Utilizing relative weight analysis in customer satisfaction research

Abstract

For customer satisfaction researchers, key driver analysis is a common practice to understand what product and service attributes are most important in driving the overall customer experience, typically measured by overall satisfaction or the Net Promoter question. To implement key driver analysis, market research practitioners often use statistical techniques such as bivariate correlation analysis or multiple regression, yet these statistical techniques have severe limitations for conducting key driver analysis. As a viable alternative, this article puts forth relative weight analysis (RWA) as an appropriate statistical technique for conducting key driver analysis. To empirically demonstrate this technique, key driver analysis was conducted using data from a B2B software provider. The analysis implemented RWA, correlation analysis, and multiple regression, and the results are compared and contrasted. RWA limitations and best practices are discussed as well as future research. Overall, this research advocates for more use of RWA in marketing research.

Comparing the Relative Importance of Predictors of Intention to Use Bicycles

“…The same can be said in the case of intentions to use the bike: indeed, it is more likely that in deciding to use the bike for daily commuting, people consider all the factors involved (e.g., presence of infrastructures, the opinion of friends and relatives) and attribute a weight to each one. Dominance analysis was employed as a valuable statistical tool to determine the importance of predictors in various studies conducted in various psychological fields, such as cognition (Gellersen et al, 2021), customer satisfaction (Garver and Williams, 2019), family (Oyarzún-Farías et al, 2021), personality (Duan et al, 2021), organization (Simonet et al, 2019), and religion (Hichy et al, 2014). All those studies have identified the variables that influence the criterion; moreover, they have the advantage of establishing which variable is the most important and, therefore, on which variable it is convenient to intervene.…”

“…Dominance analysis was employed as a valuable statistical tool to determine the importance of predictors in various studies conducted in various psychological fields, such as cognition ( Gellersen et al, 2021 ), customer satisfaction ( Garver and Williams, 2019 ), family ( Oyarzún-Farías et al, 2021 ), personality ( Duan et al, 2021 ), organization ( Simonet et al, 2019 ), and religion ( Hichy et al, 2014 ). All those studies have identified the variables that influence the criterion; moreover, they have the advantage of establishing which variable is the most important and, therefore, on which variable it is convenient to intervene.…”

Predicting Big Data Adoption in Companies With an Explanatory and Predictive Model

“…The connections allow signals to travel through the network in parallel and in series. The synaptic weight is interpreted as the strength of the connection between the nodes (Behara et al, 2002 ; Garver, 2002 ).…”

“…The first of these is applied to the network training, while the second is used to validate the performance of the model. This process also prevents an excess of training or over-adjustment of the neural network that would prevent the generalization of the results to the rest of the population (Garver, 2002 ; Deng et al, 2008 ).…”

The visualization, unbiased estimation and interpretation of distributional regression models

“…While an optimal model specification ensures the correct interpretation of estimated coefficients, many applied researchers further require the assessment of the predictors' relative importance (Green, Carroll, & Desarbo, 1978). Examples include ranking contributions to customer satisfaction scores (Garver & Williams, 2020;J. Johnson, 2000) or finding drivers of species richness in conservation biology (MacNally, 1996) This lacking ability to compare and rank explanatory covariates has motivated many researchers to propose alternative metrics.…”