Improving PLS-SEM use for business marketing research

A review of studies published in Industrial Marketing Management over the past two decades and more shows that these studies not only used partial least squares structural equation modeling (PLS-SEM) widely to estimate and empirically substantiate theoretically established models with constructs, but did so increasingly. In line with their study goals, researchers provided reasons for using PLS-SEM (e.g., model complexity, limited sample size, and prediction). These reasons are frequently not fully convincing, requiring further clarification. Additionally, our review reveals that researchers' assessment and reporting of their measurement and structural models are insufficient. Certain tests and thresholds that they use are also inappropriate. Finally, researchers seldom apply more advanced PLS-SEM analytic techniques, although these can support the results' robustness and may create new insights. This paper addresses the issues by reviewing business marketing studies to clarify PLS-SEM's appropriate use. Furthermore, the paper provides researchers and practitioners in the business marketing field with a best practice orientation and describes new opportunities for using PLS-SEM. To this end, the paper offers guidelines and checklists to support future PLS-SEM applications.

Automated summary

A review of studies published in Industrial Marketing Management over the past two decades and more shows that partial least squares structural equation modeling (PLS-SEM) is widely used to estimate and validate theoretical models. However, researchers' reasons for using PLS-SEM are often not fully convincing and their assessment and reporting of measurement and structural models are insufficient. Advanced PLS-SEM analytic techniques are seldom applied, despite their potential for enhancing results' robustness and generating new insights. This paper aims to clarify the appropriate use of PLS-SEM by reviewing business marketing studies and provides guidelines and checklists for future applications.