Data sharing: Unlocking frictionless payments and revenue with collaboration – a FLOW Summit replay

TRANSCRIPT

Introduction to data sharing trend

Stuart Gold (00:04):

All right, our next session is the data sharing trend that’s driving the next wave of payments optimization. I get to introduce the moderator he is our GM of payment solutions for, he’s the general manager of payment solutions for signified. Give him a big round of applause. Come on up.

Okan Ozaltin (00:26):

Thank you.

(00:28):

Hi everybody, and thank you for coming to our session. So we talked about this topic earlier this morning. There is a value when all the participants in the payment ecosystem work together. And Raj mentioned it, I think John Winsell from Worldpay FIS mentioned it. So today we are going to talk about a topic related to this. So there are various methods emerging in out there where merchants can share additional data with issuers and that leads to 45% reduction in false declines and couple of percentages of additional revenue conversion and also of course happier customers. So that’s what we are going to talk about today. And I just want to say to keep the excitement going, the merchants who are attending the session, I mean, if you are not looking at these options of IT sharing additional data with issuers, you may be missing out on some opportunities of additional revenue and customer lifetime value as well. So with that, I’ll start inviting my panelists. Let’s start with quick introductions. Maybe Dennis, you go first.

Deniz Ertan (01:52):

Sure. My name is Deniz Ertan. I work for eBay as a senior director of risk. So I lead the entire buyer journey from account creation, managing fraud there on transactions, stolen financials, account takeover, friendly fraud, and post transactions. Then after transaction is completed dispute management system as well as I think as a risk team. We also have a global GME enablement goal. So while managing risk, also we are responsible for enabling growth for the organization.

Okan Ozaltin (02:31):

And Amy,

The importance of data sharing in payment optimization

Amy Pierce (02:32):

Good morning. My name is Amy Pierce and I lead the merchant engagement team at Bank of America. Our team is focused on providing end-to-end payments expertise to merchants. We are an enterprise unit and very unique in that we leverage our card issuing data to provide reporting and insights and to merchants. And speaking of data, we use data analytics to really share that performance with merchants look to optimize their payments process with a focus on improving authorization approval rates.

Okan Ozaltin (03:08):

And

Hunter Roberts (03:09):

Hunter? Yep. Hey, Hunter Roberts, Capital One. I’m part of our credit card organization, our fraud team. Within our fraud team. I lead a group called Network and External Strategies, which is the easiest way to think about it is it’s sort of an in-house consulting firm that goes out into the ecosystem to sort of inform our internal fraud products and business strategies. Happy to be here.

Okan Ozaltin (03:31):

Cool. So let’s get started now. So everybody’s probably wondering what do we mean by this data sharing from merchants to issuers and how does it work? And when we say sharing what kind of data we are sharing and how does that actually help merchant actually convert more transactions into revenue? So maybe Hunter, we start with you first. What do we mean by that when we say data sharing with merchants and tool issuers?

Hunter Roberts (04:00):

Yeah, I think I’ll answer it by explaining and there’s already been some allusions to it earlier today. Raj mentioned it, John from FIS mentioned it. There’s just this massive chasm between the type of data that’s being assessed by a merchant or someone on their behalf and what issuers are using. So merchants have a ton of information on the behavior of the shopper. They have a better understanding of the customer across payment type or cards. We have a ton of information about what the customer spends their money or our money on. Those two things don’t talk if we find a way to bring those two things together. I think we’ve spoken hypothetically in the past about thinking that that could drive value, but we know it now. We’ve proven it, it matters.

Okan Ozaltin (04:56):

And to me maybe to add some more details, what are some of the examples of these data fields that really helps the issuer make more accurate decisions?

Hunter Roberts (05:07):

Yeah, I mean any type of address, we have lots of addresses, we have lots of phone numbers. All of these common elements that are present in a credit card transaction and the relationship that a customer has with a bank are really helpful both at the time of a transaction. I know we’ve heard some people talk earlier in the session around pre post or pre-transaction and post transaction data. All of those things are relevant, all of those points in the process as well.

Okan Ozaltin (05:42):

Okay, makes sense. And just to confirm, I mean there was a comments made in earlier session. I mean you as an issuer, you are also trying to constantly balance what you approve with fraud and you are also trying to make the cardholder happy here, right? Your goal as an issuer, you are also trying to, same as merchant’s goals, you are trying to improve as many trends.

Hunter Roberts (06:08):

The journey of what we’ve done from a data sharing perspective is always sort of rooted in as many wins as we could create. So this is good for our customer, this is good for the merchant’s customer, this is good for the customer. There’s no loser. I think in this ecosystem, in this industry, there can be times when we’re sort of pitted against each other. It’s all about liability. It’s like whose fault was what. And I think incentives are aligned for merchants and issuers to do more to talk where the two brands that are the most relevant at the time of a transaction, the card that you’re using in the merchant site that you’re buying something from. So for those two things, not to talk going forward, I just is a huge missed opportunity.

Okan Ozaltin (06:58):

Yeah, makes sense. And let’s get merchant’s opinion on this. So Dennis, what do you think about the data sharing and what possible experiments have you done with that?

Defining data sharing between merchants and issuers

Deniz Ertan (07:10):

So on the merchant side, we have quite rich data because from customer interaction on our website, so the payment instrument that they’re using, their IP address, geolocation, phone numbers, their purchase history with us. And I think all these information is so critical for the issuers to know because from the merchant lens we are seeing some potential risk score, low risk. But since we are not sharing that information with issuers sometimes, hey, why this is such a kind of great transaction, we don’t think that’s risky, but it’s being declined. But there’s definitely that information missing information there. And also I’ve been in fraud for a little over 20 years, so I think it’s fascinating how there is no competition in fraud area. So I think sharing definitely is a win-win for all sides. So I think we need to just build, establish that kind of effective line of communication and way of sharing data.

(08:09):

I think from a few examples, I already gave some, but in addition to kind of phone number, email address, billing and shipping and most of the time for uss are not shipping to the address that they stole the identity from. But I think that information is so critical for issues to match what we are seeing versus what they have on file as well as some IP phone number, but also customer purchase history, maybe not a lot of details, but for example, if that payment instrument is seen for the first time at eBay is different than, hey, this person has been shopping with us more than 10 times in the last 12 months. I think it’s definitely a different story than kind of first time user as well as I know it can be kind of tricky depending on the merchant, but the models that we have, we have very sophisticated models that’s using behavioral metrics, all kind of type of data and we can maybe share low, medium, high risk with issues so that they can have that lens saying that, hey, this transaction according the merchant is the highest risk transaction and they can take immediate or just relative action.

(09:20):

So I think there’s definitely loss of benefits and a win-win for both sides.

Okan Ozaltin (09:26):

Cool. And I think we talked about direct data sharing a lot here, but networks have created some products to facilitate this. So from merchant perspective, what are your thoughts about, for example,3DS? Have you tried that path or your experiences there?

Deniz Ertan (09:50):

Yeah, so 3DSis a protocol that really adds additional authentication to reduce fraud by adding additional kind of authentication methods. So on paper it sounds great, but from the issue perspective, I’m on the merchant side, but from the issuer perspective, I think moving liability to the issuers without being mandated to share some critical information is putting them at a disadvantaged position because I think you’re sharing, hey, just kind of take the liability. But on that other side, they don’t have much information. So of course merchants can choose to send those, but since these are not mandated, I think it’s not very common to share. That’s one piece. The second piece is actually we are still not clear on whether we should be sending all traffic or only risky traffic. Of course sending all traffic means cost, additional cost, but if we only send risky traffic, which I have an example we tried last year, if you send only risky traffic, issuers might only see the risky part and they may start to decline more of your transactions.

(11:03):

So we always have a kind of concern of, hey, is it going to impact our conversion rate? Which is kind of a big important topic for us. Yeah, I think there is some big knowledge gap in my opinion. I dunno from the merchant perspective especially. And going back to the example and when 3DS first came out in the US and we said, Hey, we want to pilot it, it’s a great opportunity for us to move the loss liability. And I’m glad that it was a pilot. We were trying and we decided to use a fraudulent set of transactions that we know that fraudulent, but we thought that hey issues would challenge them and decline them. And unfortunately 80% of those transactions got approved because issuers have chosen not to challenge those transactions because also they are concerned about conversion rate. So it was a 15 days of disaster.

(11:57):

But from eBay perspective, of course we didn’t have the lost liability. It was good for us, but our fraud rate of course in that population was so high. And during that time, even with some issues, and they’re not here by the way, did not Capital One or Bank of America, they identified some issue on their integration as well and we learned that we are supposed to send a flag to them saying, Hey, this is a risky transaction, just treat them accordingly. But it was an interesting experience that we stopped and we never turned it back on. So

Merchant and issuer perspectives on data sharing

Hunter Roberts (12:28):

Plus one, all of that, I think the incentives for 3DS are really misaligned. It’s more if you did it the right way and see that in the regulated markets where everything invokes it, it’s cost-prohibitive for merchants. To your point. So that leads to the adverse selection that you mentioned, which is it’s just going to be the risky stuff. Well then years and years of issuer fraud models, learning that this type of transaction is inherently risky is going to drive down approvals. So then issuers are really left to make the decision. What we’ve done is we just challenge everything because we can’t begin to understand the multiple 3DS strategies of every merchant in the payment ecosystem. So it’s like, I don’t know if you want me to step up everything or you want to go frictionless, so we’re just going to, we’re do the thing that the thing was built to do, but we know that’s suboptimal. So I think there’s some data-only products that we’re going to talk about that I think are an improvement upon that. So they sort of reduce the friction, take the friction out of it, remove the complications of the economic, the risky economics of the liability shift and just start to get better data into the decision.

Okan Ozaltin (13:45):

And Amy, any reactions from yours?

Amy Pierce (13:47):

So we do see that 3DS , it does its job of reducing fraud, but I don’t think at the same time that we fully have adopted it as an ecosystem. So there’s opportunity there for issuers, merchants, acquirers really to come together and the networks to determine what is the best use, who should be using, when should they be using it. The other item I will add, talking about data, there is another option to3DS , 3DS data only. So sounds like a good solution. As an issuer, we would receive additional data elements that we could use to authenticate the transaction. It goes through the frictionless flow. So sounds great. We did build out that functionality, however, we haven’t seen the merchant adoption yet. So it does take all parties for it to be successful. So while we are ready to test that, we need to have the merchants ready to participate as well.

Okan Ozaltin (14:48):

And Hunter, maybe you can comment on this. So in terms of the direct data sharing that’s outside of the 3DS rails, right?

Hunter Roberts (14:58):

There’s probably a hierarchy of free and frictionless, then there’s sort of the network, the solution for frictionless, but probably at some incremental cost and then sort of the less desirable, pretty expensive, pretty friction-filled 3DS flow. I think that’s probably the easiest way to organize how we feel about what the opportunities are here.

Okan Ozaltin (15:26):

Makes sense. And speaking about this network solutions, so Visa just announced, actually they announced it a while ago, but now they have this compelling evidence program. So I’m sure many merchants heard about this. So it’s around preventing first-party fraud where merchants can collect some data points about transaction and if they can submit proof that the same person transacted with the same merchant in the certain period of history, they can actually challenge the chargeback. Any thoughts around that? So Dennis, are you exploring that at eBay?

Deniz Ertan (16:13):

Yeah, of course. Friendly fraud is a big issue in e-commerce and retail business. And it’s very difficult to detect because these are legitimate people and there is no exceptions or just suspicious activities around them. But what they end up doing is I didn’t make this purchase type claim. So this is a big issue because of its nature, but I’m super happy about this visa compelling evidence program because it’ll allow us to one, not to stop those transactions because as the buyers are just utilizing this friendly fraud and abusing the system and our models and rules, we have started to decline them more. And now with this program and we already started to adjust our strategies to allow those transactions instead of declining them at the first place because we know the evidence that we have shipped to that address before, that payments instrument has been used on our site before, an IP address, et cetera, a few other elements. So that will be another great way for us to grow our business because we don’t have to decline those transactions or just take up some losses. So we are very excited about this program.

Okan Ozaltin (17:27):

And from issuing side, hunter, doesn’t this put you under pressure because now merchants can collect some data and push back on some of these first party frauds?

Hunter Roberts (17:40):

Yeah, it might be a surprise. We’re pretty into it from an issuer perspective, it’s sort of controversial. Anytime you’re shifting, again, just goes back to the whole fight over liability. This makes sense. This stuff is not good. We don’t want this for merchants. For us, we’re all spending way too much money and losing way too much money on this. So if this is the right thing and we’re aligned definitionally to how Visa has defined it, a little bit of concern on some of their execution decisions. We can talk in the hallway about that. But the big thing, I think the thing that the biggest net positive that we expect to come out of this is kind of second order. We think that this is probably 30 to 40% of our car not present claims. That’s doing a tremendous amount of damage to our models. We just like you’re putting that much noise, where’s the signal? So we don’t really completely understand what the upside could be from a modeling perspective once we’ve stripped out illegitimate claims. And that goes just beyond Capital One. I mean the network models are being trained on a significant amount of things that are not fraud.

Okan Ozaltin (19:06):

Okay. Amy, any reaction?

Amy Pierce (19:08):

Yeah, I was going to say at Bank of America, we had already started down the path of trying to pull in our own data and use that data in conversations with customers as they call in to file a claim. So now when Visa’s compelling evidence 3.0 came out, it really just projected that it is a huge change for our claims teams and the claims applications. I’m happy to say all of our requirements have been updated. So we are ready to go with the mandate. In fact, there is a beta test with a certain number of merchants that starts this month. They’ll run through June. Now I think the adoption’s going to be depending on how the merchants use it again. So we’re hearing it is a complex build on the merchant side as well. In addition, merchants have to participate in order insights, so to provide that additional data to issuers. So we can use that when working on a claim. Now I think the good piece is under the regulations we have to do an adequate investigation, but now we can use that data in that investigation to determine the legitimacy of the claim and then either talking through with the client, make that determination, do we pay that claim or do we deny a lot of opportunity I think here for claims deflection and to get that fraud out of the ecosystem.

Visa’s compelling evidence program and its impact

Okan Ozaltin (20:33):

What I’m hearing here is at least with Bank of America and Capital One. So what you’re saying is you will get the evidence and you will push it back to the consumer, the cardholder if right evidence comes to you from the merchant. So you are not concerned about constant friction between the cardholder and you and saying, oh no, this is not me, but the evidence is there so you have to take it as an issue. Any comment

Hunter Roberts (21:03):

On balance? I think that that’s right. I mean there’s edge cases that we’re a little concerned about. Obviously complaints and regulator attention or pressure is certainly something that we’re being very thoughtful about. But I think even in places where maybe we as the issuer don’t hold the customer liable and take the loss to my earlier comment, I think the value that we get in other places is going to completely overtake whatever incremental losses that come from it. Yeah, makes

Okan Ozaltin (21:47):

Sense. Any question.

Amy Pierce (21:49):

And it gives those claims associates, those front end associates that are talking with the customer now, they will have the data to talk with the customer, walk them through, have they seen those data points and prior transactions have a meaningful conversation with the data to back it up?

Hunter Roberts (22:08):

Yeah, the whole thing is just data wherever we can get it. I think that’s sort of the thesis of at least our position on this panel. If it’s at the transaction, if it’s after the transaction, we just need more and we can sort of get closer to where merchant issuer incentives are aligned.

Okan Ozaltin (22:31):

Makes sense. So maybe we have a few minutes remaining. So in terms of advice to merchants, merchants are trying to improve their authorization rates, they want to reduce the event of declining good customers. So
what advice would you give them? How should they work better with you? So Amy maybe

Amy Pierce (22:57):

Sure, I can jump in on that one. As we work with merchants on a day-to-day basis. So a couple things. I mean we talk about information sharing and not only data, but additional information. So with our merchant engagement team, if merchants and issuers can partner together to share information, I’ll share some examples. If a retailer is getting ready to do a hot new product drop or maybe a streaming service is going to stream a popular event, pre-Black Friday sales, things like that. If they can let us know, we can then work with our fraud strategies partners or authorization partners. So if they see a spike in purchase volume and transactions from a particular merchant, they expect this. They know it’s not a fraud attempt. They can monitor, they can adjust fraud rules as needed. So very important relationship. The other thing I would say is just going back to the importance of data, different indicators. So we see a lot of clarifying. The merchant name field seems pretty basic, but we need a clear field. So when that displays for the customer to see, they know what it is, avoids calls into the merchant, avoids calls into our call center using card on file indicators if that card is stored, using recurring indicators if it’s a recurring or subscription type transaction. So all of that information, the more that merchants can share with issuers, we can partner together to make a better decision and it benefits our mutual customers.

Okan Ozaltin (24:36):

Thank you. And Hunter comments from your side?

Advice for merchants to improve authorization rates

Hunter Roberts (24:40):

Took a bit of a turn. I think this is, it’s easy to sit up here and be like, ah, you should share your data with us so we can do a bunch of cool stuff for you. We know that that is not as easy as just doing it, that everybody has lawyers at home that are going to tell you all of the reasons that that is a really bad idea and probably an impossibility. We’re pretty confident that the privacy concerns can be alleviated. There’s a lot of specific regulatory privacy carve outs that we feel really comfortable about using things for fraud purposes. So the thing I would ask is just, you may not be ready yet, but I really genuinely believe that the people who move here are going to see the value and just start having the conversation about what would need to be true for you to be a part of more data going to issuers for better decisions at every point in the lifecycle of a transaction.

Okan Ozaltin (25:56):

Okay, cool. And Dennis, from your perspective, any comments in terms of how you would want to see the relationship working?

Deniz Ertan (26:06):

Yeah, I think we need to establish an effective way of communication between issuers and merchants. I know issuers have many merchants, but especially with the big retailers, e-commerce companies, because we have lots of resources and tools and models that we can mitigate fraud wherever they see a concern. But whenever we see increase in our declines, we don’t know specifically why it’s happening. And I think that effective line of communication would really allow us to share the controls we have that would give the issuers the comfort that we are managing it well or based on their expectations. I think this is all about understanding their risk appetite and adjusting our strategies. So we can definitely just go into that win-win situation again if we can establish that kind of effective way of communication between merchants and insurers.

Okan Ozaltin (26:59):

Cool. And maybe in closing, so as you heard, there are various different models of data sharing emerging in the ecosystem and signified sitting in the checkout and pre-authorization can act as a hub and connect all these data flows to different issuers and different issuers and different solutions may require different data fields. So signify plays that role of communicating these data to issuers on behalf of the merchants. So I think we are not able to take questions in the session, so please find us after the session if you have any additional questions. And with that, please give a round of applause to our panelists.