Expense fraud detection rate

Expense fraud detection rate measures how effectively the organisation's controls catch fraudulent spend. Fraud takes many forms: fictitious expense claims, inflated receipts, personal purchases disguised as business expenses, collusion with vendors, and stolen card credentials. Studies estimate that the average organisation loses 5% of revenue to fraud annually, with expense fraud being one of the most common types.

The metric is inherently difficult to measure precisely because undetected fraud is, by definition, unknown. Organisations typically estimate it through periodic forensic audits that sample historical transactions and extrapolate the undetected rate. The goal is to push detection rates as close to 100% as possible while keeping false positive rates manageable.

Expense Fraud Detection Rate = (Fraudulent Transactions Detected / Total Fraudulent Transactions) x 100

For example, if a forensic audit of 10,000 transactions identifies 50 fraudulent items, and the real-time controls had already flagged 40 of them, the detection rate is 80%. Supplement the rate calculation with the value of fraud prevented and the average time to detection. Catching fraud within 24 hours limits loss; catching it during the annual audit means the money may be unrecoverable.

Implement automated rules that flag common fraud patterns: duplicate receipts, round-number claims, transactions just below approval thresholds, weekend and holiday purchases, and rapid successive transactions. Use machine learning to identify anomalous spending patterns that deviate from an employee's historical behaviour. Require itemised receipts for all claims above threshold. Cross-reference expense claims with calendar and travel data to verify that claimed expenses align with actual activities. Establish a confidential reporting channel so employees can report suspected fraud without fear of retaliation.