Customer experience is more important than price. American customers have indicated in a survey that their experience influences 75% of their decision to purchase1. I believe eCommerce merchants who can deliver those experiences gain market share. As merchants have created defenses, fraudsters have increased the intensity and sophistication of attacks. This calls for a layered data insight approach to prevent, detect or mitigate fraud while maintaining a good customer experience.

Today’s customers have high expectations. To meet those expectations, brands need to prioritize convenience when designing customer interactions. Convenience should not come at the expense of accepting more fraud because fraud is expensive. The LexisNexis® Risk Solutions 2019 True Cost of Fraud Study Retail Edition estimates the total amount of losses a merchant incurs per $1 of fraudulent transaction is as much as $3.13. The cost of fraud has increased by 6.5% since 2018’s study.2

Based on internal analysis done by LexisNexis Risk Solutions in collaboration with merchants, we see that good customer experience can improve the sales conversion rate, and in turn, bring more revenue. Friction caused by fraud prevention can take away from customer experience, and negatively affect revenue. Maintaining a balance between the two is crucial. Making better decisions on what is fraudulent and what isn’t is critical to the customer experience. Customers don’t want to be treated like criminals. To make more accurate decisions, merchants should follow a layered approach which is focused on contextual data, machine learning insights and a passive form of behavioral biometrics.

Make data your secret weapon: not just more data but the right data.

Most merchants have two big challenges: either they do not have the right data or do not have the data in a usable format. Evaluate the data you have and collect the right data. Once you have the right data, work with vendors or other merchants with the right expertise to enhance the data and provide relevant insights. One way to collect more relevant data is by participating in consortium models and collaborating with trusted partners. Incorporating additional data can help you complete a 360-degree view of your customer’s identity and provide a better customer experience. Your new customer could be someone else’s best customer. This approach also helps to pro-actively block fraudsters who have already committed fraud on other websites.

Leverage machine-powered insights.

There is a need to combine human discernment and scalability of machines to eliminate human bias and create scalable repeatable processes. These machine learning models work great when we provide feedback into the models – the output is only as good as the input. If you have history on users’ behavior, the model can detect patterns that have been seen before. But what happens to zero-day patterns? There are two possibilities, your zero-day patterns have been seen elsewhere at other merchants or they are completely new. Leveraging consortium-based machine learning models can reduce the uncertainties associated with behavior that may be new to you but has been seen before globally. Anomaly detection models can be insightful in detecting zero-day new behavior or attacks, but it will always be challenging to detect new patterns no matter how good the anomaly detection tool is. Therefore, using consortium-based machine learning, will make the most of local and global shared data to reduce uncertainties associated with zero-day patterns.

Unleash The Power of Passive Behavioral Biometrics

Behavioral Biometrics is the field of study related to the measure of uniquely identifying and measurable patterns in human activity. This is in contrast to physical biometrics, which involves innate human characteristics such as fingerprints or iris patterns.3 Behavioral biometric tools can help build trust relating to good customers by building strong user scores over time that increases confidence in specific good behavior. This can reduce false positives by modeling behavior on a per user basis. It’s important to note that looking at the entire customer journey will help better differentiate suspicious and good behavior and result in more accuracy. Combining the way a user interacts with their device and information relating to the trustworthiness, integrity and authenticity of that device can form a compelling way to help detect high-risk scenarios more accurately.

To find your data insight gaps and to ultimately improve your customers’ experience, ask yourself the following questions:

  • How actionable is the data I am currently receiving?
  • How do I benchmark my fraud prevention compared to the industry?
  • Are my other actions improving my customer experience?
  • What data am I missing that could be used to identify anomalies or good customers?