illustration of credit card going through a phone

Payments, Friction, Customer Service, and Fraud

Ali Paterson from Fintech Finance recently sat down with David Bailey, Head of Business Development and Customer Innovation at Santander UK and Feedzai’s own Andy Renshaw for “Payments, Friction, Customer Service, and Fraud.”

What follows next is an insightful conversation that looks at customer friction pre-and-post COVID-19, the future of cash, fraud during Coronavirus, the emergence of contactless payments, AI and so much more.

 

Below, please find an abridged transcript which has been edited for clarity.

Ali Paterson: What were customer expectations pre-pandemic and what’s been the biggest issue in meeting those expectations? Basically, what are the pain points in customer experience for payments?

David Bailey: If we’re talking about pre-COVID-19, I think the reality is that customers have a pretty balanced view when it comes to payments. They do want it to be as seamless as possible, but I’m pleased to say customers are wary of anything that’s going to cause them inconvenience or loss of money, or anything unfamiliar. That’s good from a fraud perspective.

The reality is customers make decisions within milliseconds as to how they’re going to pay for stuff. That’s part of the challenge when you try to introduce new things. In the payments arena, it’s quite hard to get customers to change their behavior. Sometimes it needs a bit of a nudge from industry because customers can be a little suspicious of new stuff.

Interestingly, that’s true when it seems too easy as well. If you think about when contactless payments first arrived, I think people felt there must be some kind of catch. It was only when there was an industry incentive for free underground travel, and people saw that nothing bad happened, that we started to see things ramp up. I think people were almost expecting there to be more friction in the journey.

Andy Renshaw: I think post-COVID-19, the expectation was that something would change, but customers weren’t sure how it would change. To a certain extent, customers have been exploring this on an almost week-by-week basis. Contactless payments and the raising of the credit limits are a great example of this.

We’ve also seen customers who haven’t been digital or e-commerce users become digital customers now. A good example of this would be new online supermarket shopping. In some ways that’s been reassuring, but in other ways, it’s created a need to reestablish trust. And we’ve seen customers trying to do exactly that. I think that’s probably what we’ve seen happening over the past few months really — exploring trust with established people in different channels or with new companies that they perhaps hadn’t used before.

AP: Since the COVID-19 lockdowns, there’s been a massive spike in online payments and e-commerce, especially at the start of the lockdown. Do you think there’s been an increase in fraud, or do you think that’s been counterbalanced by the public being more cautious?

David: We’ve definitely seen a lot of different kinds of fraud appearing since lockdown. From a bank perspective, we’ve had to be careful and encourage our customers to take care when it comes to fraud because there’s all sorts of scams that have suddenly appeared. It’s astonishing how quickly the fraudsters picked up on the opportunities that COVID and lockdown have provided.

The other side of the coin is that we’ve seen contactless payments have just taken off, even with late adopters to contactless. Pre lock-down one in four consumer transactions were contactless. That has skyrocketed even more. In some ways, the industry should be quite proud that we were so quickly able to push that limit up to 45. I think that made customers relax about using contactless.

Andy: The typology or the method of fraud hasn’t changed, but the language around it has. Fraudsters are leaping upon the weaknesses and the exposure points that COVID-19 has brought, and they’ve shifted the narrative around why and how they’re trying to scam the customers.

From an e-commerce and general perspective, we need to be aware that in a recession, or in a downturn more broadly, we know the different types of fraud that start to come to the forefront. Fraud types like hardship fraud, refund chargebacks and claims for non receipt of goods will become prevalent. I wouldn’t say we’re seeing that yet, but we’re aware that as this continues, those things may come to the forefront.

AP: Do you think we’re starting to see an aversion to cash, but in a way that isn’t leaving people in a financial island where they can’t transact at all? It’s a tricky question. We want people to be able to transact and not be left out. Yet, there seems to be a bit of a decrease in the acceptance rate for cash.

David: Clearly, the use of ATMs declined hugely during lockdown. It’s starting to go up again, but we don’t see it going back to pre-COVID levels because people have become accustomed to paying in different ways.

There is a risk for people who are not accustomed to paying for things or managing their money using methods besides cash. And the truth is, they’re going to have to adapt because a lot of merchants just aren’t accepting cash anymore. In particular, I feel sorry for charities who have been reliant on people giving them their small change. We see a bit of a move for charities to be able to take contactless payment, which has been great to see. But I am worried about that. I hope they can adapt more quickly.

Within Santander, we work with some charities to help them start to do that and to take contactless payments on public transport, etc. But I think there’s going to have to be a change in people’s behavior because the merchants are going to start to encourage it.

AP: With such a movement towards online, electronic, and ecommerce payments, particularly in terms of transaction volume, how can technologies like AI and machine learning help with managing that transaction volume both in terms of reconciling and in recognizing something looks suspicious?

Andy: I saw an article from MIT Technology asking that exact question. From my perspective working at Feedzai’s, there’s encouragement. Fundamentally, machine learning models still look at individual behavior and whether that is normal or not. The models are used to doing that. For example, payday peaks, Black Friday, Cyber Monday, etc. — these are big changes to behavior that the models don’t have a problem with.

Although the coronavirus is unique in terms of the length of its occurrence, some of the underlying principles behind it are not unusual for machine learning models. And fundamentally, what we’ve seen is that they’ve stood up well to the challenges. Although customers are behaving in a different way, the transactions they’re doing are still relatively normal. They’re still using merchants they used before; they haven’t suddenly started using a whole new cohort of merchants. Where they were making high-value payments, that’s still the case, and where they’re making low-value payments, that’s still the case. In that sense, when you look at the level of normality, it’s still very high.

The second part of the machine learning model is looking for anomalies or fraud behaviors and outliers. Fraudsters are still, in many ways, committing the same types of fraud, and those transactions still look unusual relative to normal behavior. However, the mix within the normal range has changed. What we see is a big difference between fraudster behavior and authentic customer behavior. Because of that, what we see is actually more stable AI performance. The ability to segment and separate the bad from the good has remained consistent throughout.

David: Increasingly, we’re looking toward machine learning and AI, particularly on financial crime and fraud, to help us identify suspicious patterns of behavior. It’s been in a complementary fashion for the tools that we use. We’ve always had to combine that with quite a lot of human interaction and human intuition in terms of figuring out what we should direct the machine learning capability to do, and what we do with the outputs of it as well.

The other thing from a banking perspective, when you’re integrating machine learning into a highly regulated environment where everything has to be auditable for decisions and outcomes, that makes it a bit slower to integrate some of these things. Still, there’s no doubt that machine learning and AI are the way forward to help us cope with increasingly sophisticated fraud patterns and fraudsters. It’s not going to get any easier, so we have to continue building our capability.

AP: Do you think people almost need a certain amount of friction to give them the peace of mind against fraud? For example, I don’t need any friction, but my mother doesn’t really trust the process if there’s no friction.

David: That’s a really interesting question because the reality is that many people expect there to be a little friction. The analogy is, you lock your front door, right? I think people feel the same way about their finances. People expect there to be some friction in their journeys, although they are happy to remove the friction when it’s a trusted brand. They’ll leave their card on file with you, but if it is somebody new, people will be naturally hesitant to do that.

The reality is that legislation is pushing us more towards security as well. Where there has been friction in the past, we’re actually starting to see a bit more friction. For the most part, I think people are OK with that.

What I see coming from an open banking perspective are opportunities for there to be less friction. If you look at some of the journeys now, they’re becoming quite slick. If you think about some of the big, digital platforms that will start to use open banking in a frictionless way compared to what they may have to do for cards. There’s real potential for removing friction from those journeys where people trust the brands.

Andy: We tend to think about friction as customers are intervening. But my experience in the past was really just about telling customers what you were doing. I was fortunate to participate in one app launch where payments happened too fast. We actually had to tell customers through the payment phases that we were doing security checks. The customer didn’t have to do anything different. Informing them, going a little slower, and effectively communicating that we needed them to wait a few seconds while we conducted security checks in the background, added a huge amount of reassurance. There’s friction in terms of intervention, but you can also do it in a passive and informative manner as well.

I think we are just starting to see the little acorns of this, especially with some of the challenger banks who bring things like the ability to turn off gambiling transactions or perhaps international transactions. Maybe we haven’t thought about those as friction, but things like card controls or certain payment limits that customers could set might be viewed as the customer starting to choose where they’re OK with friction and where they’re not. It gives them control and pro-active choice.

Generally, I don’t think we see a differentiation where person A has a different journey to person B, but we are just starting to see these things around choice and allow customers to set some of the limits on their side rather than the bank side.

AP: If you’re differentiating for each customer, does that leave you relatively exposed for fraud if you’re doing it at that kind of scale?

Andy: Yes and no. In some ways the choices that people make would actually be useful data and in essence a lead indicator. Customers who choose high levels of friction or to not participate in certain transactions, you could argue that, in many ways, they are easier to protect.

On the flip side of those, are people who are more permissive or who’ve chosen to remove friction. That might be where you choose to focus your machine learning models and rules on a day to day basis. In many ways, it could help you be more targeted in your strategies.

David: I think there could be some complexity around where customers choose a level of friction and they get something different because of fraud rules that are in place or because the transaction has been highlighted for some reason. And also some complexity around liability that you’d need to be very aware of as well.

AP: We’ve got to talk about the future. Let’s assume there is a lot more physical POS than we’re experiencing at the moment. Do you see things like Amazon Go models becoming the future? And, when it comes to online payments, what are some of the tools that we’re going to be experiencing at the required scale?

David: The way I think about the AmazonGo model is that change will be bigger than people expect, but also slower than people expect. There’s a real potential for that model to appear in the future, but I think it will happen quite slowly. I’ll give a personal example. I’ve started using the filling station apps when I fill us so I don’t have to go into the shop. But I still feel this compelling need, after I’ve filled my car and paid for it on the app, to wave my phone at the attendant before driving away because I’m uncomfortable with the thought of driving away without paying. And I know they get a notification, but I still want to wave at the person. I think it will take some time for consumers to get over the thought of just walking out without paying. That sort of thing will definitely come and we’ll see payments as part of broader journeys become much more integrated in a way that feels much more seamless to the customer. You’re going to see that more and more, particularly within big digital platforms and marketplaces. Transport will become much more seamless, and also other things such as micropayments or media consumption.

If you think about the media, that’s a really challenging model for many media companies. They’re probably going to need to get to a point where they can take micropayments for content instead of continuing to rely on the advertising model. It will be interesting to see whether things like that happen.

AP: I’m not sure. Looking at Uber, whenever I’ve taken a black cab, I get out and then remember that I have to pay; then I get back in.

David: As always, these things will take time for people to get used to. Once again, it will probably take an industry to push things or at least some of the big brands to really get people comfortable and incentivizing people to change their behavior. They need to see that nothing bad happens. I hope to see traction starting to occur in the open banking space. Customers don’t know enough about it, but they’re starting to see a bit of growth in open banking, for instance in the charity sectors. There’s real potential with open banking to remove some of the friction.

AP: Andy, what are your thoughts on the Amazon Go model and what are the tools to roll this out at a big, big scale?

Andy: We’re definitely seeing the convergence of social media with digital financial interactions. We know that the phone is becoming more homogenized as a payment vehicle. Those things are only going to continue in the low-value payments. And because of this, I think we’re going to see cash almost being removed because micropayments can be made in simple fashion.

It’s slightly different when you look at higher value payments, deeper transactions, or transactions where you expect to have a long-term relationship. For example, say you’re buying a car and you’re going to return to that dealership on an annual basis. Is that something that you’re comfortable doing using a relatively contact-free payment? I’m not so sure.

We’re seeing that the wallet or membership type model is definitely taking off. People are comfortable with their payment details registered with someone they trust. Then that becomes the vehicle for payment because they’ve already established trust. I think that will continue to be true.

One of the problems with cash being removed will be budget management. For example, some people are used to taking out their money one day a week and that lasts the week. People who can step in and help others manage their budget and help give transparency around virtual money or spending could be successful. My hope is that open banking can present an opportunity in that space where people can actually have their budget managed as well as their payments made convenient. Those two things coming together could be a powerful proposition.

Want more information on post-COVID strategies? Watch the panel discussion:  Crisis, Fraud, and the use of AI.

Subscribe to stay infomed

  • Digital Transformation