STRUCTURAL EQUATION MODELING DENGAN ALTERNATIF COVARIANCE BASED (Studi Kasus: Analisis Faktor-Faktor yang Mempengaruhi dan Dipengaruhi Kepuasan Pelanggan Marisini Coffee di Kota Yogyakarta)

Structural Equation Modeling (SEM) is a multivariate analysis technic developed from factor analysis and path analysis. SEM’s software came from developed covariance analysis, its first software released was LISREL (Linear Structural Relationship). Covariance based SEM is one of SEM’s types requiring a construct and other indicators to corelate each other in a structural model. Generally, CB-SEM aims to estimate a structural model based on firm studies to verify causality relation between constructs, also to measure model properness and confirms its accordance to empirical data. Consequences of using CB-SEM is it insists firm theories, complying parametric assumptions and qualified model properness (goodness of fit). Therefore, CB-SEM precisely used to verify theories and justification upon those testing using complex analysis. In this study, writer researches about structural equation modeling using covariance based alternative with reflective indicator towards customer satisfaction Marisini Coffee coffeeshop located in Yogyakarta on 2020. The result from LISREL 8.80 output that direct effect between latent variables is positively influenced. Indirect effect between latent variables occurs negatively influenced, cost variable isn’t significantly influencing loyalty through satisfaction