Online fraud is a major concern for ecommerce businesses. Nearly 50% of retail loss prevention professionals are getting bigger budgets to help quell fraud, according to a report released by the National Retail Federation in June. In the race to prioritize fraud prevention, ecommerce leaders can miss the resources they need to build an anti-fraud team.
Signifyd Director of Risk Consulting Tim Davis has some advice on how to make fraud protection work for any ecommerce business, by maximizing your resources on hand and making wise investments to build new resources from scratch.
Davis’ advice covers four main areas on how to build a strong ecommerce anti-fraud and fraud detection team:
- Hire and train dedicated fraud staff
- Maximize your people resources
- Outsource what you can to machine learning
- Grow your team wisely as revenue increases
First, refine your understanding of the basic do’s and dont’s of ecommerce fraud protection operations. Then, learn from the pain points of failed fraud strategy. Start your fraud operations overhaul on the right foot.
Hire and train dedicated fraud staff
A common solution for including a fraud team in ecommerce operations is pushing fraud review duties onto an existing department, usually the customer service team. This is a bad move for a few reasons. The first being adding extra work to a team already likely burdened with a heavy workload. The second, and more difficult to reconcile, is how fraudulent transaction review runs counter to everything a customer service team stands for.
Davis points out the customer service team’s conflict of interest when it comes to fraud. Customer service staff are the ones talking to consumers every day. They’re trained to help the customer no matter what, which conflicts with making good fraud decisions.
“Fraud decisions will always upset someone,” Davis says. “You have to decline transactions or make decisions a customer might not understand. It’s easier to do that if you’re not invested in trying to make the customer happy.”
The value that your CS team delivers is far different from that of the fraud protection team. “Good fraud teams require some specialized knowledge,” Davis says. “Customer support teams probably don’t have that kind of knowledge.”
The specialized functions and knowledge required of each team aligns with the second point in Davis’ advice for building a strong fraud operations team. Let CS do what they do best, and turn your investments of time, money and training toward maximizing the human side of your dedicated fraud prevention team.
Maximize your people resources
As ecommerce businesses seek ways to get smarter on their merchant chargeback strategy, they face challenges when hiring the right team. A small talent pool and unpredictable fraud decision making strategies pose significant hurdles for retail leaders looking to build a fraud operations team from scratch.
Davis advises that it starts from the top, with hiring experienced leadership for the team. It’s challenging to find fraud professionals for any position, especially in areas without a strong ecommerce business network. That means it’s best to target experienced fraud people to lead your team.
“It’s not always possible to hire experienced people,” David says. “The person leading the team needs to have knowledge to share with the rest of the team. Hire smart people who can learn quickly, and lean on the experienced people you have.”
It’s also best to budget training resources for the right team balance. Davis proposes two ideas for fleshing out a fraud operations team: training a few people to gain a high level of fraud and risk management knowledge, or training your entire team at a minimum level. “You’re better off training the smaller team to a higher level of proficiency than trying to get everybody in customer support to also be able to do fraud work,” he says.
The combination of experienced leadership and smart people will help your fraud operations team take on another important anti-fraud piece: the human side of fraud protection, which sometimes includes manual review.
Davis has guided teams through some of the most challenging problems in fraud operations. One of the challenges is how to get the fraud team to make the right decisions for the organization without preventing them from using their judgment.
“A lot of the time your staff doesn’t feel empowered to make decisions because their leadership doesn’t agree with the choices they make on fraud cases,” Davis says. “When leaders rely on strict guidelines, you get a cycle of increasingly convoluted rules that prevent your staff from making judgment calls. It’s important to allow your fraud staff to use their judgment and empower them to make decisions on their own.”
Davis’ advice for empowering fraud staff doesn’t mean a hands-off approach. He asks fraud leaders to coach their staff while giving them room to fail and learn from their mistakes. Especially when a good decision can still be the wrong choice.
“Anyone can look at an order, see all the boxes are checked, think the order looks good, and then your company still gets a chargeback,” he says. “It’s not about whether the decision was right or wrong, but if the fraud agent made a good decision based on what they knew at the time.”
In a world driven by uncertainty, Davis relies on collaborative learning and communication within his team to arrive at the best possible fraud decisions. He also asks his team to be more cognizant of the fraud ecosystem, which includes talking to others in the industry and understanding how machine learning supplements humans’ best efforts to provide chargeback protection and fraud prevention for ecommerce businesses.
Outsource what you can to machine learning
It’s an unfortunate truth that ecommerce leaders often can’t address fraud on their own. The problem is too complex for one team to tackle. Davis laid out how the limited fraud operations talent pool hurts organizations looking to build their own team. This is why machine learning-based ecommerce fraud prevention tools are becoming more popular.
“With all the complexities of fraud protection, it helps to work with a solution that provides fraud review so you don’t have to worry about it,” Davis says. “Companies using machine learning can fill in the gaps that would be covered by hiring an experienced fraud team.”
Machine learning-based tools help an anti-fraud team save time and money because the artificial intelligence (AI) can quickly make a lot of easy decisions. As Davis points out, this covers most of the fraud decisions that come up. Humans are better at identifying fraud cases that require tougher decisions: the borderline situations. Machine learning improves the anti-fraud process not by eliminating fraud, but by freeing up human teams to look at the tougher cases that a computer can’t effectively address—another success factor for Davis’ teams.
“This combination of humans working with machine learning will always be an important input,” he says. “The fraud team must provide the information the machine needs and find the things that could augment the machine’s current decision making process.”
It’s also important to retain the human gaze to go beyond what metrics can give. “It’s easy to hold your team accountable for a production metric,” Davis says. “When you decline a transaction, you lose the opportunity for a merchant chargeback. You miss your chance to get feedback. It’s impossible to know if the decline was a bad decision.”
Our understanding of ecommerce fraud protection is growing at a rapid pace. The last piece of advice from Davis is how to wisely grow a fraud team as revenue increases.
Grow your team wisely as revenue increases
Retail is a largely seasonal business. While most buying activity spikes during the winter holidays, ecommerce businesses can see a buying rush any time of the year depending on their product category. These order spikes often drive growth as well. Too many orders is a great problem to have—until it becomes clear that your fraud team isn’t staffed up enough to handle the demands of more orders coming in faster than ever.
Davis recognizes this as a main challenge for fraud operations teams. Just as quickly as sales heat up, activity on your ecommerce website can cool off. It’s tricky to predict how fraud protection staffing demands will handle the ebbs and flows.
“The holidays are always a stressful time for a fraud team because your employees don’t want to show up for work on Thanksgiving,” Davis says. “But you know your transactions will double over that entire weekend.”
You may think it’s the right time to hire more people. But Davis cautions to think twice about adding staff to meet seasonal demand. “It’s hard enough finding experienced fraud analysts,” he says. “It’s even harder to find experienced fraud people who are willing to work for a couple of weeks or months as a seasonal employee.”
Another time to be cautious about hiring is when revenue steadily grows. When a company makes more money, it makes sense to expand, right? This is not always a best practice for hiring in fraud prevention.
“It shouldn’t be a direct scale,” Davis says of growing a fraud team as revenue increases. “If you have 50 people today and $15 million in revenue with the goal to double revenue, don’t double the ops team. The goal is to increase your margin, not keep it the same.”
This is where a smart anti-fraud hiring and growth strategy makes the most sense for ecommerce leaders. Davis says the best way to plan headcount for smart growth is to be more efficient about making fraud decisions.
“Handing off a larger portion of your decisions to a machine will come naturally as your business grows,” he says. “You’ll become more confident and knowledgeable as you come to understand the gray areas. That’s where outsourcing to machine learning becomes valuable for a growth strategy.”
Build the right fraud team for your business
We’re learning more about ecommerce fraud protection every day. As we enrich our knowledge and partner with technology, ecommerce fraud becomes a bit easier to understand while the solutions get closer within reach. The days of living in fear of fraudulent transactions are over. Building a better fraud team is possible through a combination of experienced leadership, open-minded professionals ready to learn and AI that does the heavy lifting that humans can’t. Now that you have perspective from one of Signifyd’s anti-fraud leaders, you can take ownership of your company’s ecommerce fraud strategy.