08/14/2018 | Episode 27
Patrick Presto is Senior Project Manager at Kickstarter.
Evan: Welcome to “Trust and Safety in Numbers,” presented by Sift Science. I’m your host, Evan Ramzipoor. You’ve probably heard of chargebacks. It’s essentially when a customer asks a bank to issue them a refund on an item. The bank usually investigates and if they think the customer has a valid case, then they transfer funds from the merchant to the customer. The customer is under no obligation to do anything, return their item, contact the merchant anything at all. It sounds like a good thing, right? And it can be but here’s the problem. Many chargebacks are actually what’s called, “Friendly Fraud.” That’s when customers claim purchases are fraudulent, but they’re really not. Depending on your sources, friendly fraud resulted in losses of about $25 billion to $35 billion in 2017 alone. To take a closer look I’m sitting down with Patrick Presto. You might remember him from a previous podcast on cryptocurrency fraud. But before he started writing a thesis on currency at NYU’s graduate program on Interactive Telecommunication, Patrick worked on risk and disputes at Square, where he got up close and personal with friendly fraud. But before we move on let’s warm up with a quick fraud fact.
Did you know that businesses and entrepreneurs are exploring Blockchain, that’s the online decentralized ledger underpinning cryptocurrencies. They’re exploring Blockchain as a potential tool for fraud fighting. To learn more check out cryptocurrency’s new fraud problems, or solutions on the Sift Science blog. Now, on to the interview.
Patrick, you’ve done a wide variety of work in the fraud and risk space throughout your career. But a good part of your career was focused on chargebacks. Can you give us a high level overview of how you think about chargebacks?
Patrick: Chargebacks really, in the context of credit card transactions or just any plastic, you know, credit, debit card transaction, it’s the process of consumers initiating like a reversal of the transaction with their bank for various reasons, and usually this is under the umbrella of consumer protection.
Evan: Something I was surprised to learn when I was new to the fraud space is that chargebacks aren’t as simple as they seem. If you can figure out why you’re falling victim to friendly fraud, and why your customers are issuing honest chargebacks too, you can get a snapshot into how users are interacting with your site. How can we go about treating chargebacks like a tool rather than an obstacle?
Patrick: It’s an interesting question because I think people really think about this knowing that everything’s digital nowadays. But it all comes down to the context of your site and the platform you operate. The way I think about it, especially when I was working at Square at the time, was chargebacks is one data point and you can actually learn a lot from it. I often think about like other companies, other people who run websites, often the common denominator and a lot of these is, the fact that you can transact using a credit card or debit card. So going back to what I said about, knowing that chargebacks are a process consumers take, usually that data point of chargebacks comes down to this proxy of someone is unhappy or they’re upset with something they bought. So chargebacks in some way is almost like a support issue but grounded in the transaction of the services that was provided between your site and whoever bought from you.
Evan: Patrick makes a great point when he says that chargebacks are fundamentally a support issue even when and especially when they’re fraudulent. Think about it. There are a number of reasons someone might file a chargeback: They’ve got buyer’s remorse, they don’t want to go through the return process or they missed the deadline to make a return. They want to avoid the restocking fee They didn’t understand when the delivery would be made, or they didn’t want to wait that long. Each of these issues is something solvable, something that could make your user experience better. Sherlock Holmes in your way into each chargeback to really dig in and figure out the customer’s thought process can be an integral part of improving your overall user experience.
Patrick: And I think that’s kind of how chargebacks can act as a window, to answer your question. Like the best way to illustrate this is, kind of, just how we learned about this focusing on like, risk management and chargebacks at Square. Solely, imagine it’s 2011, 2012 when Square was around and you as a cardholder are now able to like, buy tacos from a food truck or like, a small flower vendor and then they swiped your card on this random device on their mobile phone.
Like back then, when you think about it, it was kind of relatively new. So obviously if you’re diligent about your finances, you’re going home you’re looking at your credit card statement and you don’t recognize a name of some random charge and you just call your bank and like, “Please remove that I’m not paying for that.” What we found out is that early on is, people were charging back a purchase that was completely legitimate. They just had not known the brand, Square, at the time. And that was our first learning…is, “Oh, wow, we made it easier for sellers to sell stuff.” But we were still in the beginning stages of getting buyers to recognize the brand and understand like, “Oh, this whole thing, this is gonna come off as Square and whatnot.”
Evan: So, as Square grew, they were able to incorporate that insight into their product design. And help them avoid fraud obstacles, they scaled and even made them aware that they needed to do a better job amplifying their brand.
Patrick: This problem actually really kind of grew when we started introducing a lot more sellers onto the platform across different business types.
Evan: Okay, so let’s say you’ve gathered your data you’ve identified a pattern and the types of chargebacks coming in. How do you go about distilling that data into something actionable?
Patrick: It’s both quantitative and qualitative. I think the approach you take, from a quantitative standpoint you can see the types of chargebacks are coming in like, the average dollar but also like the associated chargeback reason code because that’s a nuanced detail is typically when you’re working with the payment process or a bank, there’s like a reason code. And that reason code is mapped to a database of why the customer initiated this to begin with. So I think that’s one helpful point. But then the qualitative stuff is like the harder things to go about. Sometimes when you’re running a business where you provide services, it’s really tough to know if you provided services and full or like, if you really met the expectation of that buyer.
Evan: This is partially because of the nature of chargebacks. Remember, at the customer’s request, the bank takes money from the business’ account. The customer isn’t in touch with the business directly. Sometimes the business doesn’t even know a chargeback has been filed until after the fact.
Patrick: Dividing a strategy to fix that grievance does go back to what I was taking about, understanding the qualitative nature of what’s going on. If you see a consistent pattern of, “Hey, people really don’t like the way I’m handling this particular part of my business. Maybe it’s the shipping, you know, maybe customers are receiving the products and they’re not happy because it’s not specifically described as it was on their website.” I think you really need to take these little data points and go back to see what you’re offering as your business, because I think that’s the only way you can really alleviate having these chargebacks coming in to begin with.
So, it’s weird because on the forefront, there’s whatever you’re running as a site, make sure you’re completely disclosing exactly what you’re selling, you know, dimensions of it, or if it’s a service exactly what you’re going to give.
Evan: What Patrick is getting at here is, while some chargebacks are unavoidable, a business can take steps to alleviate the risk that someone might file a chargeback. For example, an E-commerce site that allows users to sell items to other users, should have terms of service in place. There should be clear guidelines and rules that incentivize users to transact honestly, and all E-commerce sites regardless of whether they allow user-generated content should be sure that item descriptions, shipping and return policies are described clearly on the site. By minimizing disruptions and surprises for your user, you minimize disruptions and surprises for your business.
Patrick: But on the backend, when you’re already [inaudible 00:08:28] handling chargebacks, I mean, the whole point of having customers is establishing trust and credibility. So you wanna make sure, you know, you can resolve this in a way that isn’t either detrimental to your brand or just you know, you want to be able to leave this sort of dispute on good term for both parties.
Evan: Thank you for joining me on “Trust and Safety in Numbers.” Until next time, stay vigilant fraud fighters.
Learn more about what sets Sift Science’s machine learning apart.
With billions of compromised credentials already in criminals’ hands, how do you protect your users’ accounts, your brand, and your bottom line?