Understanding practitioners’ perceived effectiveness of lean tools across non-value adding activities: evidence from ordinal and machine learning models

The precast industry provides standardisation, quality and speed to the infrastructure sector, but there are still many non-value adding (NVA) activities that are causing the industry to struggle to deliver efficiencies, such as waiting time, excess inventory, unplanned maintenance, and workflow disruptions. Even though lean construction is widely used, it’s not well understood how effective it is for each NVA category in precast construction. This research aims to identify practitioners’ perceptions of the effectiveness of three lean tools (Just-in-Time (JIT), Continuous Improvement (CI), and Total Quality Management (TQM)) in NVA categories by creating an integrated analytical framework. The study utilizes cross sectional survey data, measured on an ordinal 5-point scale, which is then used for descriptive analysis, cumulative link modelling (CLM) and complementary machine learning approaches to explore cross-construct relationships and their predictive consistency. The results suggest that the effectiveness of lean tools is different between NVA categories and that there are context-depended cross-construct relationships, but no uniform effect. TQM exhibits a relatively high perceived effectiveness in various NVA categories, and JIT exhibits a relatively high range. The results of ordinal regression analyses highlight significant asymmetric relationships between the constructs of lean-tool and NVA. Within sample predictive evidence complements the findings from the ordinal regression using machine learning analyses. The study thus reframes the measure of the effectiveness of the lean tools as an ordinal and interaction-based phenomenon and offers practical guidance for prioritizing specific lean interventions by NVA category. This study establishes practitioner’s perceived effectiveness of lean tools as an ordinal, interaction-driven, and predictively stable phenomenon.