Big Data

Western Union uses big data to stop wire transfer fraud

David Thompson of Western Union

Above: David Thompson of Western Union

Image Credit: Michael O'Donnell/VentureBeat

Big data is coming to the rescue in stopping fraud in wire transfers, according to Western Union’s chief information officer.

David Thompson, the chief information officer and executive vice president of global operations and technology, said in a fireside chat at VentureBeat’s DataBeat 2014 conference that the company has to tap massive databases to sort through customer information and figure out the risk that a particular transaction might be the result of a scam. When they find a risky transactions, they can block it. But they lose money and possibly a customer if they block a legitimate transaction.

“The fraud prevention challenge is staying ahead of the bad guys,” Thompson said. “There are a lot of bad people in this world. In financial services, we have financial scams, lover scams, and elderly scams. They prey on consumers, and we are put in a position of trying to protect the consumer from these schemes.”

Western Union is 162 years old, starting out as a telegraph company. But it’s a big financial services company now that handles a huge number of cash transfers on a daily basis. It has more than 200TBs of data, and that data is growing at a rate of 100TBs a year now. The data comes in from 174 sources including digital platforms, retail locations, and banking partners.

To protect consumers from scams, Western Union has to analyze a huge stream of data for anomalies. It uses its Hadoop open-source software atop commodity hardware to deal with the complexity of it data queries. The Hadoop solution, which it installed in three months, is a lot less expensive and more capable than past data warehouse solutions. The fraud prevention is fast and comprehensive thanks to big data.

“I finally feel I have my hands something that I can really use to drive business,” Thompson said. “I have data that can answer the questions that the business side has been asking for 10 years. I can respond better than ever before.”

Western Union has 9,000 employees, which sounds like a lot but isn’t enough when you think about millions of possible fraudulent transactions. In compliance, the number of employees in 1,200.

They use the visualization tools of Tableau and Tibco to figure out the flow of money to figure out hot spots for fraud.

“They can get deep into the transactions,” Thompson said.

Thompson said the company can draw inferences from different kinds of databases. If you are sending money to a person in another country, Western Union will check to see if you have done this before and whether it is going to a country where fraud rates are high. It will check your email to see if it has been recently created, and it also will do the same for the recipient. It will check your social network to see if you really do know the person that you are sending the money to.

“With this information, my people can make a better risk decision,” Thompson said.

Western Union can even be proactive. If you usually send money to a mother in a distant country, but you haven’t this month, Western Union may send you a reminder asking if you want to do so. That verges on a privacy violation. But Thompson said that each person is different. Some would welcome such messages, while others wouldn’t want them.

“You have to know your customer,” Thompson said. “I ultimately want to know your personal preferences, and how I can serve you better.”


VentureBeat is studying mobile marketing automation. Chime in, and we’ll share the data.