (0.7 for Factor I and 1.8 for Factor II) determine the importance of each factor, and the factor loading of Attributes A, B and C of (0.65, 0.77 and 0.83 respectively) on Factor II reveal the relative importance of these attributes. This model tells us that Attribute C is more important than Attributes B and A, and Factor II is much more important than Factor I in driving the loyalty index.
There are a number of analytic techniques that may be employed to determine the relationship between each individual attribute (predictor variable) and outcomes such as Loyalty Index or Overall Satisfaction (outcome variable). Some of the commonly used methods are listed below:
Derived importance framework can be used in various ways with relationship and transaction surveys, to prioritize resources and business initiatives. A common application is to determine the products and services that have greatest impact in driving customer satisfaction. Reverting to the example of retail banking, the coefficients (b1, b2, b3 … and d1, d2, d3 …) of the following multivariate regression model, indicate the importance of the banking services in driving overall satisfaction, and the importance of relationship manager (RM) attributes in driving "Overall Satisfaction of RM":
To diagnose issues, it is often useful to construct derived importance models for a specific group of respondents (e.g. a loyalty, behavioural, or demographic segment). For instance to determine the factors contributing to dissatisfaction among respondents in the "dissatisfied" segment, a derived importance model would need to be crafted for that loyalty segment.
It is also desirable to test for interactions that involve two or more variables. For example, if a customer transacting with a relationship manager experiences good personalized service (PS), and during the same interaction receives useful tips on market developments Knowledge of market (KM), the combined effect may be greater than the sum of the parts. This synergistic impact is captured with the use of an interaction term (PS × KM). Since these interactions do exist, modellers need to test for their presence.
One of the issues with customer satisfaction data is the high incidence of collinearity, i.e., the existence of significant relationships among two or more predictor variables. Signs of collinearity include the presence of coefficients with reversed signs, large margins of error on the coefficients and hypersensitivity. The impact of harmful collinearity is mitigated by using analytical approaches such as ridge regression and principal component regression.
Note: To find content on MarketingMind type the acronym ‘MM’ followed by your query into the search bar. For example, if you enter ‘mm consumer analytics’ into Chrome’s search bar, relevant pages from MarketingMind will appear in Google’s result pages.
Reporting, analysis and visualization solutions for customer satisfaction research.
Is marketing education fluffy too?
Marketing simulators impart much needed combat experiences, equipping practitioners with the skills to succeed in the consumer market battleground. They combine theory with practice, linking the classroom with the consumer marketplace.