Data-Driven Defense: Transforming Loss Prevention Using Analytics

By Allie Falk

Technology is evolving at a rate that can be overwhelming. Extraordinary concepts are developed daily, with modern technologies rapidly advancing and trends constantly fluctuating based on the needs of businesses and the demands of consumers.

Every day, there are stories in the media about developments in technology, such as artificial intelligence (AI). Even data analytics, which have been around for a considerably longer amount of time compared to AI, are changing and improving at faster speeds than imaginable a few years ago. Those of us who are less familiar with data analytics may be asking, ‘What really are analytics anyway?’ To be precise, analytics are the systematic computational analyses of data or statistics. Analytics are applied with the aid of specialized systems and software.

Inevitably, concerns arise due to the rapid advances in technology like analytics. How do I measure its accuracy? How do I maintain data privacy? How do I implement new strategies based on data extracted from analytics? Will analytics cut employee headcount? Moreover, as findings emerge through the consolidation and cross-analysis of data, LP professionals will discover both advantages and disadvantages throughout the process.

To offer you a deeper look at these issues, we reached out to a group of industry-leading solution providers to get their take on this evolution. We asked them to talk about both the positive and negative ways that analytics are changing the world of loss prevention. Here is our summary of the topics discussed along with their valuable and varied points of view on each.

Product Advancement and Shifting Demand

Inevitably, as analytics evolve, products will continue advancing at a rapid rate to fulfill consumer demands. In response, there is a need for solution providers to remain on the cutting edge to ensure their offerings continue to evolve to meet those demands.

“AI and analytics have become more efficient and straightforward; they are speeding up the customer experience,” said Axis Communications Segment Development Manager James Stark. “As analytics have progressed, I think the biggest advancements in our products revolve around our chipsets, enabling intelligent data to be processed at the edge—or far edge of cameras. The camera is only transferring back critical pieces of information that you’re looking for, also known as metadata in this case. It also enables our partners who are in the AI space to reside on the edge with us, getting that information back for the video. At the end of the day, a good customer experience is what it’s all about.”

Loss prevention practitioners also want easy-to-use and accurate tools. LP teams have expectations that their weapons to fight crime will be as straightforward as the other digital devices in their lives. Data analytics should be no exception.

“We can now control bandwidth to where people can pick up their cell phone using a 4G or 5G connection and look at live recorded CCTV systems without a bandwidth issue for their company,” said Keith Aubele, chief security officer at Salient Systems. “Being able to assess the analytics while simultaneously viewing video footage without a bandwidth restriction is phenomenal, and this progress in technology is only going to continue.”

Multi-purpose solutions are in high demand. Integrating analytics to create solutions that accommodate this provides a more efficient experience, and supplies retailers with the functions they need to prevent loss, such as clearer store communication.

“Multi-use and multi-purpose solutions are what consumers are after,” according to Robb Northrupp, director of marketing and communications at siffron. “As technology has evolved, we have added a device that can transmit wirelessly. It can then communicate to retailers’ systems, whether that’s the electronic article surveillance system, cameras, or any sort of store communications. By analyzing those triggers and aggregating results with analytical data, we will only continue to see this sort of intelligence and make a shift on the LP side of things.”

Vendor management system (VMS) platforms are a good example of a multi-use solution.

“If you want to add a facial recognition analytic, if you want to add a bottom‑of‑the‑basket or cart detection analytic, or if you want to add a shopping cart analytic, you can do that all in one platform, and most of the time it costs you absolutely nothing,” said Aubele. “Just as important as having data analytics available is the opportunity for retailers to digest shared data and insights to improve customer service.”

“Nowadays, the ability to review and act on data across stores, regions, and channels has increased,” reveals Pedro Ramos, chief revenue officer at Appriss Retail. “As delivery options have grown, retailers have an increased need to push recommendations to the customer service desk to counter ‘item- not- received’ and ‘did- not- arrive’ claims. Finally, having a streamlined way to present insights—not just data to different audiences, has also been driving more conversations about standard libraries, reports, and dashboards.”

Staying current with the rapid change of pace in technology, embracing the capabilities it holds, and flexing to use this critical tool in other areas of the business are all crucial.

“Our platform has evolved significantly with the advancements in analytics, embracing the power of AI and machine learning,” Ramos continues. “These innovations have enabled our platform to offer retailers even more precise and real-time insights into their operations. AI is used to detect and flag anomalies both within and across the stores, letting loss prevention experts work more quickly to investigate. We have also moved to offer solutions in other aspects of the retail business, including claims and appeasement processes within the customer service department.”

Customer Satisfaction Starts with Safety

The bottom line for every retailer is keeping their customers and employees safe. Identifying threats sooner makes it possible to prevent incidents from occurring altogether. Facilitating quicker emergency response is also a significant advantage and can save lives. As this technology continues to improve, more retailers are investing time and money in analytics to ensure a safer shopping environment for everyone.

“Whenever a retail organization is looking at deploying defenses and being safe and secure when ORC comes into play, understanding the concept of the positive customer and the negative customer is crucial,” Aubele stresses. “You have two types of people who walk in the store: a positive customer with the intent to purchase items and to pay for those items, and a negative customer with the pure intent of shoplifting or executing their ORC scheme. You must put together programs that will impact both the positive and the negative customer.”

For example, if there is information about a known ORC offender in the area, the store can insert data into the camera or surveillance system— including specific characteristics—thus creating greater potential to identify that criminal. As a result, the positive customer feels safer, while the negative customer is more easily detected and can either be apprehended or deterred.

“The ability to apply forensic analytics is so crucial in moments of violence. You don’t want employees or your customers getting hurt trying to intervene and take these people on because the threat of violence is such a risk,” added Stark. “It’s becoming very prominent, and I think there are a lot of retailers out there looking for solutions that help train staff to react appropriately or can forensically put a case together and show law enforcement, ‘Hey, this is the group of individuals and here’s the information we have.’”

Identifying ORC can be a complicated, difficult process. With the help of data analytics, obtaining the information needed to identify offenders not only becomes easier, it can also turn into solid evidence.

“There’s no silver bullet with ORC because it looks very different,” Stark reveals. “But AI and analytics have gotten to a point where they’re providing us with the data, we need to ensure we can act safely and accordingly to each scenario. For example, say you have surveillance footage of someone in a brown top and a blue hat who stole from your store. You can then tell your analytics search, ‘Show me everyone with a brown shirt and a blue hat’ and it’ll supply me with all video footage of a person wearing a brown shirt and a blue hat. Analytics collects all that data and presents it in a much more straightforward way to the end user so they can then build a case.”

Customers want to shop in safe environments where they can find the items they want; analytics contribute to the overall customer satisfaction experience.

“When used to reduce shrink and fight organized retail crime, analytics help ensure in-stock inventory accuracy, which makes shopping easier for consumers,” said Ramos. “Advanced analytics play a role in identifying emerging external crime patterns, allowing retailers to allocate resources more effectively and mitigate the impact on both shoppers and employees. Retailers can also use advanced technology, powered by predictive algorithms and statistical models, to identify and deter return fraud and abuse without isolating consumers making a good-faith return.”

Measuring Data Accuracy

There are various methods of measuring analytics’ accuracy, but no analytics system is perfect and there will always be some degree of uncertainty. Through regular evaluations and iterations, analytics systems will continue to improve in accuracy and reliability.

“Measuring the accuracy of data analytics and predictive models comes down to continually analyzing a diverse range of data sources and comparing it against actual incidents of theft or fraud,” shared Ramos. “Continuous monitoring and validation against real-world outcomes, as well as incorporating feedback from loss prevention experts, contribute to an ongoing assessment of accuracy and the refinement of analytics strategies.”

Stark recommends a “proof- of- concept” period with one selected store or small group of stores.

“Once the analytics are live and calibrated, it is a matter of monitoring the process to ensure desired results,” he stated. “Much like other programs, once you have your baseline and success criteria established, you can then test functionality and build out reporting to ensure the analytic is doing what you have been told it will do, and if you are getting the desired results.”

Accurately measuring the value of data analytics is critical. Here, Auebele breaks down the evaluation process:

• Inventory Accuracy: Retail analytics often rely on inventory data to generate insights. Comparing the reported inventory levels from analytics with actual physical counts can help identify discrepancies and assess accuracy. Retailers typically conduct inventory once a year in the hardlines area (grocery and cost departments are monthly). Sometimes, segmenting off an area (say, over- the- counter areas of pharmacy) and conducting specialized inventories to validate impact, e.g. after a thirty-day run, is ideal.
• Forecasting Accuracy: If your retail analytics system includes forecasting capabilities, measure the accuracy of its predictions. Compare the forecasted sales, demand, or inventory levels against actual results to gauge the system’s performance.
• Comparison with External Data: Validate the analytics results by comparing them with external data sources or industry benchmarks. This can help identify potential biases or errors in the analytics. Additionally, lean on research groups like the LPRC at the University of Florida to enlist their help or guidance in real-world analysis of specific analytics.
• Return on Investment (ROI) of Analytics: Measure the impact of the retail analytics insights on business outcomes. If the insights lead to improved sales, reduced costs, or other positive effects, it indicates the analytics are providing valuable and accurate information.

Maintaining Data Privacy

Compromised personal information can lead to multiple forms of fraud, such as cyberattacks, identity theft, and financial fraud against your customers. It can damage a company’s reputation and undermine consumer trust and loyalty. Not surprisingly, data privacy remains a concern when evaluating data analytics tools.

Aubele offers a list of best practices to ensure data privacy while using data analytics:

• Specific Data Usage: Only collect and store the data necessary for the retail data analytics purpose.
• Clean the Data: Anonymize customer data before using it for analysis.
• Encryption: Ensure that all data collected and stored is encrypted through its life cycle, to protect it from unauthorized access.
• Access: Implement strict control to limit who can access the data.
• Audit Trails and Regular Audits: Maintain detailed audit trails of data access and utilization to track any misuse or unauthorized access. Frequently audit processes and policies to ensure compliance.
• Secure Storage: Use only the highest-level secure data storage solutions with robust security measures and redundancies in place, such as firewalls, intrusion detection systems, and regular security updates.
• Data Retention Policy: Establish a clear data retention policy that outlines how long data will be stored and when it should be securely disposed of.
• Training: Train employees on data privacy best practices and the importance of protecting customer information.
• Third-Party Vendors: If your third- party vendors are utilized for retail analytics, ensure they have strong data privacy policies and adhere to relevant regulations. While collecting and retaining consumer data may be an essential practice to identify suspicious patterns and stop repeat offenders, it has its risks.

Ramos cautions, “To collect and utilize data, there must be a balance, including adhering to relevant privacy laws and ensuring consumer trust.”

Stark added, “I strongly recommend partnering with your IT security teams and ensuring your company’s data privacy standards are met prior to any technology deployment involving data governance.”

Making Informed Decisions

Informed decision-making is crucial for the success of any retailer. Insufficient data can lead to decisions being made based on assumptions alone. Having accurate and complete information provides a more thorough understanding of the context and decisions are made based on facts, rather than emotions or assumptions, which inevitably saves time and money.

“Predictive analytics can help retailers anticipate potential loss events and make informed decisions on how to reduce shrink and improve profitability,” said Ramos. “In recent years, data analytics has improved by using advanced AI and predictive algorithms to analyze transactions in-store and online. This data is vital for retailers to identify potential theft or fraudulent behavior in real-time, enhancing their ability to respond swiftly to threats such as returns fraud or internal theft.”

Analytics have created a world in which decision-making has become much more deliberate. Being able to see data in real-time and analyze past performance enables retailers and employees to also significantly shift the way they approach safety measures, shopping behavior, the customer experience, and more. Employees with real-time information can prevent loss and increase sales.

“Analytics offer growth potential, providing the means to drive data and in turn, support strong decision-making,” Northrup added. “This can happen both at the corporate level and on the floor for the everyday employees. They are now better equipped to understand what’s going on in the store each and every day.”

“Let’s say a theft incident occurs, and the perpetrator heads to another store in the same market,” shared Stark. “With License Plate Recognition (LPR), the car license plate is loaded into the system, and staff would immediately be alerted as the car pulls into the parking lot. This type of technological advancement prompts the staff to engage and be more aware of what’s going on, improve store safety, and detect and deter any potential illicit activity.”

When choosing solutions, decisions should never be made in a silo or based on a single criterion. When evaluating data analytics solutions, determining which solution to choose must be based on the overall value of the product and how it will be used across the enterprise.

“Companies must consider a number of different factors when investing in a data analytics solution,” Aubele said. “First, you must consider the return on your investment. One potentially negative aspect is that analytics solutions do not come cheap, and there are several different options to choose from. Additionally, there are fractional analysts needed to gather and interpret the information. As a result, cuts may need to be made somewhere else to compensate for the added expense. Ideally, we find solutions that balance our needs and the cost of the solution with minimal impact on other resources, such as headcount, resulting in lower shrink and less risk.”

An Optimistic Future

As data analytics aids in simplifying operational functions, performance will continue to improve, and most importantly, it will create safer working and shopping environments.

“The outlook for loss prevention is promising,” said Ramos. “With the integration of AI, machine learning, and real-time data analysis, the field of loss prevention is poised to become increasingly proactive, precise, and simple for retailers to implement. Through actionable insights from data, retailers will see improved efficiencies. The continued advancement of analytics in loss prevention holds the potential to revolutionize the way retailers safeguard their operations and assets.”

source : https://losspreventionmedia.com/