Read on Twitter
1/ Folks voted to hear about Stripe Treasury. This was one of the last things I worked on while at Stripe and it's an exciting and compelling product. But it's not the end of the BaaS arms race and its launch creates some interesting dead ends for Stripe.
2/ I joined Stripe in 2016. The company was in the middle of getting its U.S. money transmission licenses and wanted to get the work off the existing team’s plate. I was at Square, where I managed their money transmission exams and did a bunch of other regulatory odds and ends.
3/ I spent the next two-ish years helping Stripe with U.S. money transmission and other regulatory issues. It was, and remains, a pretty great place to work. Along the way, I worked with @shreyas @jay_ssh @JorgeO and others on the Connect team.
4/ Connect was going to be the wrapper for Stripe’s U.S. money transmission licenses. We spent hours and days and weeks working on what could be possible and how we could use the licenses to hold and move money for customers.
5/ Like most complicated projects, getting the regulated payments part of Connect took some time. But we finally got things launched that used the money transmission licenses. Today you can make payouts to over 30 countries using Connect.
6/ The work handled by the Connect team was growing more complicated, and eventually the team got split up to better handle the work. Around this time, a lot of us rotated to other projects in the org. Work continued on regulated money transmission products.
7/ In early 2019, my manager came to me and asked if he could rotate me onto some key skunkworks projects. I had a rep for high-quality, high-volume work. I could help a team identify regulatory paths to support product launches. Or this is what my manager said.
8/ One of the projects I came to work on was a bank account service via API - what eventually became Stripe Treasury. Working with the very talented @amfeng and @cjc, who were project leads, and folks like @athyuttamre and @tarstarr was a lot of fun.
9/ One of the decisions we spent a lot of cycles on was whether to do white-labled DDAs or build the product as a stored value wallet using Stripe’s money transmission licenses. The bank partner said we could do either. For the deal I worked on, we did stored value wallets.
10/ I eventually handed the project off to another lawyer. They had a banking background and had a hard time with why we had structured the initial bank partnership to support a money transmission wallet. They complained a lot.
11/ So fast forward to today - Stripe Treasury is live! It’s awesome. It offers the ability for Connect platforms to offer BOTH stored value wallets and bank accounts. And a whole host of other functionality. So let’s unpack.
12/ This is a slick product. Those of you at Square, Chime, BlueVine, Current, PayPal and my current company http://Privacy.com (we’re hiring, btw) know how hard it is to build a product on top of a bank, especially card and DDA products.
13/ Stripe Treasury allows companies to programmatically open accounts at Evolve and Goldman Sachs (I understand Citi and Barclays aren’t live), and set key features like interest rate. Accounts offer ACH and wires. And you can use Stripe Issuing to spin up payment cards.
14/ This is not a BaaS category killer. Shopify, FB, etc. weren't going to work with Bond, Wise, Unit in this space. Stripe, and in many cases Marqeta, already work with these companies. They have enough internal talent to build without BaaS. And can command really low pricing.
15/ Stripe Treasury isn’t a BaaS category killer due to other reasons - scale, human processes and channel conflicts. First, a detour to unpack.
16/ If you want to build a neobank, you need a bank sponsor, an issuing processor like http://Privacy.com , someone (you or maybe @usealloy) to do KYC/AML/sanctions, and someone (you or maybe @ModernTreasury ) to manage all your ACH and reconciliation.
17/ Each of these elements has tons of complexities. Working with a bank requires a compliance officer, financial partnerships hire, project manager and usually a really good lawyer. It takes 6 to 12 mos to launch a product on top the bank. And noobs get hosed on bank pricing.
18/ Similarly, issuing payment cards is complex. Issuing processors generally have super complex fee schedules, crappy maintenance windows and can be slow to work with. It’s why we built a better offering at https://privacy.com/card-issuing (with simpler fees and faster launch times).
19/ BaaS companies are built on the promise of making the bank partnership and issuing processor piece easier - dare I say turnkey. Companies like Wise and Bond are building a software layer for FinTech and embedded finance companies to more easily work with banks and processors.
20/ Stripe isn’t promising a better way to interact with a bank - they’re offering an abstraction layer. Abstraction is nice, but if you want to differentiate you’ll eventually want to get closer to the metal and that means working directly with the bank and your processor.
21/ This means Stripe isn’t really offering BaaS. They’re offering abstraction and bundling. Come for payment processing. Come for Connect. Come for Radar. Once you get scale, flip on lending and banking for new revenue. It’s Square in a box!*
*customers sold separately
*customers sold separately
22/ BaaS companies, on the other hand, are offering a way to reduce headcount needs. Use Bond, Wise or Unit and you can have one financial partnerships person instead of three. Two compliance hires instead of four. Maybe no need for an in-house lawyer.
23/ There’s ultimately a need for both sets of companies in the market. Stripe Treasury is also likely to have some market and operational limitations.
24/ Limit 1 - ops bandwidth. Stripe has done great things with Atlas. But have you ever had to get a product launch approved by a bank? Now imagine Stripe is having to get 100 different product marketing campaigns, each by different companies, approved by the bank partners.
25/ I think ops bandwidth and partner approval is the reason Stripe’s BaaS offering is invite only for now. Those in the industry know partners are picky. They won’t take everyone. They all want the next Shopify (they used to want Square, but that’s a dirty word due to the ILC).
26/ This is where the stored value wallet has advantages over the DDA account. Stripe - not its corporate bank partner - controls what happens with the stored value wallet. Partner branding and marketing can be approved by Stripe.
27/ That’s an Andy Grove Breakfast Factory waiting to happen, if Stripe can find the right lawyer and risk manager to run the process. But it's not a DDA, so it’s not BaaS.
28/ Limit 2 - channel conflict. Stripe offers its own term loans. So why would lending focused cos want to work with them? They offer their own charge card, so why would Divvy, Brex or others work with them? They have checkout + an investment in Fast . . . you get the point.
29/ The channel conflict also seems to cut Stripe out from the next level of channel partners - BaaS companies.
Shameless plug, if you are a BaaS co and want to compete with Stripe on tech, make sure to contact http://Privacy.com . We have lots of cool stuff to help.
Shameless plug, if you are a BaaS co and want to compete with Stripe on tech, make sure to contact http://Privacy.com . We have lots of cool stuff to help.
30/ Same goes for Stripe Issuing. Stripe Treasury creates another layer of channel conflict that makes the sales job for Issuing harder. There’s also lots of decent issuing processor alternatives such as Marqeta and http://Privacy.com .
31/ Limit 3 - Product. From the sample API calls, I suspect Stripe can’t support consumer DDAs. If they can, they’ll still have limits on oversight over customer service and consumer compliance. Who’s staffing the phone lines? How is the bank approving 100 platforms CX scripts?
32/ You want to know a secret? If I’m John and Patrick, I don’t think I care to solve those limits above. If I’m either of those devilish lookers from Limerick, I’m playing a longer game.
33/ If Stripe wants to really have its lofty $100B market valuation, it needs to make sure its revenue is coming from more than payment processing. SHOP has its software subs. PayPal has the margin from its wallet and on-us transactions. SQ has Cash and all the services revenue.
34/ Stripe has . . . payment processing. Atlas has to cost more than the $500 per corporation- lawyers charge $20K for that. Capital is hopefully growing, but likely early. Radar is nice, but a few bps and most customers probably get IC+ pricing on processing.
35/ Enter Stripe Treasury. There’s two really interesting places this can go. A PayPal-like wallet. And the network DOJ seems to think Plaid is building.
36/ First the wallet. If Stripe can get a lionshare of its Connect platforms to offer a stored value wallet, it can kickstart a PayPal like on-us ecosystem and also optimize to earn yield off the funds. PYPL has great take rates + BILLIONS in FCF. Not a bad thing to emulate.
37/ Was this why I nudged my product colleagues to pick the stored value wallet? Was the lawyer I handed the project off to too small minded to see this opportunity? I’m bound by confidentiality obligations, so I can’t say.
38/ The other tantalizing opportunity is to build a network of key bank partners. By connecting Evolve, GS, C and Barclays, Stripe will be able to programmatically open accounts and move funds between one key FinTech bank, two prominent U.S. HQ’ed money center banks and Barclays.
39/ What if you added Wells, who already works with Stripe on payment processing? And maybe super-hungry super regional PNC, who bought Stripe partner BBVA? Would that get another top 10 U.S. bank to work with Stripe? And what about HSBC and those outside of the U.S.?
40/ Suddenly Stripe can open accounts, know balances and help move funds into and out of all those banks. AND Stripe has a massive merchant network to move those funds over to. Last I saw publicly, some 80% of U.S. residents had shopped on Stripe.
41/ The interesting thing for Stripe is whether it can get to these destinations before late-stage brain drain occurs and/or before all of its new outside managerial hires force the builders in the ranks out. Even if they don’t, I’m still glad I’m a shareholder.
42/ One last point. Given the invite-only nature, and the bandwidth limits noted above, Stripe seems to be getting away from its indie developer roots. If I’m a 23 year old first-time founder in a Palo Alto garage, can I get access to Stripe Treasury? I doubt it.
43/ So to recap:
→ Stripe Treasury is awesome . . . if you’re a big co
→ BaaS is still going to be a thing
→ If you can't score a Stripe Treasury invite, you can still build your own neobank with APIs from http://Privacy.com
→ Stripe Treasury is awesome . . . if you’re a big co
→ BaaS is still going to be a thing
→ If you can't score a Stripe Treasury invite, you can still build your own neobank with APIs from http://Privacy.com
44/ Back to the top! https://twitter.com/regulatorynerd/status/1335841935521505280?s=20
