Transcription
Welcome back to Behind the Checkout, the podcast where we explore what really happens behind the payment button. I'm your host, John Lun, CEO and co-founder of Gravy.
For years, merchants built payment stacks around a fairly simple assumption. They owned the customer relationship. They controlled the checkout experience and decided how transactions flowed. But that assumption is starting to break down today. Platforms, marketplaces, wallets, operating systems, and increasingly AI-powered experiences are inserting themselves between merchants and customers. The question is no longer just how you optimize payments, it's whether merchants still control them at all.
Today's guest is Michael Habman, Senior Director of Operations at Radial Commerce Solutions. Michael has spent years helping some of the world's largest retailers navigate the operational realities of modern commerce from checkout performance and payment infrastructure to fulfillment and customer experience at scale. Michael, welcome to Behind the Checkout.
>> Thanks, John. It's great to be with you. It's truly an honor uh to get to talk to you today.
>> Thank you. It's going to be great. So, in this episode, we're going to ask some simple but possibly uncomfortable questions, right? If you had a chance to rebuild your payment stat from scratch today, knowing everything you've learned over the past decade, what would you do differently? What technologies would you keep? What assumptions would you throw up? And how would how should a merchant prepare for a future where more and more purchasing decisions happen outside their own channel. So, we'll break that down. I'll break it down in some some questions and and we'll see where we go.
So, let's start. First, most merchants inherited payments infrastructure that was designed 10, 15 years ago. What's the first architectural challenge that you would challenge you would change?
>> So, I the first the first thing I would change and I can I can speak to this pretty accurately because I was on the merchant side for quite a while. So, I did inherit a a legacy system, I'll call it. So I would I would rip out the single acquirer type architecture where you're just completely doing static routing and that's things have changed over the years obviously because the globalization and just I think the options that are out there for customers use payment method wise have changed significantly and the speed has changed quite a bit as well.
Okay, next question. Between Apple Pay, Shop Pay, Amazon Marketplaces, social commerce, AI, Agentic Payments, are merchants becoming essentially fulfillment providers rather than commerce owners? Have they already lost control of their checkout without even realizing it?
>> Yeah, this is a great question. They definitely have have lost some control. Uh, but it's also helped them optimize and grow their revenue. Quite a bit. So the way I look at this and having, you know, been been on the fulfillment side in the past and being on the fulfillment side now, we can see that revenue has definitely increased for those merchants that have chosen to sell, take for instance a social on a social commerce site because essentially they're they're able to the social commerce site takes all the all the heavy lifting off that merchant. So they know who this person is. They've transacted with them transacted with them before taking all the liability and risk and and and taking over you know essentially the payment process on behalf of that of that retailer. So they have lost control but I think it's it's a a fine balance and a fine line to walk when you talk about outsourcing you know essentially your commerce and your customer experience to these different platforms. Then you take so you take the AI component into it. Going to change things because essentially you're going to have this agent making these decisions for you and you're not really going to have a whole lot to do or to be able to do to control that whole shopping experience. So, in my opinion, it's really going to be about can you commit to say the inventory levels? Can you offer the best pricing out there? And what does that experience look like? Because that's really all you can control at the end of the day. So, I guess what I'm trying to say is you're giving up control, but you're trading that for conversion.
>> But what does this mean sort of that SMB midsize large retailer? Cuz I think a lot of a lot of the talk have been out there around sort of why this is great for SMB because their products will be as discoverable as the products being sold by major retailers. Does it actually change anything or is just the economies of scale always win here?
I think the little guys are going to get squeezed, the SMBs and the medium-sized businesses because this isn't going to be a free service that's offered by a platform. They're going to charge for this. And I think I think the the little the littleer merchants, whether that's SMB or or midsize, are going to get are going to get crunched.
>> Interesting. That's absolutely the opposite that that some of the other shopping platforms are saying. So that that's interesting to hear. So, what's the most expensive mistakes retailers still make when selecting payment providers?
>> I've seen this again play out in real time. I think retailers going after um the headline pricing and they're they're they're looking to shave off a couple basis points in return for great pricing, but they're not taking a full 360-degree view of their entire payments structure and organization and how things function. That's the one I've seen probably happen the most. And I usually can see this by when I'm sitting down with a retailer or a merchant and they're almost telling us that we're too transparent in our pricing because they've been burned before. So, I think that's that's one of those things that I think that's where merchants need to really understand the the trade-offs. And the trade-offs are things like auth rate potentially, I guess, but also um you know, some of the legacy providers are still requiring you to log into sort of websites and and download reports and do things in very manual ways that often merchants don't don't expect, especially if they've come from a very modern provider and and then they get offered a better price and then realize that actually they're going to need six more people to deal with, I guess, legacy legacy solutions.
>> Yeah, correct. Almost buyer's remorse.
>> Yeah.
>> You buy You buy the car you've always wanted for a long time and you get in and drive it after you've already purchased it and you've got some some remorse that occurs.
>> Yeah. Yeah. I I had that with an Alfa Romeo. It just didn't fit. If you were going to optimize one metric, would it still be author rates or is other metrics becoming more important in modern commerce such as resilience, checkout speed, customer lifetime value, operational efficiency? What's the the number one metric?
>> So, I'm pretty I'm pretty passionate about this and what I've learned over the years is that auth rate alone doesn't just pay the it doesn't pay the bills. I think your more accurate number you should be looking at is your successful revenue capture.
>> Okay.
>> So what that is is it's take into account your auth rate still because that's still an important number but you want to look at your what your retry success is. So think about in a in a multiple acquirer environment what your fraud false positives are. That's also an important number. What that end-to-end conversion looks like and then really what's your uptime and resilience. So I think if you take you take those five together that's a super accurate number.
>> Yeah. A mistake we see a a lot of merchants doing is they look at the top line rate figure and go it's terrible and then you dig it right and often it's the same person trying over and over and over again. So it reduces your overall auth rate, but your conversion is actually fine because uh it's just the same person who's perhaps not getting any money out of their bank account or you know worst case they're afraid to try stuff. But once you dig up data, you actually end up with a much better calculation. I think conversion rate for me is is much more powerful. And then you know if you are billing someone on the 31st of the month, right, it's they don't get paid till the first usually. So maybe try tomorrow, right? Or try the day after. Really simple things like that make a huge huge difference.
>> Correct.
>> So, everyone's talking about agentic commerce at the moment, but what actual infrastructure changes become necessary when software starts buying from you instead of people? What what what would you need to change from an infrastructure perspective?
>> So, you definitely have to have an API-first architecture. So you essentially have to if your if your architecture isn't API-first, it's going to need to be because that those AI agents have to be machine readable and that it sounds that's that's a a pretty long pole in the tent. I I will say at the same time you have to have your your payments have to be programmable. Your identity like when I think of identity I think about like the wallets like the wallets are almost becoming the the credential like that trust credential. You've got to have real-time fraud signals almost end-to-end decisioning that's accurate and that's you know doesn't drop to a manual queue but is is decisioning in real time and then you almost you have to have I'll call it deterministic fulfillment. So, you have to be able to understand when that has to be fulfilled and follow through with the with your your your customer and making sure that your fulfillment times and delivery speeds and reliability are all in line. So, there's there's quite a lot to do.
>> Yeah. I think I mean the the biggest thing for me is data really. Is your data ready for it? And I think that comes in a in a few different places. I think from a front-end perspective, is your inventory data good enough? Right? Is it up to date? Is it available? Is it even in a format where you can get it into an agent? And I think a lot of people are are looking at that area at the moment cuz in order to do agentic, you need to be able to surface your your data. And it's been a problem in omnichannel for years like where you know you end up having to end a season sales for items that would have sold out or they're online cuz they just happen to be in a store. And so you see see that happening and I think then your data's got to be able to deal at an incredible speed of delivery, right? These agents don't operate at human speed, they operate at agentic speed. So, if an item's on sale, being able to update your data on availability really, really fast, it could kill you. Like, if you sell the same item four times and you don't have stock, then it's going to be an awful mess to try and roll out. And that could happen in milliseconds. So, I I do wonder whether a lot of merchants are ready for the speed of agentic commerce and what that will actually mean to the infrastructure. And, you know, a lot of these things actually look like denial of service attacks because they're so cool. Can tech can cope with it?
>> Yeah. Yeah. For years, for years we've been trying to to to stop those denial of service attacks and now
>> we're welcoming with open arms.
>> Yeah. Okay. What does a payment stack designed for AI-driven commerce actually look like?
>> So we're we're building out our AI payment stack right now and we're getting ready to build it out. So on the on the I'll say the on the payment side the orchestration component is super important. Being able to do dynamic routing and even have redundancies built in is super important in in the AI native world. Obviously the identity components as well as the the tokens are also have to be you know kind of reconfigured to your AI native languages that you're going to have. Risk and trust. I talked about that a little bit earlier. I'm I've been a fraud guy most of my career. I love fraud. Love talking about fraud. And then you have to have this API API native kind of headless commerce. So gone are the days of of obviously with an AI uh with an agentic setup. You're you're not going to be going through all these different pages. You're going to it's going to be you click and you buy and that's and you're done. So you have that headless and invisible commerce component.
>> Yeah. So retailers obsess over merchandising, pricing, marketing. Why do you think payments are still dealt with as a reactive def function despite like aim cap revenue so heavily? Why is it always the last thing you think about when you're building out ecommerce site?
>> Yeah. So I I'm I'm an older guy. So with with you know being being an older guy comes some some knowledge. So I I think there's there's three things I'll talk about. There's these organizational silos that exist today. So you've got finance or operations as of kind of you know your payments organization is sitting under um payments need to be product-led for sure and I also feel that the second thing is pro payments are invisible and when I say payments are invisible they're looked at more of like a cost center and nobody knows payments exist until something goes wrong or there's there's an outage. We've had outages well when I was on the merchant side before and it's just it's just not a good place to be at all. And then lastly, I think payments need to be looked at as a as a revenue driver and not as a cost center and just a a cost of doing business. It needs to be part of your infrastructure that helps generate revenue for your company.
>> And I think like I think buy now pay later has shown us that actually payments can be a revenue driver, right? Because suddenly people can afford things or >> feel they can afford things that perhaps they couldn't before. And like we've seen experiments where if you move the payment button further up the stack at least the the payment knowledge further up the stack then maybe people can buy you know you get go to the 45-inch TV and then it goes well you know if you if you pay this off in four then you can afford the 72-inch TV right like and so the op options to sell upsell people and give people the flexibility might you know I think actually payment shouldn't right be at the end of the stack. In many cases, it showing people what they can afford actually potentially increases the sales. So, um I think that's one of the areas that we're going to see a lot of fights between the wallets going on because they all want to get further up the stack. But I think merchants are a lot of merchants are missing a are missing a trick there where they could probably increase, you know, cart size or um cart value payments as a marketing tool.
So Google announced universal checkout or universal cart trying to make checkout vanish completely, right? So you're in you're in in your your your chat with Google in Gemini and and you can add two items from this retailer, four from that retailer and press go and and and off you go. Is that future merchants want?
>> So I think there's real upside to this. Merchants will experience, you know, higher conversion rates, faster experiences, um, and less friction. But I think that what merchants don't realize is they're losing that differentiation. In my opinion, nothing makes them unique if they're just going through through Google like you mentioned. There's also loss of data and there's loss of, I think, insight into the the the lifetime value of that customer in some situations. Um, so again it comes back down to do you want the data to make informed decisions or do you want the conversion that is usually offered by going through a platform because really a platform you know it's not super unique to your brand or your your the experience that your customers demand. So I think it's an interesting interesting topic for sure. I mean they're I mean to use since we started on cars earlier to use the car example, right? You you walk into a car dealership and you and they ask you how you're going to pay and you say cash and they're a lot less interested in you, right? Spray it away because they make more money on selling loans, leases and and so and the same applies to some parts of retail like in white goods retail for example. you know, they make more money selling you the warranty on the on the washing machine than they do on the washing machine, right? That goes away that goes away with universal checkout, right? So, fundamentally, the economics of some of these businesses are actually in danger and I'm not sure people are thinking about that.
>> Yeah, totally agree.
>> Okay, so we've told merchants for decades now, customer experience is king. Customer experience is king and checkout. like optimize your checkout, make sure it's smooth, don't put anything in their way. And now we're being asked to outsource checkout to a partner, right, to a platform. What changed here?
Again, I'm using the experience I've had in the past. I mean, and I I want to kick this over to you too once I kind of get through this first part, but I think the compliance component of payments fraud, dealing with fraud chargebacks and non-fraud chargebacks, and then really the localization of, you know, the payment methods and where you're doing business in has really shifted things and it can be brutal to keep up. So merchants can say, "Hey, it's a lot easier for us to go to experts that know this business and let and we can just focus on selling our product and making our customer experience the best it could be." So I just wanted to see what your thoughts were on this one cuz I it's an interesting question.
>> Yeah, I think for me it what changed it in some checkouts being run by some, you know, platforms there is definitely mixed incentives, right? So if they've got their own payment method which is more profitable to them then they're going to preference that payment method over the others because that's how they're making their money right so and I'm not going to name names but there's companies out there and I could think merchants were a little bit suspicious of that like why is this button appearing behind above credit card button when it costs me more money to process than credit cards um and why would I use your checkout if that's the case I think what's happened is companies like ourselves has come along and and we neutral we don't care. Um so so you know it's however the merchant wants to configure it for themselves without having to to deal with someone else's business model around payments. And I think that's that's fundamentally changing things for the better. I mean giving the merchant control of the checkout without having to build the checkout themselves means that you can preference towards whatever works for you. And we've seen merchants that have, you know, promoted debit, promoted open banking in Europe, promoted other payment methods when the risk is low, right? So if the item isn't needed, it won't be delivered for 30 days or shipped or whatever and you can take a prepayment on it, then why wouldn't you preference a payment method that's going to be better for you as a merchant? But if you know, you got to ship the item immediately, you want to you want to bias towards a payment method that gives you um liability shift and a risk and all the rest of it. So I think part of it is is neutrality, right? And being having a neutral platform rather than the incentivized platform. And you look at what's going on with Agentic right now. And it's a it's a power struggle, right? Behind the scenes, it's power struggle of who's going to essentially be the Agentic wallet of choice. And you've got, you know, Stripe very much saying Link is the way forward and that's going to be the agentic choice. And then you've got Google saying it's Google Pay. And you know, we're waiting to see and Visa/Mastercard obviously saying it's cards and and through what we're doing. So that's an interesting power struggle that's happening right now. And I don't know if any of them are the right things actually. I think for for human in the loop payments, yeah, well, it's definitely handy and probably safer, but credit cards have done a great job in creating a framework that makes it equally as good, if not better. But then agents don't need to use those old school rails, right? That agents are going to be making transactions with multiple parties and chains of payments and thousands of payments happening in microseconds across agents paying agents paying agents and like that's the payment system that we use today for humans. Is that the right payment system for a future where which looks much more like a trading platform like what you
>> Yes.
>> Correct.
>> And and so I think that's going to be fascinating where we're going to end up in a system which maybe it's stable coins, maybe it's not, it doesn't really matter, but you're going to have a big ass ledger that's working out who's paying who and it's got to work really really fast. So yeah, good times. I'm I'm looking forward to it.
>> Yeah, for sure. This is an interesting question because it's more of an international question than a US only but it's around click to pay which so the the Visa Mastercard Amex Discover program called click to pay where they're essentially getting rid of the card number and the field where you type in a card number and we've seen great success in in Australia so far and we're starting to see great success in in Europe. If you're rebuilding your payment stack today, would you would you be looking at click to pay and thinking um I should definitely architect around it or or do you think it's never going to happen in the US?
>> I I think it I think you said something key there is is international and how this is adopted internationally. Um, I think it's heading towards being mandatory in the different card ecosystems and it's it's pretty I mean it's a good thing because you've got you've got the network-based identity, default tokenization and really that that cross-merchant portability which is all super powerful. So I think once it gets adopted in the US it'll take off. I just don't know if it's going to happen anytime soon.
>> Okay. What's the biggest checkout innovation everyone is talking about that you think will fail? And what's the quiet infrastructure change nobody is paying attention to that you think will matter more?
>> I I wouldn't say I wouldn't say dead actually. Can you read that question one more time? I'm sorry.
>> Sure. What's the biggest checkout innovation everyone is talking about that you think will fail? And what's the quiet infrastructure change no one is paying attention to that will matter far more than we think?
>> I I wouldn't say fail. That's that's a strong word and I don't like to to use that word. I'll say over I'll say overhyped.
>> Yeah.
>> I think universal checkout buttons are pretty overhyped right now because I don't see a single winner emerging anytime soon. And no matter what website you go to or where you shop through, whether that's through a Shopify store or another place.
>> Yeah.
>> I think the checkout button that universal checkout button is pretty overhyped.
>> There's just so much competition with it.
>> Yeah. I mean, there's a big wallet war on the way for sure. Like relevant and you've seen >> some people's stock prices not doing well because because of this war. So
>> yeah.
>> Yeah. Interesting. And then I think that the the really quick win uh that's that looks very favorable or that's considered underestimated is network tokenization. So in every implementation that I've been a part of, it usually gets higher approvals, the fraud is lower and you've essentially got better control. So I think it's I think from an ROI perspective, it's it's probably one of the the highest, you know, ROIs you can you can get today.
>> How about them? uh you know it I think it remains to be seen it's suit I it's I think it's very important in certain verticals depends on what you're selling where you're transacting at who you're transacting with I think I think there's more to be I think there's more to come on vamp.
>> Okay. 5 years from now, what part of today's payment stack will feel as outdated as a fax machine?
>> So obsolete if I had to look into a crystal ball today I would say single acquirer setups. Going back to your first question, it just that's not that's not the world we live in anymore. Uh when you're set up with a single acquirer, you have no redundancy. So you're leaving approvals on the table. You've got, you know, you're even with your dynamic routing, you're leaving approvals on the on the table as well. You've got weak global presence. And probably one of the biggest things is if you're in with a single acquirer, and I think I think merchants will catch up to this, you have zero leverage. So you can't use competition in your favor. You're just you're locked into a single acquirer. And along with that obviously comes the static routing. So you're just, you know, send it here. If it doesn't work, then it fails and you're you're done. I think stored PANs are still out there. That's going to be a fax machine soon. Um and I talked about it earlier about these the manual checkout form process. So you're going through seven steps before you check out. That's going to go away. That's going to be cut down to pretty much zero I think within the next few years. And then the standalone fraud tools. Again, I talked about being a fraud guy. I think I think standalone fraud tools, I think your your fraud is all going to be built in line with that entire process. I think gone are the days of queuing up orders and and waiting for a review to happen. I think that you can see the fraud landscape has changed quite a bit just over the last 5 years. So, it's really going to be about identity-driven checkouts and and real-time decisioning.
>> Very good. Well, Michael, really appreciate your time today. It's been been a great conversation. And any for anyone listening, please uh you can follow us on on the usual podcast channels. Thanks for joining us.
>> Thanks for having me, John.
>> Cheers.