The impact of retail fraud is enormous, and it's increasing at an alarming rate. Many factors contribute to the surge in criminal activity across the UK and on a global scale:
- The coronavirus pandemic has created global economic pressure.
- Omnichannel customer services encourage tech-savvy criminals to operate across physical and digital boundaries. A new generation of theft is evolving to exploit services such as:
- BOPIS (buy online pickup in-store)
- BORIS (buy online return in-store)
- Self-checkout requires less staff and reduces in-store costs. It also creates new opportunities for determined criminals.
- The increased use of mobile devices and contactless payments expose security loopholes.
- Generous policies encourage people to return goods after using them. Wardrobing is common - where a customer wears an item before returning it for a full refund.1
This article focuses on bricks and mortar fraud prevention. For modern shoppers, a seamless omnichannel shopping experience is essential. As the number of customer touchpoints increase — so does the opportunity for fraud. In the real world, prevention must reach across online services as well as physical stores.
When targeting criminal activity, a little effort can go a long way. Prevention comes in many forms, including staff training, physical security and anti-tampering devices. Here we examine how data can provide additional insights in the fight against retail crime.
The Scale of the Problem
The figures are staggering; reports claim five billion pounds of shrinkage2 across the UK. Retail fraud accounts for global losses of $100 billion. The following table illustrates the extent of the problem.
On average, criminals are only caught once out of every 48 times they steal. Less than half of associated arrests lead to prosecution. The situation is worse for chronic offenders with arrests for only one in every 100 offences.
Serial offenders are creatures of habit exploiting specific products and scenarios. Retailers often refuse to sell goods that attract criminal attention.
When Thieves Benefit, Everyone Else Suffers
It’s not only retailers and business owners who suffer. Lost revenue affects investments in other areas:
- Policing premises comes at a cost and impacts the customer experience.
- There is less investment available to improve terms and conditions for staff.
- Competitive pricing and discounting become harder to put in place.
- Often there is no budget for improving store layouts and customer services.
Conviction rates for arrested people are low. As a result, many retailers are reluctant to report criminal activity. Additionally, the risk of abusive and violent behaviour increases when challenging customers.
False Positives and Negative Impacts
False positives are a big problem — where someone accuses an honest customer of stealing. If you get it wrong, it's likely you've lost a customer for good. Lost business leads to lower revenue.
The potential for criminal activity is significant. As customer services improve, thieves are finding more ways to cheat. It is not possible to watch everyone all the time. CCTV acts as a deterrent, but shoplifting often goes unnoticed without continual monitoring. Video footage alone is not effective against all forms of retail fraud.
Staff often assume that unscanned goods are a mistake. This approach avoids false accusations. Once notified, customers can scan and pay for the item in good faith. It does, of course, leave offenders to try the same con another day. After all, what's the worst that will happen?
Old Tricks and New Technologies
Criminals are finding new ways to exploit technology. Organised gangs add scale to the problem. The number of retail staff who steal from their employers is a great cause for concern. Customers and staff are stealing using methods that are difficult to detect.
Retailers are getting hit from all directions. The diagram shows some of the activities taking place on high streets and in shopping centres. Data can help us to address many of these criminal activities.
It's Often an Inside Job
Estimates attribute 22% of shrinkage to employee fraud. Here are some things that may help to surface underhand activities:
- GIFT CARDS: Track the number of gift cards issued by specific outlets at various times. Use this to set a baseline to compare against future activities. Mark sales that are much higher than the baseline for future investigation.
- REFUNDS: Set KPIs for acceptable levels of product refunds — review excessive activity. Investigate when staff perform refunds without a customer.
- PRICE OVERRIDES: Look at specific cashier-customer combinations where there are many price reductions. These alterations are often a sign of sweethearting. We examine this in more detail in the following section.
- POST VOIDS: Check the times between voiding transactions and re-entering new prices. A cashier may void a sale only after the customer has left to avoid ringing up a receipt.
- THE BANANA TRICK: Look for many purchases of the same item in a transaction. For example, scanning single bananas rather than a bunch. Multiple line items for the same product can indicate theft.
Data analysis can help with each of these examples, either by itself or in combination with other measures such as CCTV, security guards, RFID.
A Simple Example Using Real Data
Criminal activity often involves employees and customers working together. In this example, we look for a high number of price reductions between the same cashier and customer. If the number of reductions is excessive, security staff can review video footage when these transactions occur. The following diagram illustrates the process.
A Summary of the Data
The following insights come from a set of real-time, anonymised retail data. The data consists of more than 280,000 transaction records collected in 2017. The information represents 19 days of trading with the following details:
- Records exist for four different outlets.
- Transactions cover 112,214 unique customers and 496 different cashiers.
- There are 5000 price alterations and 22,000 voided line items.
- The total value of sales is approximately £18.5 million.
The Extent of Price Reductions
The data analysis looks for price alterations involving the same cashier and customer. On average, each cashier reduced one item per customer during the recorded period. Looking at any cashier who reduced a dozen items or more for the same customer may be worthwhile.
The table shows cashier and customer combinations where the number of price alterations is greater than or equal to 12. Over time, the optimal number of alterations to observe will change.
The diagram is a graphical representation of the data. Cashier 1525-9993 altered pricing 23 times for one customer alone.
The data includes the times for each transaction. Security staff can fast forward to find appropriate footage. Further analysis can show activities based on store, product costs or times of the day. In this way, retailers can track behaviours from many different perspectives.
This article describes common types of retail fraud and how data can help to detect it. The analysis complements physical security and needs no changes to what already exists.
The example reduces the impact of false accusations and the problems this creates. Criminals will always find ways to exploit honest traders — it's a never-ending battle. While this is true, there has never been a better time to start building an effective defence.
Our data team has helped many of our clients extract actual business value from their data. If you’re looking for ways to improve your retail business, learn more at our Data in Retail page.