THE IMPACT OF DATA-DRIVEN INDUSTRIAL ENGINEERING MODELS ON EFFICIENCY AND RISK REDUCTION IN U.S. APPAREL SUPPLY CHAINS

Abstract

This study investigated whether data-driven industrial engineering models improved efficiency and reduced risk in U.S. apparel supply chains, a domain with high clockspeed and volatile demand where causal mechanisms were often under-specified. The purpose was to estimate associations between the adoption intensity of analytics-supported line balancing, statistical process control, stochastic inventory and scheduling, and simulation-backed capacity planning and two outcome families: operational efficiency and operational or supply risk. The design was quantitative, cross-sectional, and multiple-case. A focused narrative review of 44 peer-reviewed studies informed construct definitions and item wording. The sample comprised 208 analyzable cloud and enterprise cases drawn from brand owners, contract manufacturers, and distribution partners, with managers as respondents and optional KPI uploads. Key variables included adoption, process standardization as a mediator, supply chain complexity as a moderator, and composites for efficiency (on-time in-full, order cycle time, throughput, unit cost, rework, appropriately reverse coded) and risk (disruption frequency, lead-time variability, stockout, returns or defects, compliance). The analysis plan specified descriptives, correlations, ordinary least squares with robust or clustered standard errors and a common control set (size, automation, capital intensity, market segment, nearshoring ratio, supplier reliability, demand clockspeed), moderated models with mean-centered interactions, bootstrapped indirect effects for mediation, and robustness checks including alternative index constructions and influence diagnostics. Headline findings indicated that higher adoption was associated with higher efficiency and lower risk after controls, with partial mediation through process standardization and attenuation of effects at very high complexity; results were stable across robustness specifications. Implications for practice included sequencing capability building from measurement to modeling to routinized execution, modularizing product and network complexity, and treating analytics pipelines as governed infrastructure while interpreting estimates as associations rather than causal effects.