The Application of Big Data Analytics in the Acquisition and Evaluation of Audit Evidence

Abstract

This study explores big data analytics' application in audit evidence acquisition/evaluation, revealing its transformative effects on traditional audit paradigms, practices, and trends. Global data volume surges at 27% CAGR, projected from 33ZB (2018) to 175ZB (2025), covering structured/semi-structured/unstructured data, expanding audit evidence sources. Traditional sampling audits have limitations (sample bias, insufficient risk identification), while big data analytics via full-data modeling enhances audit efficiency by 30%, achieves 92% risk identification accuracy (e.g., decision trees for fraud prediction), and expands scope to IoT/social media data. The research has theoretical/practical dimensions: Theoretically, an ERI model quantifies evidence quality via AHP based on Information Asymmetry/Risk Management/Decision Usefulness theories. Practically, KPMG's retail audit platform integrates 500M monthly POS records, identifies 3% abnormal stores via 200+ metrics, generates 10K+ evidence records; EY’s AI system shortens audit cycles from 45 to 28 days, boosts fraud detection to 89%. Challenges include 30–40% data inconsistency (Gartner, 2024), 25% mature big data teams (ACCA, 2023), 45% security vulnerabilities (Ponemon, 2024), requiring ISO 8000 governance, CISA certification, federated learning. Future research focuses on AI explainability, quantum computing’s impact, ESG evidence analysis, promoting auditing’s transformation to proactive risk monitoring for digital economy’s evidence trust systems.