ACL Blog – The Changing Seasons of Analytic Requirements

2009-10-01

The Changing Seasons of Analytic Requirements

By Christopher Stewart-Smith

When talking to prospective ACL customers, it often seems that analytic requirements come around like the seasons. This summer, I spoke to 5 large companies about retail analytics. Each of the 5 organisations has several hundred outlets across the United States and Canada. Each outlet may have up to 50 points of sale. All 5 were concerned with the same risks, and all came to us to find out what we could do to minimise loss in their business.

I wondered, “what has suddenly occurred that has made these retail companies so anxious about loss prevention at the same time? What new mandate or penalty has been introduced for a lack of controls at the point-of-sale level?”

Of course there is no new regulation: it is much simpler than that. In a tight economy where consumer spending is down, how can retailers ensure the best chance of survival? The answer is fraud detection and effective loss prevention.

For 3 of the 5 retailers, I demonstrated tests that all of them had specifically asked to see. They included (but were not limited to):

Voided transactions
  • Summarised by employee (to find unusually high totals/spikes across the staff).
  • Summarised by month/day/hour (to compare with employee attendance tables).
  • Summarised voids by store/outlet to gauge trends.
Refunded transactions
  • Show all refunds where there is has been no existing sale.
  • Show all refunds where the total does not match the initial sale.
  • Show all refunds where the method is different (Card sale vs Cash refund).
Refunded transactions
  • Compare daily POS sales data from the outlets with HQ’s ERP version of the truth

One retailer was adamant that such tests must be executed against all of their data, so they provided me with 97 million rows of POS transactions (hmm, thanks) to prove the point. So it seems that a depressed economy means that companies are focused more than ever on every single transaction in all areas of their business, despite the daunting volume.

Did I find any fraud in the data from these retail organisations? Let’s just say that I found many interesting and (as yet) unexplained anomalies involving several employees and thousands of dollars. Only the owners of the data can determine if they are fraudulent transactions, but it just goes to show that every exception adds up.

Find out more: ACL Blog

Thursday, October 1, 2009 In: Hot Topics Comments (None)

Contact

Pricing

Demo