Dr. Frank Benford, a physicist at General Electric in the 1920s, found that the first and second digits of many populations of numbers occur with a fairly consistent frequency. This has been found true, for example, of census numbers and certain accounting populations, such as accounts payable. Benford developed a model that predicted the frequency of each digit occurring in a particular location depending on the length of a number. For example, he finds that the digit #1 occurs as the first digit in about 30% of all populations, while the digit #2 occurs as the first digit in about 17.5% of all populations. On the other hand, the digit #9 occurs as the first digit only about 4.5% of the time. Therefore, digits such as 990 do not occur as often as digits such as 124. Many others have empirically verified the Benford predictions.
Auditors have found that as individuals commit fraud or make up fraudulent transactions their intuition in developing numbers for the fake documents often does not follow Benford's Law. Therefore, auditors have come to use Benford's Law to identify a wide variety of unusual transactions, including fraud, double payments, and other fictitious accounts. Audit software, such as ACL, comes with modules that allow auditors to apply Benford's Law to search for unusual patterns in populations by identifying numbering patterns that differ significantly from that predicted by Benford's Law.