(Dec 1): Every time someone uses a smartphone to buy a gaming mouse or laptop from Razer Inc.’s website, an algorithm from an obscure Singapore startup is checking across 1,000 data points to work out if it’s a fraud.
The type of device accessing the site, a user’s timezone and shopping patterns are all analysed in real time by CashShield and its database to work out whether the transaction should be stopped. It’s an approach that has attracted some high profile backers, from a unit of Singapore’s state-owned investor Temasek Holdings to Tony Fadell, who rose to fame for his role in developing Apple Inc.’s iPod.
“It’s the only company in the sector with a Wall Street-style, high-frequency trading approach to weighing fraud risk for companies,’’ said Fadell. “It’s clear that their technology will be in demand by many industries.’’
CashShield founder Justin Lie has spent a decade slowly building the company, using machine learning technology to speed up the process of detecting frauds to fractions of a second. After raising US$5.5 million ($7.41 million) in September in a round backed by GGV Capital and Temasek’s Heliconia Capital Management, he plans to raise as much as US$50 million more. The aim is to deepen a global presence by building sales teams in Berlin, Jakarta, Singapore, Shanghai and Menlo Park, California.
In addition to Razer, CashShield’s customers include Alibaba Group Holding and Vodafone Group Plc.
CashShield’s algorithm is built on data it acquires from third party apps, information that people might not realize is being shared when they skip through the terms of service agreement. Those reams of information cover how users interact with their devices that spans swiping patterns, pressure points, frequented locations and typing speed.
As a customer tries to shop, millions of pieces of data are analysed to decide whether to accept a purchase. Then the company applies similar principles to those a high-frequency trading algorithm employs when deciding whether to go ahead with a transaction.
“The point of tracking these data isn’t to track individual user behaviour," said Lie. “Our intention is only to identify coordinated behaviour.”
Lie founded the company around the time of the global financial crisis and became interested in studying more about Wall Street trading tools as a finance student. The 35-year-old is a graduate of Santa Monica College and National University of Singapore.
“I realised that I could draw parallels between trading and accepting transactions to view fraud as financial risks,” Lie said.
Jenny Lee, managing partner at GGV, has encouraged Lie to be more ambitious after years of slow but steady growth, envisioning a future for its technology beyond e-commerce.
“Definitely other players in the market don’t do it this way,” said GGV’s Lee, whose firm has invested in several tech giants, including Alibaba. “It’s a good approach that can be used in other areas like financial technology.”
