This study develops an accounting-analytics framework for continuous auditing of inventory integrity and working capital control in ERP-enabled supply chains. The empirical analysis uses the DataCo SMART SUPPLY CHAIN FOR BIG DATA ANALYSIS dataset, a public transaction-level dataset distributed through Mendeley Data and mirrored on Kaggle, containing 180,519 records and 53 variables spanning 2015 to 2018. The paper operationalizes inventory-integrity stress through a set of ERP-visible exception signals: delay variance, severe delay, negative-profit orders, high discounting, workflow-status anomalies and fraud labels. Descriptive analytics, cross-segment heat maps, Pareto concentration analysis and logistic regression are used to identify where risk is concentrated and how a continuous-auditing rule library can be designed. The results show that 54.83% of observations carry late-delivery risk, 18.71% are negative-profit orders, 11.11% involve high discounting and 2.25% are fraud-labelled orders. Positive shipping slippage averages 1.62 days and generates 34.01 million sales-days of working-capital drag, equivalent to 0.92 extra cycle days across the portfolio. Shipping mode, rather than broad market geography, explains most late-delivery variation: relative to First Class, the odds of late-delivery risk are materially lower for Same Day, Second Class and Standard Class channels. Exposure concentration is also non-trivial: a small set of product categories, led by Fishing, Cleats and Camping & Hiking, accounts for a disproportionate share of negative-profit sales exposure. On the basis of these findings, the paper proposes a six-rule continuous-auditing architecture for ERP-enabled supply chains, combining transaction screening, exception scoring, root-cause triage, remediation ownership and working-capital escalation. The contribution of the study lies in translating continuous-auditing theory into a practical control design that links transaction analytics with inventory integrity, margin protection and liquidity discipline.
Chingezi et al. (Tue,) studied this question.