This analysis was done in Microsoft Access with the graphs done in Microsoft Excel. The graph templates come with my book, and a discussion of Benford’s Law is in Chapter 2. Chapter 5 shows how to do the analysis in Access and Excel.

The original data set had 1,432,000 records, of which 1,040,000 records were $10 and over. Only $10 and over numbers are used for Benford’s Law graphs. The notebook computer took about two minutes to process all the records.

The Mean Absolute Deviation (MAD) is 0.0050 which rates an “Acceptable conformity” per Nigrini in Digital Analysis Using Benford’s Law.” If the MAD was 0.004 or less the conformity would be described as “close conformity.” I think that the fit is excellent for sales data. A company’s own sales numbers are often clustered around some value making the fit to Benford’s Law weak. For example, most air tickets sold on SouthWest airlines would have first digits of 1, 2, and 3.

Mark Nigrini, September 2, 2002.