rprend

nooooo zenobia is an ancient queen

oops

Yep you get it Bruce. Unfortunately I confused being a villain (everybody likes to bash visa) with being bad at their jobs and disrupt-able.

I appreciate you looking out for my well being, but really there was nothing shady here. We weren’t forced out by a Visa mob or something like that. I actually wish that were the case, because that would mean that we were growing! We just aren’t well connected enough to convince established merchants to switch over to a startup’s payment network, especially since the value to them is dubious.

But it’s not like there was some shadowy Visa conspiracy. We received pre seed investment from institutional investors and built a pay-by-bank network entirely fine without anybody stopping us

it's a super dangerous tarpit, because it has what id call theoretical product market fit: when you describe it to people, they're super excited by it. Everybody says that they want it and they encourage you along the way. But when it comes down to it, nobody wants the solution

> The incentive for merchants to accept cards for payment is that it'll increase number of sales this is something that ive thought about a lot, because while it is strictly true in the short run it may not be in the long run. For example, i don't have any debt, but i use a credit card for everything. Why? It's become my default to use it.

I wonder if the same thing will happen with BNPL (Klarna, Afterpay). These are higher fee than credit cards (5-7%) because they bring in new customers. But, like with credit cards, savvy customers are starting to see BNPL as interest free loans (aka free money on the float, even better than credit card rewards), and it's possible that they become the new consumer results. Merchants are left holding the bag of paying 6% processing fees for everyone, even people who can afford it.

Where do you get 5X? European interchange is regulated to .3-.4%, the network fees are the same as the US (.1-.2%) and then the payment service provider takes their fee. Online card processing in the EU costs around 1%.

Since February, Teddy and I have worked tirelessly to build a Visa+Mastercard competitor. We quit our jobs and built an instant clearing, mobile first, pay-by-bank network. However, we failed to get any adoption besides thieves and mobsters.

A couple weeks ago, we decided to pivot away from Zenobia Pay. Today, we are open sourcing all the code we wrote: the core payments service, the merchant dashboard, the ecommerce integrations, the transfer status websocket service, and the iOS app. It is all free and MIT licensed.

To go live, you will need a fintech banking partner, so this isn't a one-click-deploy. That said, I hope it will save a lot of time for somebody who wants to try where we failed. I've seen a few HackerNews commenters talk about this recently, like https://news.ycombinator.com/item?id=44676559.

In the blog post, I go through the different customer iterations we tried and the problems we faced with each. If you just want to see the code, here are the links: - https://github.com/zenobia-pay/ios - https://github.com/zenobia-pay/zenobia-pay - https://github.com/zenobia-pay/core

I think it’s important to note that this is asking rent, not what people are paying on average. Landlords do tricks to make rental history appear more expensive than it was, and every online posting should be read as aspirational price, not a real one. Listing inflation helps increase price expectations for the future, and has helped drive the crazy rent inflation NYC has seen post covid (exceeding other expensive metros like SF).

For example- my current apartment had a crazy high lease concession. The paper rent is 30% higher than what I pay. The only inconvenience is that i had to pay a security deposit at the paper rent, not actual rent. It’s like a software business giving 2 months free then multiplying MRR by 12, not taking churn into account. I am moving out after a year.

Pix, UPI, etc are nationally silo'd and have little risk of spreading around the world. They took over their respective countries because they were government backed and mandatory.

The US has its own version of this: government backed, 24/7 instant settlement payment rails (called FedNow). It released in 2023 but adoption is slow because it's not mandatory. Banks are especially slow to support "Requests for payments", which might compete with card networks for retail purchases.

It's the banks, not Visa / Mastercard, that make the most off credit card fees. Visa / Mastercard take .1-.2%, the bank takes ~2%.

The other unique bit about the US is we have uniquely strong consumer protection laws and a uniquely large amount of fraud. Irreversible payment rails are a nightmare for consumers who are used to chargebacks being built in to the payment rail, and banks are skeptical of exposing such rails to consumers because most of the time the banks are required to make good when the customer is defrauded.

it's important to note that Pix took over Brazil because it was mandatory. The US Federal Reserve built equivalent infra: FedNow, as of 2023, is the rails for instant settling, $.04 bank transfers. But adoption is meh

Ya we looked in to the wallet fund model and while it solves fraud concerns, it introduces new frictions (funding the wallet) that makes growth tough.

Cashapp pay is leading in this department, and since they own the POS terminal Square they’ve actually been able to implement QR payments in some SMBs. They get terrible adoption, though

Stablecoins do seem to address all these concerns. USD backed, and I believe Visa/Mastercard are adding stablecoin support so that there will soon be stablecoin debit cards.

Our thesis on crypto was that American consumers will never pivot to a foreign currency to do retail purchases. That’s just too much friction. It’s possible we get proven wrong, though

I’ll talk to my cofounder about open sourcing.

In payments, though, the code isn’t the hard part. The hard part is the licensing/regulation and most of all the risk analysis. You need a partnership with someone with an MTL to move money; you need anti money laundering policies and sanctions screenings and to do KYB checks. These days that’s not that hard because there’s vendors for all that, but it’ll cost you low five figures and you need to go through an enterprise sales deals that take a month or two.

Even with all of that, you’ll probably still lose tens of thousands to Russian fraudsters who present as a perfectly legitimate American business.

Payments is a tough business

The merchant would disagree, but the merchant is lying to themselves, haha. The tough lesson i had to learn was people’s stated preferences often don’t match their revealed preferences. I really wanted it to be true so i gave them a ton of benefit of the doubt, but still lukewarm response.

Regulation on interchange worked in Europe. The other option is government mandated support for bank transfer rails. That’s how UPI in India and Pix in Brazil took off- they were mandatory.

The US has FedNow, but it’s push only (not user facing, too high friction for checkout). RfPs are there but still low adoption.

I’m still convinced that someone could play our playbook and it would work. Just someone with deeper pockets and deeper industry connections.

Luxury is struggling in 2025 and is losing to the resale market which is growing 3x faster. They’re especially struggling with gen z. Build a branded luxury-only payments network that owns the category “buy to resell”. Attach purchase metadata to pay-by-bank identity for digital proof of purchase.

You have to play it right (digital proof of purchase generates social media posts and a verified stamp, ie buying it to flex). But that way you can convert young customers from rewards cards AmEx etc without having to compete on rewards

We tried with SMBs (it was our original focus, low net margin businesses). But 1. Deployment times are way worse in person, 2. Merchants still actually didn’t care (they said they did! But there was no strong need). 3. conversion at checkout is lower for small tickets than high tickets.

SMB business model just didn’t work, with tons of white glove customer support needed for a customer we’d make $20-$50/month in fees

With high ticket sellers the purchases are high intentionality enough that you can convert customers.

This is ecom, not in person, and we found these places are brand agnostic about payments— they don’t control the look/feel of PayPal, Klarna, etc. And for every Hermes, there are a dozen luxury merchants less anal about the brand.

White labeling the payments infra as “their pay” did intrigue them, but not enough

Haha i wish i were just a bean counting manager. A number of venture backed startups have tried exactly this idea, and it was a tarpit.

Confused about your hate of acronyms. The bank rails are a practical limitation on how effectively a startup can compete in this space. Do you feel the same hate when people use technical terms about coding eg “we deployed with Kubernetes on AWS”?

Speaking as someone who tried (and failed) to build a competitor, merchants don’t actually care about the fees. We built a mobile payment system using bank transfer rails (ACH debits and FedNow credits) for ecom platforms. QR payments, just like AliPay, but for the US. We studied past failures at this same idea, and targeted specifically high ticket items (jewelry, luxury), because the fees are a bigger problem, and we could solve other problems with pay-by-bank like limiting botting/scalping and providing digital proof of purchase for resellers.

We raised some money, built it, and demo’d it to dozens of different business. Instant clearing, way cheaper fees, bundle fraud insurance, digital proof of purchase. But it was an uphill battle- people didn’t want it.

Merchants might say they care about the fees, but in practice they care about conversion and getting sales. If you want them to act you need to promise to bring in new revenue, not reduce their cost of existing revenue.

Second problem is the rails. Your only option to debit is ACH, and ACH’s can be returned by the customer for any reason up to 60 days after the debit. Even with return-risk profiling, we had issues with Russian fraudsters using stolen SSNs to onboard as a merchant, do a “custom jewelry” sale for thousands, then as soon as the merchant gets paid out, ACH R10. The “merchant” has RTP’d that money away so when we try to debit it NSFs.

RfPs on FedNow (if they ever get widely adopted) would solve this, because at least they’re irreversible. Until then the only competitor is crypto, but good luck getting US consumers to use a different currency at checkout without tanking your conversion (people have tried!!)

I suspect people wouldn’t actually like the alternative. An alternative world where it’s relatively easy to get an MTL, or consequences for AML aren’t huge, looks a lot like crypto. Imagine banks and money processors doing rugpulls or going bust.