retail, chargeback

New chargeback strategies require right data, methods

The frequency of fraud and chargebacks rises during a recession. Chargeback Gurus‘ COO Rodrigo Figueroa said that addressing it properly requires the correct data and methods.

In partnership with Checkout.com, Chargebacks Gurus researched how fraud rates are affected by a recession. The results are significant:

  • Fraud to transaction ratios increase by 50%;
  • Average fraud losses increase by 39%;
  • Dispute-to-transaction ratios increase by 42%;
  • 27% of 500 global enterprise-level retailers surveyed say that their chargeback rates have increased in the past six months;
  • 30% of retailers say that high false-decline rates pose a very high reputation risk to their businesses; and
  • Disputes have been up by 35% since last summer.
Rodrigo Figueroa headshot
Rodrigo Figueroa

Figueroa knows how to deal with fraud. He was PayPal’s head of fraud risk oversight, Citi’s managing director of independent risk fraud, and WorldPay U.S.’s chief risk officer. 

COVID-19’s chargeback effect

COVID-19 produced massive behavioral shifts. The unique circumstances affected how companies tackle the problems arising from the pandemic. Top of the list is surging rates of first-party misuse. Preventing it begins with properly deploying data science.

As society abruptly left the office and cocooned at home, it changed behaviors, Figueroa said. With stores closed, many shopped online. Brands weren’t prepared for what happened next.

Those with problems only had one place to go — the call center. It was common for customers to wait for hours to speak with an agent. Amid the frustration, chargeback rates grew and grew to the point that 60-70% are deemed fraudulent, Figueroa said. Separating all transactions from any human element was a massive and sudden shift.

“Because you would (normally) have a handful of folks actually managing these process elements,” Figueroa said. “And suddenly, you’re dealing with three, four, five, six times more volume.

“We help merchants alleviate the pressure and recover a lot of this money that in a different circumstance will be considered lost, even though those were actually very good sales.”

Why many fraud detection systems don’t catch chargebacks

Figueroa explained that most fraud detection systems are effective but weren’t designed to fight first-party fraud. The customer’s name and account information are familiar. They are known to the company in many ways.

Usually, there is little correlation between fraud and recessions, but friendly fraud changes everything, Figueroa said. Understanding that people struggle financially and don’t want to pay for some transactions is a different situation.

“When I was at my previous employer, we had a hunch that was already happening initially,” Figueroa said. “We were thinking that if there’s pressure coming, you can probably calibrate the model a little bit more and address it. 

“It’s not the model. The model was operating pretty OK. It was never calibrated for first-party fraud. How do you identify early signals for first-party misuse?”

Chargebacks shift the workload

Figueroa said that the workload is transferred from the merchant to the issuer when a chargeback is initiated. That forced issuers to increase staff.

There’s an incentive to change the system and identify how this can be stopped sooner and at a lower cost. A good solution saves work for everyone. It’s also a lifeline for smaller companies on slim margins who cannot afford even a few unnecessary chargebacks.

New solutions are in their early stages but show promise, Figueroa said. They’ll rely on new data sets describing the client’s business relationship. How long have they been a client? What devices do they use to shop? What subscriptions do they have? Why are they canceling it?

Figueroa said the solution begins with enough data to identify accurate signals. Then you have to make a valuable solution at the pre-authorization stage.

Care has to be taken when compiling and storing data, Figueroa cautioned. Every piece has to be stored and secured. If you don’t need certain data, don’t compile it. Minimize your risk and allow yourself to focus on the most important.

But have those important data discussions.

“That friction is important,” Figueroa concluded. “Because friction will create balance in the system and lead us to a better place.”

Also, see:

  • Tony Zerucha

    Tony is a long-time contributor in the fintech and alt-fi spaces. A two-time LendIt Journalist of the Year nominee and winner in 2018, Tony has written more than 2,000 original articles on the blockchain, peer-to-peer lending, crowdfunding, and emerging technologies over the past seven years. He has hosted panels at LendIt, the CfPA Summit, and DECENT's Unchained, a blockchain exposition in Hong Kong. Email Tony here.