Sponsored by Appriss Retail
The retail industry is on the cusp of a transformation in its use of data to protect profits and to identify and overcome sales barriers. We’ve been here before. First, with data mining, then with exception-based reporting (EBR). Now, it’s leveraging machine learning and AI models to generate more powerful insights to identify and address problems as they emerge—rather than waiting for them to be reflected in disappointing year-end figures.
As director of data sciences at Appriss Retail, Renee DeWolf, MBA, CFE, is uniquely positioned to understand how retailers will be using data analysis in the years ahead. She sees hybridization of today’s robust exception reporting toolkits with statistical model based insights as the industry’s next great leap forward. Widespread adoption is inevitable, as it provides a reliable, economical way to fix systemic problems, reduce shrink, and ultimately improve profits.
“This represents another transformation in the extent to which analytics can drive the LP function,” she said. “It’s another example of how we can expand on the nature of the LP department itself; a solution like this, with the nature of the additional data, can be both a loss prevention and a sales enhancement tool.”
Sales reducing activities (SRAs), for example, are a part of doing business. They exist to improve customer service and to correct errors that happen from time to time during sales transactions. At the point of sale (POS), these include voids, price modifies, coupons, tax overrides, refunds, and manual entries—and it is all data that is ripe for analysis as it has always been, but can now be complemented by inventory, e-commerce and additional SRA categories from across the business, according to DeWolf.
“The shrink equation has loss in the numerator and sales in the denominator. Traditionally, retail Loss Prevention focused only on correcting the loss,” she explained. “Our goal is to use technology to drive both sides of the equation holistically, producing much better results.”
An increase or change in SRA frequency doesn’t necessarily indicate fraudulent activity—but it may. By analyzing SRAs in detail and combining them with additional risk variables, a retail organization can identify and shut down fraud schemes. LP leaders have done this very thing successfully for years via EBR systems, using their deep knowledge of the industry to establish notification criteria for when certain thresholds are met. It’s this success that next-generation solutions can build upon with machine-learning analytics. “It enhances the economy of scale of these manual interventions and puts additional computing power behind LP leaders’ deep understanding of fraud indicators,” said DeWolf. “It adds robustness to that effort.”
Tracking and analysis of SRAs can also protect profits when fraud is not the cause of a variation, by revealing weaknesses or problems that can be caused by any number of issues, from inadequately trained store associates, to confusion over a new coupon program, or even a software bug. The core of SRA exception monitoring has historically been within POS, but in the future, we will go beyond it—and include data on measurables such as customer satisfaction. It’s all worth capturing, suggests DeWolf, because it’s all data that correlates—eventually—with future sales figures.
“You can include data that is not generally within the LP function, such as customer service or inventory data and merchandise availability, and bring that into the application,” she explained. “It’s no longer simply about exception-based reporting, it’s about total retail performance—to minimize losses, drive sales and margin, and identify data that are useful predictors of positive or negative shrink results. There are lots of components to retail performance, and it’s beneficial to combine that data into one place and to react to it on a daily or weekly basis.”
Employees across the retail organization can tap into and extract intelligence from the data, from LP to operations to marketing, store-level to corporate. An LP professional, for example, will make better use of the time spent on a store audit by running an SRA report in advance. Store-level personnel might resolve the root cause of problematic SRA activity by identifying the need for and then conducting retraining for specific sales associates. Regional and corporate-level personnel can find and address broader issues and causes by using analyses for the whole organization.
Importantly, such data sharing and collaboration is becoming seen as foundational to enhancing retail performance, according to DeWolf. “It’s something that is even stronger in Europe, this focus on total loss performance across the organization” To date, however, retailers’ efforts to achieve this important goal has been hampered in part using tools that were not specifically designed to do the job. This has now changed, said DeWolf, and why we’re poised for an exciting new era for the use of data in retail.
For a retail organization there is a compelling case for expanding data analysis efforts. For LP, the value includes getting a reliable upstream indicator of shrink. Not only will these models for SRAs indicate where shrink may have taken place, but they can provide an early indicator of future shrink. “Shrink is essentially a lagging variable that you only really see once or twice a year, but there are many data components that are not lagging,” said DeWolf.
“We’re now able to add additional power to risk profiles, and from a LP director’s position they can see, ‘What are the things that my field team can triage?’ and also get early indicators of how a new store manager, or store format, of some other change is impacting shrink.”
To extract maximum value from the data they capture, LP, operations, and other functions across the retail enterprise should not wait for results to roll in. “You want to think about what data you can view today to suggest how your results might post down the road,” explained DeWolf.
