ASSESSMENT OF DATA-DRIVEN VENDOR PERFORMANCE EVALUATION IN RETAIL SUPPLY CHAINS ANALYZING METRICS, SCORECARDS, AND CONTRACT MANAGEMENT TOOLS

Abstract

This quantitative explanatory study assessed data-driven vendor performance evaluation in retail supply chains by integrating multi-KPI scorecards with downstream operational outcomes. The measurement architecture and variable logic were grounded in a structured review of 32 empirical papers on retail vendor evaluation, KPI taxonomies, scorecard construction, and outcome modeling, which informed the selection of delivery, quality, cost, and flexibility indicators. Operational data were compiled from ERP purchasing tables, WMS receipt logs, TMS carrier scans, POS demand records, supplier portal submissions, and claims databases. Of 312 vendors initially identified, 47 inactive vendors and 23 vendors with insufficient transaction histories were excluded, yielding 242 active vendors observed from January 2021 to December 2021. After data cleaning, 3,805 vendor–period records remained for analysis. Descriptive results showed strong average delivery performance (on-time delivery mean 91.4%, SD 6.8; order fill rate mean 94.1%, SD 5.2), while lead-time deviation displayed wider dispersion (mean 2.6 days, SD 1.9). Quality performance was generally stable but risk-concentrated (defect rate mean 1.9%, SD 1.4; return ratio mean 2.7%, SD 2.0), and flexibility exhibited the highest volatility (rush-order acceptance mean 76.8%, SD 14.9; recovery time mean 4.1 days, SD 3.3). Reliability tests supported dimensional consistency (Cronbach’s α 0.86 delivery, 0.83 quality, 0.79 cost, 0.76 flexibility; overall scorecard α 0.88). Correlation patterns confirmed KPI coherence (on-time delivery–fill rate r 0.74; defect rate–return ratio r 0.69) and strong alignment between vendor scores and retail outcomes (overall score–stockout r −0.58; overall score–shelf availability r 0.62). Driver models indicated that forecast variability (β −0.27, p < .001), distance (β −0.14, p = .010), and product complexity (β −0.17, p = .003) reduced vendor performance, while vendor size improved scores (β 0.18, p = .002). Outcome regressions demonstrated temporally ordered effects, with delivery strength reducing stockouts (β −0.36) and quality strength lowering return processing costs (β −0.29). Overall, the study showed that data-driven scorecards reliably differentiated vendor capability and explained meaningful variation in retail service and cost outcomes.