top of page

Who Will be Displaced by Amazon’s New Fraud Detection Service?

Fraud solution providers need to be prepared to competitively position and message against the new Amazon Fraud Detector risk modeling services. Here’s what risk vendors and sales professionals need to know about Amazon’s new risk management service offering and how to prepare for prospects or current clients considering this new service.

So many industry incumbents have been displaced by Amazon that business leaders and investors have a term for this competitive threat: it’s called being “Amazon-ed.” Are fraud prevention vendors next? Amazon is now offering a fraud prevention service of its own, called Amazon Fraud Detector, which uses machine learning-based risk modeling to detect payment and identity fraud as well as loyalty and promotional program abuse. The base models and machine learning techniques are derived from the company’s own experience managing online and mobile commerce risk.

This sounds scary on the surface, but the sky isn’t falling. Amazon’s new fraud prevention modeling service will undoubtedly make ripples, but in this large, crowded and growing market, it’s a lot harder to make waves. There are hundreds of risk vendors out there focusing on different aspects of fraud prevention and risk management with relative strengths and weaknesses among their groups.

The Fraud Practice’s philosophy has always been to group fraud service providers into relative categories to make an apples-to-apples comparison, then look at the features and relative strengths of each vendor within their broader and more narrowly defined groups. From a consulting perspective, our approach is “technique-driven,” where we assess the technology, techniques and features an organization needs to effectively manage risk then compare vendors that can offer the right mix of these tools and techniques.

Amazon Fraud Detector sits between the more broadly defined Fraud Scoring and Modeling/Analytics provider groups. They are not offering a fully customized modeling solution, rather from a selection of base models users can choose one or multiple models that will be implemented and tuned with machine learning based on the merchant’s historic data and new data over time. Amazon is not offering the market anything new, as it is similar to existing vendor solutions such as Sift, Feedzai and others.

Amazon’s risk services fit well for SMBs, but in the enterprise space there is more of a preference to create and manage completely bespoke models, whether that is with in-house expertise, managed services, or a combination of the two. A relative strength for Amazon Fraud Detector services is going to be their pricing model, which involves no up-front fees or long-term commitments. It is a hosted SaaS solution where users only pay for actual use.

Amazon is leveraging their 20+ years of eCommerce experience as a strength supporting the quality of their models and machine learning algorithms. While nobody can deny Amazon’s experience and breadth of data, here is what vendors competing against Amazon Fraud Detector Services should focus on when interacting with prospective merchant clients:

“Amazon is a unique case and so are you.” Amazon is the largest online retailer in the Western Hemisphere and represents around 40 percent of total US online retail sales. A major benefit Amazon enjoys that most retailers do not, from a risk management perspective, is that an extremely high share of their purchase volume is recurring customers using saved accounts and previously used payment instruments. Fraud detection is a lot easier when the same payment card is used to send goods to the same shipping address as it has been, successfully, for the past several months to years. We are not doubting that Amazon is good at recognizing high risk, but a very high share of their orders do not require the same level of scrutiny as a typical online merchant who deals with more guest account and one-off transactions without stored and trusted payment cards. How do you know the models that worked so well for Amazon will work as well for you, as a smaller volume merchant with a high share of guest account checkouts?

From a sales and marketing perspective, this is known as creating FUD – Fear, Uncertainty and Doubt. Yes, Amazon has developed based models and machine learning techniques based on their vast experience and volume, but their volume and risk patterns are very different compared to most other merchants. Their business success shouldn’t be extrapolated to all potential use cases.

Focus on where you fit in the overall equation. Most risk vendors are not an all-in-one solution, most offer various point tools that provide risk signals to be used in conjunction with other risk signals. Be very specific about what your organization offers and how that fits in the overall context of a risk management strategy. Modeling solutions like Amazon Fraud Detector serve as the hub of risk architecture, but still rely on point tools and signals derived from many different techniques and services to make effective risk decisions. Remind prospects that there is no silver bullet when it comes to fraud prevention solutions and communicate how your solution provides value in the broader context of a risk management strategy, regardless of what rules engine or modeling solution they are using. Understanding and conveying this message is explained very well in our online training course titled eCommerce Fraud: Moving from Tools to Solutions.

Focus on your domain and niche expertise. Whether it’s proprietary technology and patents, or deep experience with merchants in a certain vertical market or of a certain volume, most risk management solution providers have their “sweet spots.” Focus on those areas of expertise and define the key differentiators for your organization. This is where it pays not to be just an expert in sales, but an expert in understanding your organization including how the solution works, how it is best leveraged and applied as well as the pain points and use cases where it provides the most value. Remember that stopping fraud isn’t the only KPI that defines a successful risk management strategy and toolset. Look beyond just stopping fraud to show various points of value and ROI with switching to, or sticking with, your solution.

Most organizations are looking to complete any major changes to their risk management strategy by the end of October to be prepared for the holiday season. Now is the time to focus on retention and new sales. You may be, or have just hired, a sales expert but that is not the same as a fraud and risk services sales expert. The Risk Vendor Sales Training program provides sales professionals the knowledge and expertise to operate in a consultative sales capacity specifically related to e-payments and fraud management.

For More Information:


bottom of page