On June 30th, we invited senior risk and product executives from some of Latin America’s leading mobile apps to a virtual roundtable. Participants had a discussion on the ways in which their fraud strategies have evolved and how to know which users to trust, and which users to keep out. Guests included executives from Rappi, PedidosYa, Stori, and more.
To kickstart the discussion, executives spoke about the key fraud challenges they were facing. One participant brought up the different faces of fraud. Since fraud can take many forms ranging from promo abuse to user-merchant collusion, it’s important to know the app and what makes it attractive to fraudsters. Take the example of a food delivery app becoming a superapp. A superapp offers a multitude of services such as ride-hailing and e-wallets in addition to food delivery, so the opportunities for fraud are amplified. Fraud teams would need to find and analyze the cracks fraudsters can slip through.
Another executive discussed the downside of the rise in smartphone adoption. Mobile apps offer convenience and speed, which is why companies worldwide have been going digital, including high-risk services such as fintechs. But this also means companies gain a competitive advantage by verifying actions such as microloans in a matter of minutes, which could make them complacent in noticing the signs of fraud.
Participants then spoke about the trade off between protecting their businesses and hitting growth targets. Balancing safety with growth is a tricky task, and participants agreed that launching new products can create friction between business and risk teams. They stressed the importance of all teams needing to be aware of the risk of fraud when launching new products and incentives.
One executive voiced their main concern in determining if a user is good when onboarding new customers. They shared that fraud teams no longer have to rely on just geolocation and IP address to determine if a user is trustworthy. There is a lot more data available now which makes it important to first know what data you have.
But not all attendees agreed that having more data is beneficial in knowing if a user is genuine or not. One attendee felt that fraud teams are overloaded with data. Whether it’s device data, historical data, data from marketing teams, or data from third-party tools, data is only useful if it’s been sorted and brings value.
The conversation then steered towards how the evolution of technology has helped in stopping fraud. One participant shared that while technology is great, fraud can come internally or externally. And because it can come in many different forms, it’s imperative to be able to differentiate between them first before leveraging machine learning to generate insights.
Another executive spoke about how technology has been vital in analyzing large volumes of data. Without technology, there would be insufficient resources and time. Participants found that a big issue was making associations. For example, associations between IP addresses and users. This is where the value of data correlations come in.
To round off the discussion, participants spoke about the future of fraud in Latin America. Many agreed that blockchain technology would be a key component in fraud prevention.
