Place your fraud team in the right place within your business
Fraud is a growing problem that touches the financial, operational and technological parts of your company. As such, its actions are never confined to itself alone and impact your customer experience and conversion as well as the bottom line. Address this by considering the interfaces of the team and create effective communication channels between the fraud team and the business.
People before tools
Hiring the right people is key to building a successful team. By focusing on people with analytical skills you can make sure your team figures out the specific fraud patterns that your company is facing. While experience in fraud is helpful, some of the best analysts came from backgrounds ranging from business to biology. The key is their ability to ask questions and answer them methodologically. Balance skills in the team by hiring people with backgrounds across domains, allows team members to learn from each other. If you only hire people with similar skills and experience you are only ever going to get the same solutions.
Fraud methods are widely shared, new fraudsters join the flood every day and new attack vectors are added daily. Training is essential and on-going to keep your team ahead of the game. As new fraudsters learn trade secrets from their forums on the dark web, your analysts can be trained to understand fraudster behavior based on their actions. Build an internal training program that includes fundamentals of payments systems, phone and internet communication and solid overview on how fraudsters interact with your systems. Allow the team to professionally grow by encouraging the team to run routinely internal training sessions, sharing findings across the team and company.
Using the right tools
The tools that businesses used are often selected by trial and error. We see everything from the basic in-built functionality that their payment services provide to Excel sheets, text files and various database queries
While those tools will take you some of the way to tackling fraud, providing the right set of modern tools to the team dramatically improves their efficiency. It reduces the time they spend investigating each case and improving the accuracy by providing decision support information exactly where it’s needed. This translates to less fraud and less customer friction.
By focusing on these principles, you can find the optimal solution for your team. Monitoring suspicious accounts should be done automatically, without a the need to actively search for them. Blocking suspicious activity should be automatically and instantly. Repeated attempts of fraudsters to create more accounts should be blocked without further human involvement, as duplicate accounts will appear and you don’t want to create extra work for the team.
While a modern system function is fully automatic, you want the ability to change business flow based on risk scores. For instance, you may want to require more information or direct to 3DS orders that receive high risk score. You also want to be able create business policies as rules to control which customers are caught by your fraud solution. Often your business requirements change overnight and you don’t want to rely on software to learn what you can easily tell it to do.
Even though you want to leave your analyst free from manually reviewing all orders, you do want to allow them to investigate some fraction of cases to keep the system in check and improve it. You want a system that shows a prioritised list of candidates to investigate, reducing the time it takes to find interesting cases. You also want to enable exploration and search for patterns you see as interesting or high-risk, not limiting your search criteria to the list the vendor has pre-compiled for you.
When an analyst investigates a suspected fraud incident, you want your system to present all of the relevant information in a single page - you don’t want them to manage multiple notepads, excel sheets and various database queries to keep track of the relevant information.After an analyst has spend time reviewing an account, you want to capture their decision so you can rank both their performance and the system’s performance, creating visibility and accountability with your team.
Measuring the right things
When setting the team's goals, it’s important to align their focus with the company’s goal. Measuring their capacity in terms of cases resolved is not sufficient, as it needs to take into account sales conversion, accuracy, time and the level of customer satisfaction. By creating two benchmarks - reviews numbers, reviews correctness - measurable improvements can be quantified and each analyst can see where they need to improve. Analysts can then suggests an improvement to the processes, and having a framework to evaluate the value of such improvements eases the implementation process.
While actual fraud levels may change depending on how hard you make it for the fraudsters as well as their motivation, you always want to stay ahead of them. One way of doing it is by challenging your own system in a process often called `red-teaming`. In this exercise, you a team member will be trying their best to commit fraud and the other team members will be trained to catch it. If does get caught by your system, it’s a great way to reevaluate your solution and find your own weak spots before someone else does.