Higher-Order PLS-Path Modeling: A useful tool for the reduction of the length of the Customer Satisfaction questionnaire

Abstract

Customer satisfaction is a key issue for every company wishing to increase customer loyalty and thereby create a better business performance. As such, considerable research and revenue has been invested in developing an accurate survey process for the assessment of consumer satisfaction. The questionnaire represents a vital part of the survey process but its length, very often, affects cooperation rates in surveys. This work focuses on using Structural Equation Modeling (SEM), specifically Partial Least Squares- Path Modeling (PLS-PM) estimated through the PLS-PM Regression Approach, as an alternative method for the analysis and study of Customer Satisfaction. In particular, the aim is to reduce the length of the questionnaire, by employing a classic questionnaire with questions related to the general satisfaction of the user, a classic PLS-PM model (using the Manifest Variables (MVs) of the higher-order Satisfaction block) and a higherorder PLS-PM model (not considering the MVs of the higher-order Satisfaction block). The objective is to demonstrate that, by eliminating the MVs related to Satisfaction and, consequently, by reducing the length of the questionnaire, a higher-order PLS-PM model produces results similar to those obtained using a classic model, both in terms of the validation of the model and the interpretation of the results.