Why FinTech Product Managers Should Work in Fraud Prevention
This post was inspired by a tweet from @robleathern that I whole-heartedly agreed with. In fact, I have adapted the advice from this Tweet to my pitch to my team and prospective candidates:
Why work in fraud prevention?
- The mission and purpose are very clear: you’re fighting to keep money away from the bad guys. Some of the money lost to fraud goes to things like funding terrorism, sex trafficking etc.
- The work is fast paced and ever changing: Sometimes its a day to day battle against the bad guys, while sometimes its more strategic where we need to think about the next wave of innovation to beat them back.
- The financial leverage is enormous: McKinsey estimates that global credit card fraud losses were $23B in 2016, and likely to rise to $45B by 2025! Helping to bend this curve by even a little bit can lead to a significant P&L impact for banks and merchants.
- Fraud is one of the few places in FinTech where alternate data (beyond the bureaus) is actually living up to the hype e.g. the NYT had some this awesome article on the use of behavioral biometrics in fraud prevention.
- Strategically, fraud prevention is only going to become more important: with increasing penetration of e-commerce, digital banking, and consumer expectations around real-time frictionless experiences.
Why is knowing Fraud important for Fintech PMs?
Simple: Fraud could break your product and make it unviable.
Fraudsters are devious and creative individuals. They know that early versions of products often do not have fraud prevention measures built in, and they take advantage of this. Famously, Apple Pay was hit by an industry-wide fraud attack fairly early on; Fuze, one of the many “whole wallet in one physical card” products, found that early users were Fraudsters who used the product to load stolen credit cards. The company then needed to spend a bunch of its time working with the Secret Service to sort out this issue.
As an aside, the ability of malicious actors to use the features of your product to create havoc is not unlike the conversation we’ve been having as a society in the aftermath of election interference by state actors using the core capabilities of social media platforms. This dynamic also means that the techniques used to fight credit card fraud are also applicable to how Facebook fights fake news and information operations.
Full disclosure: there is a big downside of being a PM in fraud
Sometimes it’s disheartening and frustrating to see fraudsters exploiting holes in our defenses. We often see a “balloon squeeze” effect, where if we secure one part of our business, fraudsters will turn their attention somewhere else. We have this morbid half-joke “Fraudsters have to feed their families too…”
Why I still love working on product on fraud
- You’re often solving for three types of customers: the fraudsters, the real consumer whose identity or credit card is stolen and many different types of internal customers (fraud investigators, operations, investigators etc.). Given this dynamic — you get to flex many different kinds of product development muscles. ML driven products to fight fraudsters, lean product development/design thinking for external / end consumers and co-development for internal users. Personally, for me — building products for internal users is especially fulfilling because these folks are colleagues and friends and being able to make them more efficient at their jobs is an awesome feeling.
- Building products in fraud is an amazing blend of AI to catch the bad guys and human-centered design to cushion the blow for real users. On the AI / ML side, many super interesting techniques are being deployed across the industry e.g. network/graph modeling to detect fraud rings, human in the loop AI to leverage human investigator to identify trends and then have the machines automate defenses via Natural Language Processing, Deep Learning, Robotic Process Automation etc. To support these applications, fraud PMs also have opportunities to work on the infrastructure/platform side on topics like distributed computing, machine real-time decisioning, various database technologies etc. If you think about it, running a credit card transaction, or a mobile app login through a complex fraud risk assessment algorithm involving hundreds of data feeds in a few mili-seconds is an incredibly challenging technical task.
- There are some super interesting problems to solve on the human-centered design side: Having your identity or credit card stolen can be extremely upsetting to consumers (even if they are not liable for the fraud losses). Building products to guide users through this extreme state of heightened anxiety is an exceptionally challenging task. False positives (e.g. a credit card transaction that gets erroneously declined) can be extremely frustrating to users and a “moment that matters” in their overall relationship with the brand. Finally, “friendly fraud” where customers file fraud claims on legitimate transactions is an industry-wide problem: Javelin estimates that ~15% of fraud claims could be of this nature. Typically, this behavior stems from confusion, and solving this problem requires completely new frameworks for developing products.
- Direct P&L impact: Our work has direct P&L impact and is hugely fulfilling from that perspective. It’s also useful for PMs to flex their analytical and P&L management muscles when navigating the pathway to higher-level Product GM roles.
- There is a vibrant start-up eco-system in this space which is constantly pushing the state of the art and driving a lot of energy.
All views, opinions and statements are my own
