This paper discusses the development of goodness-of-fit indices for partial least squares (PLS) path modeling. The authors evaluate the behavior of the goodness-of-fit index (GoF) and the relative goodness-of-fit index (GoFrel) using simulated data and compare them with fit indices from covariance-based structural equation modeling (CBSEM). The simulation shows that GoF and GoFrel are not suitable for model validation. However, GoF can be useful to assess how well a PLS path model explains different data sets. PLS path modeling is popular due to its flexibility, ability to handle small samples, and ease of use with software like SmartPLS. Unlike CBSEM, which focuses on unbiased parameter estimates, PLS path modeling aims to maximize explained variability. This difference in objectives means that PLS path modeling lacks a global scalar function for model validation, a limitation addressed by the GoF, which considers both measurement and structural models. However, the GoF has limitations, especially when dealing with formative indicators or single-item measurements. The GoFrel, a normalized version of GoF, compares PLS results with those from principal component analysis and canonical correlation analysis. While GoF and GoFrel are not suitable for model validation, they can be useful for assessing model fit in specific contexts, such as validating models with different indicators. The paper concludes that while GoF and GoFrel provide some insights into model fit, they are not reliable for model validation. CBSEM is better suited for this purpose. The authors recommend further research to improve the GoF for formative measurement models and to enhance its applicability across different types of models.This paper discusses the development of goodness-of-fit indices for partial least squares (PLS) path modeling. The authors evaluate the behavior of the goodness-of-fit index (GoF) and the relative goodness-of-fit index (GoFrel) using simulated data and compare them with fit indices from covariance-based structural equation modeling (CBSEM). The simulation shows that GoF and GoFrel are not suitable for model validation. However, GoF can be useful to assess how well a PLS path model explains different data sets. PLS path modeling is popular due to its flexibility, ability to handle small samples, and ease of use with software like SmartPLS. Unlike CBSEM, which focuses on unbiased parameter estimates, PLS path modeling aims to maximize explained variability. This difference in objectives means that PLS path modeling lacks a global scalar function for model validation, a limitation addressed by the GoF, which considers both measurement and structural models. However, the GoF has limitations, especially when dealing with formative indicators or single-item measurements. The GoFrel, a normalized version of GoF, compares PLS results with those from principal component analysis and canonical correlation analysis. While GoF and GoFrel are not suitable for model validation, they can be useful for assessing model fit in specific contexts, such as validating models with different indicators. The paper concludes that while GoF and GoFrel provide some insights into model fit, they are not reliable for model validation. CBSEM is better suited for this purpose. The authors recommend further research to improve the GoF for formative measurement models and to enhance its applicability across different types of models.