Andy Freedman, Riskified: "The solution to constantly evolving fraud is a constantly adapting solution"

Riskified is a relatively young company, founded in 2012. Could you introduce Riskified to us, and tell us what Riskified’s goal is?

Riskified’s goal is to ensure that card-not-present (CNP) fraud never gets in the way of retailers’ business goals, and to work tirelessly to turn fraud management into a growth engine for our customers. In the three years since founding Riskified, we have become a trusted partner to Fortune 500 companies, global brands and rapidly growing mobile commerce platforms including Wish. All told, Riskified has reviewed and approved purchases from over 186 countries around the world.

Riskified is first and foremost a data driven, technology company. Our system processes terabytes of data every day, and the majority of our 100+ employees are developers, data scientists, and engineers. Riskified is always pushing the envelope looking for more efficient ways to use data in order to eliminate the inefficiencies of ecommerce and grow revenues for our customers.

You have just won the MRC METAward, in the emerging category, with your newly introduced service Spektra. Can you tell us what this new service entails and the vision behind it?

The vision behind Spektra is to bring deep learning, the same technology used to create autonomous cars and automatic speech recognition, into the field of fraud prevention - protecting online retailers with the most cutting edge technology available. Deep learning is the next level of machine learning, in which a computer actually ‘comprehends’ the entire context of what it is learning.

Spektra uses behavioral analytics deep learning models, focusing on a users clickstream to quickly identify and distinguish between fraudulent and legitimate shoppers with increased accuracy.

Spektra analyzes data about their behaviour in real-time, giving it the potential to allow for real-time risk modelling that could make dynamic changes to the checkout process advantageous. For example, unnecessary friction could be avoided by eliminating steps for safe customers, while additional steps could be added for risky ones. 

Tapping into patterns hidden in online behavior adds an entirely new dimension to fraud management. The more thorough the understanding of a shopper’s behaviour is, the higher the overall accuracy of decisions become.

How many transactions have you reviewed in the past year? Are you predicting growth for the coming year? If so, how much?

Riskified grew 400% in 2015, processing over USD 3 billion in approved and guaranteed transaction volume. We continue to experience unprecedented growth this year, and our adoption rates continue to validate the incredible global demand for a flexible fraud prevention solution for enterprise brands. We look forward to expanding on our customer success and continuing our rapid pace throughout 2016. We recently closed a USD 25 million funding round, and plan to double the size of our team this year as we open a US office in New York City.

Looking at the current fraud environment, what are the most important developments in the fraud schemes, and what solutions should be created in light of that?

The solution to constantly evolving fraud is a constantly adapting solution. For example, although desktop orders are currently riskier on average than purchases made via mobile, fraudster adaptations will likely change this in the future. That’s why Riskified already has mobile SDK’s and is analyzing real-time data from millions of mobile devices to uncover fraudster patterns.

Fraudsters are also exploiting new shopping flows to their advantage, such as ‘buy online, pickup in-store’. In situations such as new and innovative products, the solution is to analyze and adapt. For example, Riskified processes terabytes of data every day, so our machine learning algorithms quickly and continuously learn new fraudster methodologies and change our risk models accordingly.

Ultimately, the most important thing that merchants and retailers should keep in mind is that the greatest fraud-related risk is being overly risk-averse. The vast majority of consumers are legitimate, and most online purchases are valid. Research by Javelin Strategy shows that US businesses lose USD 8,79 billion to falsely declined CNP orders every year. Bottom line, businesses should not let fear of fraud stop them from achieving their revenue goals.

About Andy Freedman

Andy Freedman is the CMO at Riskified, driving global market and demand strategy and strategic partnerships. He brings over a decade of branding, marketing and strategy expertise in building Riskified’s brand and partnership department. Prior to joining Riskified in 2014, Freedman served as the Vice President of Strategy and Revenue at LevelUp, leading strategic ventures and partnerships. Andy previously held multiple leadership roles at Fortune 500 companies, including Visa and General Mills. Andy received an MBA from the University of Wisconsin-Madison, and a B.B.A. in Marketing from Emory University.

 About Riskified

Riskified is the ecommerce solution for risk and payment professionals. By reviewing, approving, and guaranteeing orders, Riskified helps hundreds of merchants prevent fraud. Retailers decide which transactions are reviewed and pay only for approved orders that deliver sales to their business. The quick review process eliminates delays and friction - keeping consumers happy. Riskified’s 100% chargeback guarantee allows merchants to boost sales and expand confidently to new markets.