The Many Faces Of Data

It may seem like every department in your retail operation is demanding more and more information. The most valuable source for this data comes in the form of customer demographics and sales transactions - all information that is collected by the point of sale (POS). The good news is that many departments from loss prevention (LP) to marketing and from sales audit to store management can use the same data to complete their jobs. The bad news is you have to efficiently give them access and maintain an organized database to help them make the best use of terabytes of information.

Michael Corinotis, VP of loss prevention at Odd Job Stores (South Plain Field, NJ), has been working with Brad Jarrett, CTO at Information Management Software, Inc., to transform the retailer's POS system into a data-collecting and reporting machine. Loss prevention is just one of the departments that will benefit from the new exception-based reporting capabilities.

1. Is there a wrong way to use data?

Brad Jarrett: There is the old adage that numbers don't lie, but what if you are looking at the wrong numbers? You can get numbers to say anything you want. You might have sold $2,000, but if you spent $3,000 getting there, then you didn't have all the information. The only way to misuse information is if you don't capture all of it.

Michael Corinotis: If the parameters are not clearly defined, the data becomes so voluminous that nothing gets accomplished. As with most exception reporting programs, it is trial and error until you find the parameters that work best. If loss prevention (LP) looked at every refund that totaled $25 and above, there would be too much data. The average dollar per transaction in our stores is just under $15. By looking at refunds of $75 or more, we will get a quick snapshot of whether there is fraudulent activity.

2. How many different ways can retailers use the same collected information?

Brad Jarrett: I don't think retailers realize how much of the same data can be used for multiple things. LP takes sales data and looks for inconsistencies such as abnormal voids or refunds. Marketing uses the same information to see selling trends of certain products and to decide the best time for promotions. In sales auditing, the information determines how to best manage a store's labor, for example the appropriate time to open the most register lanes. A store manager looks at how his store compares to others in terms of sale goals. The same information could be used four or five different ways.

3. How can information be managed in a way that everyone can access it?

Brad Jarrett: Retailers should use data mining and data warehousing to make sense of it all. It is nice to put information in a bucket, but who is going to sort it, who is going to manage it, and how do you get what you need out of it? What I've learned is that most companies keep their data accessible for two years, but they will warehouse six to seven years' worth. People want to surf this data the same way they surf the Net. One way to accomplish this is to use Internet-type capabilities to access the data through Explorer-type screens. Even though most large corporations don't want to put their information out on the Internet, they can still use Explorer-type screens to access their own network intranets.

4. Why is using the Internet to access data good for retail employees?

Brad Jarrett: When each department within the retail operation logs on to the network, they can only access reports specific to their job descriptions. In traditional environments, the system slows if several people try to simultaneously request information from a database. If you have an application on each computer, it grabs hold of the database and doesn't let it go until it is done with the query. You might have several people requesting database information in the form of queries and reports. A Web-based approach makes deployment and optimization easier. In order for this way to work the best, retailers should look at a higher level database such as a SQL or Oracle server.

5. Do retailers need an additional software piece to make sense of all the data?

Michael Corinotis: Most POS software programs now have modules that can decipher information. Our POS software did not have a transaction log, which we needed for exception reporting. Installing the transaction log involved writing code on a store by store basis. We had to add an additional server and the LP software as well.

6. How will data help your LP department?

Michael Corinotis: In 365 days, we apprehended 183 dishonest employees from our 79 stores before we had exception reporting. Once the LP software is installed, we hope to identify internal theft patterns at the POS (point of sale) quickly. We also want to identify training issues before they become a shrinkage problem. The quicker we can spot these problems, the easier we can rectify them. We hope to see theft patterns within two weeks to a month. Once we establish the database history, we will really be able to identify telltale patterns.

7. How will you measure the return on your investment?

Michael Corinotis: We recently apprehended an assistant store manager in one of our stores who stole $70,000. It took four months to develop the case and we had to conduct a lot of research. This would have been easily identified and the loss would not have been as significant if we had this program installed a year ago. If we identify a lot of big cases, it will immediately justify its expense and then going forward it will be more of a preventive tool. If we get two cases for $100,000, it has already paid for itself.

8. Do retailers need brand-new systems to run successful data analysis?

Brad Jarrett: No, not anymore. With additional software, retailers can get more out of a lower-end POS (point of sale) than they did out of a higher-end. With additional software, retailers can use even simple cash registers that have been installed for years to capture information electronically at the detail level, mine the data, and use it for reports not available through the POS system itself. And, at the same time, they can add functionality such as loyalty programs or time keeping.