- 3.5% more orders than in the previous year. Average order value up by 10% to €275.
- Promotional week attracts young customers in particular: average age of buyers 1.5 years younger than the annual average.
- Lower fraud rate due to more regular orders. Higher fraud rate expected at Christmas.
Black Friday is considered the peak period for e-commerce, when retailers try to outdo each other with high discounts, and ever-increasing numbers of consumers place orders. According to an analysis by information service provider CRIF, online retailers recorded almost twice as many orders in 2024 than the annual average (198%). Compared to Black Friday 2023, the order volume recorded by CRIF rose by a further 3.5%. In addition, the average order value rose by 10% to €275 (previously €250). Young people in particular took advantage of the promotional campaign: the average age during Black Friday week was 1.5 years younger than in the rest of the year.
Lower fraud rate despite increase in orders
For fraudsters, however, Black Friday in itself is not a reason to intensify their activities: It is true that every 133rd order (0.75%) is a potential fraud attempt. However, compared to the previous weeks, this is actually a decrease of 19.5%. On average over the year, almost every 100th order (0.93%) is a potential fraud attempt. "The price of goods is not important to fraudsters, but it is for legitimate buyers. Therefore, promotional events such as Black Friday increasingly motivate genuine customers to make purchases, reducing the ratio of fraud attempts to regular orders," explained Dr Frank Schlein, Managing Director of CRIF Germany. "The situation will be different during the week of Christmas, when fewer regular orders are received while fraudsters remain active. This will increase the fraud rate," according to Schlein.
Typical fraud methods and protective measures
Fraudsters in e-commerce use a variety of scams and present companies with considerable challenges. One frequently used method is identity theft, where fraudsters misuse real identities to place orders. In addition, there are also frequent fraudulent purchases where orders are placed despite the clear inability to pay. Equally widespread is the creation of fake identities, where false customer profiles are constructed to conceal fraudulent transactions. Another problem is returns fraud, where return policies are manipulated or fake receipts used to illegally obtain refunds or replacements. "These diverse fraud techniques show how important it is to implement targeted prevention and detection measures," emphasized Schlein. "Retailers can only protect themselves through comprehensive risk assessment, by combining internal and external data with intelligent modelling. The targeted use of machine learning technologies and manual control procedures offers the most effective protection against fraudsters."
Tips for retailers and consumers
Both retailers and consumers can recognize and avoid fraud attempts at an early stage by paying attention and taking targeted measures. Retailers should be particularly vigilant in relation to conspicuous order patterns. These include unusually high order values, which are often favored by fraudsters, as well as recurring email domains that occur more frequently in certain regions. By recognizing such patterns early on and consistently checking suspicious transactions, retailers can minimize the potential damage. Consumers should watch out for unexpected order confirmations, charges for digital goods or invoices for purchases they have not made. "Such signs can indicate the misuse of personal data or identities," noted Schlein.
Artificial intelligence is both an opportunity and a risk
AI plays a central role in fraud prevention. CRIF uses data-driven machine learning models to identify fraudulent networks and identity fraud. But AI can also be misused: It facilitates the creation of deceptively genuine fake stores or manipulated product reviews. CRIF relies on flexible platforms and an excellent database. Thanks to the adaptability of its AI models and the use of tried-and-tested tools, the company is able to recognize suspected fraudulent orders with a high degree of accuracy, and adjust the models to new fraud patterns. ML (machine learning) models also play a central role. These models use complex decision-making logic to identify fraudulent orders even more precisely and distinguish them from legitimate orders. Their ability to recognize patterns in large amounts of data and continuously learn from new information enables a more effective and more accurate prediction of fraud attempts.
