Building Stablecoin Payments at Scale

Transcript

Hello everyone. Thanks for being here. My name's Justin Ghazely. I lead stablecoin payment acceptance at Coinbase. I've been doing this for 4 years, which in crypto is fairly long time, but in payments is quite short. But I want to talk about what we've learned building stablecoin payments at scale selling to enterprise and payment service providers. What we've learned along the way and what what I still think we need to accomplish. So, I want to start by zooming way out and sharing this Reddit post from Coinbase founder Brian Armstrong. It's actually Hacker News, not Reddit. But in the post, which was in 2012, before he

applied to Y Combinator with the idea for Coinbase, he's writing that he thinks he can use crypto to remove interchange, to remove that 2 to 3% tax that Visa and MasterCard have on every payment. What's so cool about this is before he even had the idea of buying and and selling Bitcoin as what would ultimately become a a massive company, he was thinking about crypto as a payment method. And so, crypto as a payment method has been ingrained in Coinbase from the very beginning. And when I took on Coinbase Commerce 4 years ago, the mandate was effectively the same. It was how can we make crypto payments work at scale? And we tried an

incredible amount. We tried Bitcoin payments. We tried doing every token on every chain from every wallet. We leaned into meme coins. We tried all sorts of cashback experiments. We tried point of sale, invoicing, online checkout. But none of it really moved the needle. And for a long time, there was a period that we weren't sure if crypto payments were ever going to be a big business. Sure, maybe a niche startup would serve the demand, but at Coinbase's scale, was it really a viable business to be in?

And we just had this worry that maybe crypto wasn't right for payments. But the signal was there if we listened closely. We spoke to hundreds of buyers and merchants and asked them, "Hey, why are you using this? What What excites you about this? What are you trying to get out of it?" And as we listened closely, we kind of found there were a couple of different camps of people. Some of them wanted to spend their gains, some of them were just really excited about Bitcoin, but for the most part, there was a group of people who were really excited about crypto as their cash, as something that they're earning into that they then actually want to spend. They wanted it to be stable. They didn't want to deal

with gas or volatility. And what we found is maybe more assets didn't mean better outcomes. And so we made a bet. We stripped out Bitcoin, Ethereum, all the volatile assets, and we went all in on stablecoins. And that may sound obvious in hindsight, but I'll tell you at the time, it was absolutely not consensus. This actually led to a ton of backlash. Even internally, we were debating. A lot of us didn't believe that this was the right move. Right? People thought, "This is antithetical to what we're trying to accomplish as a company."

But we were hearing from our customers that this is probably the direction to go, and so we wanted to take that bet. And the counterintuitive result was that all of the metrics went up. Conversion went up, volume went up, merchant satisfaction went up. Even though we no longer had 99% of the crypto market able to pay. And the reason for this is that what people actually wanted was to hold digital dollars and spend digital dollars without fees.

And instead of building a crypto product, they really just wanted to use a payments product. And so that was the turning point for our business. And another stat that I think really kind of nails this in is that 64% of the wallets that paid on the stablecoin only product held no other assets. That's fascinating. It's a fundamentally different cohort of people than who was using our Bitcoin payment product before. These aren't crypto traders using their dry powder to spend. They're savers, people who are using stablecoins a way to save cash, to earn.

It's a fundamentally different group of people. And I think that this customer group really didn't exist a few years ago. This is a very recent phenomenon. So, I want to talk about what we actually built. I want to start by telling a story. So, we made this pivot to stablecoin only uh and we built what we thought was really the best stablecoin checkout in the industry. You could use pretty much any crypto wallet. You could use it on any chain. All the major stablecoins. No gas ever on any on any uh token. Um and so we were really thrilled about this. We thought it was the best. And so we brought it to Shopify who's a partner of ours. And we're really proud of it. And

so we get their whole team into a room and we say, "Hey, we want to show you this new thing that we built. We want to bring it into your checkout." And I see the hand raised from Shopify's global head of payments and checkout. And he says, "Well, what's your plan for off and capture?" And as embarrassing as it is being someone in payments, I had never heard those words before. I was like, "What's that?" And what followed was a year of working with Shopify as a really fantastic design partner to understand everything that a real payment system would require.

So, often capture, the idea that you can separate holding the funds from the receipt of the funds. So, you think about if you check into a hotel, they swipe your card, they put a $1,000 hold on it, and then you can take from the mini bar, and they'll pull only the amount they need at the end of your hotel stay. And so, that's often capture. They also needed automated voids and configurable expiries and refunds and different settlement schemes and different reconciliation reports and the ability to handle disputes. And so, all of these things were prerequisites for a PSP or enterprise to even consider stable coins, and the checkout experience was just a fraction of this overall puzzle.

And one of the things that we learned in doing this was just not to underestimate how much thought has gone into the traditional payments system. We love to say that it's broken and extractive, and of course, parts of it are. But, decades of thought have gone into building systems that work for billions of people. And if we want to have payment service providers and enterprise take stable coins seriously, we can't ask them to change each part of their payment flow. We can't ask them to change the reconciliation or rebuild the reporting.

We have to plug into the places that they already exist and make it so easy for them to say yes. That's our wedge. And so, we built the commerce payments protocol. I know some of the audience is technical, so I thought I would dive into the interesting smart contract that we built. Um and what we did is we shipped this escrow protocol on base. It's fully open source, and it powers Shopify's USDC checkout that's now live on hundreds of thousands of sites. And so, the core is this escrow smart contract. And on authorization, funds move from the buyer into escrow.

On capture, funds move from escrow to the merchant. The contract enforces that once funds are authorized, they're guaranteed for capture. So, we had to build this novel escrow on the blockchain to make it so that there's no double spend problem. Businesses need to make sure that they can give the user a check mark and ship the goods. And of course, there's no assumptions here about the wallet type or the receiver, which allows us to be super flexible as we continue to scale the product. We also created this idea of an operator. So, in traditional payments, someone has to submit the transaction, typically the merchant's payment processor. And we call this role the

operator. Anyone can build an operator. Uh and it's effectively giving the operator the keys to the car without the keys to the house. And so, they can facilitate each of these key actions. They can't modify a payment intent. So, the payer signs over this hash with the payment details, the receiver, the token, the amount, the expirations, and any mutation would produce a different hash and the contract would reject it. The operator also can't lock funds. And so, if the authorization expires and the operator hasn't acted, the payer can reclaim them directly on the contract without any intermediary.

And of course, operators can't impact other operators. So, a bad one can only hurt their own set of merchants. And the whole protocol is just six operations. Authorize, capture, charge. Charge is the ability to combine off and capture into one step. Void, which is moving funds back from escrow to the buyer. I'll jump ahead, refund, which is moving from the merchant to the buyer. And then reclaim. Reclaim's a really interesting one because it's pretty blockchain native. The idea is that the buyer can pull their own funds back from escrow after the authorization expires. And this doesn't exist in traditional payments. And so, you can imagine if your credit card hold expires at that hotel,

you have to wait for the issuing bank to release the hold. And that could take days. But, in this case, if the operator chooses or forgets to send the funds back to you via a void, you can actually go to the contract yourself and reclaim your funds, which is pretty interesting. Now, there's a lot more we had to build to get enterprise to say yes, as well. I consider these the boring, but required. They want predictable fees. And so, they would not be happy if one day their gas fee is 8 cents and another day it's 12. They want predictability.

And so, we absorb all the gas fees, we flatten it, and we give it to our enterprise as a per fee fixed fee. Settlement files. Each payment service provider has their own settlement requirements. They have their own way to read in data and give that to their customers. And they're not going to be reading the block explorer themselves. And so, we built these serializers to make it so that each payment service provider can have exactly what they expect when it comes to reporting. Fiat settlement, of course, not everyone is ready to hold stable coins. And so, we are happy to deliver them fiat in their bank account every day on whatever schedule they'd like.

And working with their various entities. There's a lot of compliance that goes into payments, as I'm sure you can imagine. And so, there's a lot of this kind of boring, but required stuff that we need to do to make it just so easy for them to say, "Hey, I'm ready to dip my toes into stable coins." And the result is that we're just another alternative payment method. You get familiar APIs, consistent reporting, simple buyer UX. Their finance team doesn't have to do anything new. Their developers just can toggle it on in their existing payment service provider portal. It works just like any other alternative payment method that they already support. And that's a great thing cuz they don't want it to be a change in their system.

So, what do we still need as an industry? Well, maybe this is a hot take. As much excitement as there always is for new blockchains, new L1s, new L2s, faster finality, higher throughput, I think the blockchain is pretty fine for payments. Now, of course, there is need for investment. But, base already settles in 200 milliseconds for less than a penny. And for the most part, the chain is not the problem. For us, the things that we're seeing is that it's the wallets, the on-ramps, the signing UX. These things that were built for the crypto trader sitting at their desk, not the individual who's saving their life savings in

stable coins, who's earning it to stable coins, and wants to be able to spend that to pay for their life. And so, there's a few places that this kills us today. Wallet connections. Unfortunately, some of the wallets with the most assets actually have the worst UX. They can have really flaky connections. I attribute this to the the same as being a declined card. If the wallet connection drops and they can't spend, they have to go back in, try again, we're going to lose them. On-ramps still need to be better. A meaningful percentage of people want to get more money into stable coins. It can be very challenging to do that outside of the US. And so, there's lots of progress we can make to make the on-ramp experience better.

Signing UX. Raw hex signatures are not what individuals who want to save their life savings in stable coins want to see. We need to make it easier for people to understand what are they doing and avoid unnecessary friction. And then, I think smart wallet adoption, or just adoption of new standards in general, we need to push our wallets to be better. These are ecosystem problems. They're solvable. I don't think it's necessarily that we need to go raise another billion dollars for a new chain, but rather really lean into these pieces that sit around the wallet and the payment infrastructure.

I'll end on a bright note here, which is I think 2 years ago today, no one was really talking about stablecoin payments. They've gone from this curiosity to something that is in production now. Millions of businesses, millions, hundreds of millions of individuals are holding stablecoins and using them for as a better rail. All of the primitives exist, which is amazing, and the demand is real and it's growing. I think we have our work cut out for us, but it's very encouraging to see how far we've come in such a little time.

With that, I want to thank you all for being here and for listening today. You can check out the Coinbase payment suite at coinbase.com/payments. You can find me on Twitter at Kinsley Justin. I'll be around if you want to chat after the talk. Thank you again.