Helping customers fight fraud and prevent online abuse: Sift Science

In 2011, Sift Science was created to bring the power of machine learning technology to businesses of any size. Jason Tan, who was part of the Y Combinator incubator program, talked to several e-commerce websites, online marketplaces, and payment networks about how they were fighting fraud and abuse. He discovered that existing anti-fraud solutions were difficult to configure and use, involved setup fees, and required long-term contracts.

With those challenges in mind, the team designed Sift Science to be user-friendly and accessible to all, with easy integration, a beautiful UI, a pay-as-you-go model, and a free trial. But Sift Science’s biggest advantage over legacy fraud-prevention solutions was better technology. Other solutions relied on fixed if-then rules, which are difficult and costly to maintain, easy for fraudsters to outsmart, and result in too many false positives where you accidentally turn away good customers.

At the heart of Sift Science’s solution is real-time machine learning. Machine learning, in which computers learn from huge swathes of data and suggest likely future scenarios, can detect new fraud patterns automatically and pick up on nuances that rules cannot. The more data you send, the more accurate the machine learning model becomes. Machine learning for fraud and abuse prevention delivers three strong advantages: businesses experience less fraud, get more sales and happier users, and lower their costs for managing fraud.

Owning the machine learning fraud prevention space

Sift Science was the first to market and is recognized as a thought leader in the machine learning fraud prevention space. It was launched as a freemium product that ingested multiple data points and scored all of a business’ users based on a variety of signals, including device fingerprint, IP address, number of users per cookie, etc. Businesses using the product could give feedback to Sift Science about which users they would or wouldn’t ban that would improve the accuracy of their machine learning model.

Sift Science was extremely well-received by early customers, who were excited by the high accuracy of the company’s machine learning technology, plus the product’s ease of use. Today, the company has an impressive list of flagship customers, including Airbnb, Twitter, Yelp, Jet.com, Twilio, and Harry’s in the U.S. and Cabify, Traveloka, Destinia, and Jabong in international markets.

Sift Science has maintained a near-zero churn rate for years, and the company’s customer support is seen to be head and shoulders above the competitors. One of Sift Science’s 5 company values is “Start with the Customer” – and it uses that value as its North Star. Industry-leading technology, volume and quality of data, easy integration, a single platform approach, and a customer-first mindset are the biggest assets of Sift Science that keeps it ahead of its competitors.

Customer feedback has largely guided Sift Science’s evolution from a company serving startups and smaller businesses to one equipped to serve mid-market and enterprise customers. Customer feedback also informs key product decisions. Here’s an example: the company had been hearing for a while that its customers were seeing an increase in account takeover (ATO) attacks. In March 2017, Sift Science introduced its Account Takeover Prevention product, which is custom-built to prevent that exact problem. Similarly, in 2016, Sift Science added Account Abuse, Content Abuse, and Promo Abuse products after hearing from customers that they wanted to use Sift Science’s machine learning to prevent those growing types of abuse.

The road ahead: from fraud prevention solution to trust platform

Sift Science monitors changes in the overall fraud landscape, as well as its customers’ on-the-ground feedback, to expand the company and offerings. Fraudsters are always evolving, growing more technologically sophisticated and looking for new ways to exploit businesses and consumers. Over the past few years, the tactics used by fraudsters have become more diverse, spanning not just credit cards, but also fake accounts, compromised accounts, scammy content, fraudulent use of promotion codes, and more.

As a single-platform solution, Sift Science monitors activity from all possible attack vectors across a website or app, so online businesses can prevent every type of fraud and abuse. An additional advantage is that Sift Science provides a clear view into who a company’s good users are. Armed with that information, businesses can optimize their login and checkout processes to provide the best possible user experience – which builds trust and brand loyalty.

Sift Science believes that trust is one of the most important competitive advantages any business can have. In November, the company will formalize this message by announcing the launch of the Sift Science Trust Platform. Instead of talking about how Sift Science prevents fraud, the company is focusing on how Sift Science uses real-time machine learning to accurately predict which users businesses can trust, and which ones they can’t. Sift Science facilitates the largest trust network of online businesses and consumers on the internet, powering the trust necessary to compete and win in the digital marketplace.

Meet the man behind Sift Science

Jason Tan is the CEO & Co-Founder at Sift Science. Before moving to San Francisco to start Sift Science in 2011, Jason Tan was an early engineer at a few Seattle startups: Zillow, Optify, and BuzzLabs. Jason is an engineer/entrepreneur with a passion for great products and amazing people. He loves to learn, whether it be from books or people, and constantly challenge himself and others to be better.

“The greatest benefit of using Sift Science is not having to ever worry about fraud again. If you integrate properly and train the system to adapt to your unique circumstances, the word ‘fraud’ disappears from your business altogether.”Adam Ijaz, CEO dbrand