Fintech 🧠 Food - 21st Feb 2021 - Citi's $500m error, Marqeta IPO, Brex Charter, and why procurement is the enemy of change

Hey everyone 👋, thanks so much for coming back for more brainfood. A space to learn in public and hopefully process everything happening in fintech.

Thanks to the 108 people who signed up in the past week for this free newsletter.

If you haven't joined yet, you can subscribe right here 👇

We also have a ton more content over at the 11:FS Youtube

** Snap survey - What do you want from your partner bank? **


Weekly Rant 📣

In case you missed it, a federal judge ruled this week that Citibank is not entitled to the return of $500m it sent creditors last August in error. Citi intended to send $7.8m, but the complexity of what their client was trying to do combined with the confusing interface of the Software Citi used led the bank to accidentally pay back the entire loan.

Meet Oracle Flexcube 👋

This software runs a decent percentage of global financial infrastructure.

Citi gave the work of entering this transaction into the software to a subcontractor in India.  Like many banks, Citi requires three people to sign off a transaction of this size (lovingly called the "six eyes process," which makes it sound like a superhero collective instead of a stupidly inefficient process).  

A day later, Citi noticed they had sent out $900m, some creditors sent the money back, but others refused. By the time this got to court, the judge ruled that it was reasonable to believe a bank as sophisticated as Citibank wouldn't send this money by accident.

This judge has never worked in payment operations.  Most humans haven't worked in payment ops.  

There are three things to unpack here.

1. Shitty UX

2. Shitty Process

3. Shittier Culture & Procurement behavior

That UX

Good design hasn't come to the back office of banking.  

In the 90s, when we began building UX over Mainframes, the UX took its cues from the underlying system. So what you see are things that look like Microsoft Access database Macros. The software was written before the development of modern UX and software engineering techniques and likely by a person who happened to work on that team back then.  It's not the software's fault that the UX is terrible by 2021 standards.

When Flexcube was built, it was built to manage the complex back-office operations of a bank. That software had value outside the bank and was spun out into what we now lovingly call Flexcube before being acquired by Oracle in 2006. (It turns out Amazon wasn't the first company to turn a cost center into a profit center after all). But this history is essential.  When the bank built the software, it was an improvement on how things used to run.

It's important to emphasize how common this level of horrible UX is in banking back offices.  It is a testament to the poorly paid, often underappreciated subcontractors that this kind of headline isn't much more common.  Instead, we have to look at the processes and how they evolved.

That Process

A payment operations team performs several functions depending on the payment being processed. A high-value payment can include checking KYC documentation, that the payment is lawful, and is for the right amount and going to the correct beneficiary.  If you are a large corporation trying to make a multi-million transaction, you don't want it to go wrong, and this high-touch process is fairly standard.

Payment operations have evolved over decades from their original paper equivalent. Banks needed large payments ops teams in the pre-internet era when moving money was complex. Over the decades, banks have gradually improved these processes, brought in new software, and digitized those paper processes.

When you digitize a paper process, you also fix it in place. By making creating an efficiency improvement, organizations bake functional fixedness into their culture.  Functional fixedness is a cognitive bias that limits a person to use a tool or object only the way the person traditionally used it. This applies to organizations too.

Each process was built in a silo. Each silo was digitized over the decades with the best tools at the time. So what banks now have is expensive software systems that were never designed to talk to each other, with humans filling the gaps with spreadsheets.

The software was built to manage the existing process. The software evolved over time. The software is working as intended.

As 11:FS CTO Ewan explains, banks weren’t built to be platforms.

The complexity of changing core banking software is non-trivial. Core banking software is welded to thousands of other systems and runs trillions of dollars in notional and real-world value through high-value payments every single day. So the complexity of change, both perceived and real, is significant.  

Therefore when banks "improve" payment operations, they're making marginal gains in the process.  Back-office functions become a "cost to manage" rather than an area of real product design and genuine innovation.

Still, when the mistakes can cost as much as they did for Citi, why haven't banks changed sooner?

The procurement gap

Everything works against back-office transformation. The suppliers they take advice from (who monetize the status quo), how banks implement change and make decisions, and how they buy their incentive structure.

All of these are chapters for a book someday, but how banks buy is incredibly broken. Banks don't buy things; they procure the shit out of them. Procurement is the enemy of innovation because, much like the software banks use, banks didn't design bank procurement for the world we live in today.

Banks designed procurement for 10-year contracts with mainframe providers with 100s of services and line items from a single provider. Banks use a cut-down version of this process to buy innovation or from new providers. Some banks even have specialist fintech procurement teams, but they’re really hacking the old process rather than creating a new one.

Banks want much more efficient and digital processes. But banks consistently view their back office as a cost to be managed by the lowest cost bidder. Until that mindset changes, and culture changes it will be hard for banks themselves to change.  

In the past two decades, large organizations' spend on outsourcing is staggering (and not unique to banking). But when you outsource a problem, it got cheaper, but it didn't go away.  An outsourced problem is now harder to change.  What incentive does your outsourcing provider have to make your processes more efficient if they're selling your bodies to run your operations?

Wanting innovation and executing it are two different things.

The biases and assumptions banks carry run deep into the organization. No one CEO or executive can change them; they're baked into the very processes and software. But like any system, there are tipping points, and procurement is a big one.

If banks can change how they buy, they can change who they buy from. 

The past year has seen a wave of new fintech providers who can help you build any bank function better, faster, and cheaper than any time before.

The view that fintech is a shiny toy has evolved. Banks understand partnerships with B2B fintech providers are key to tangible bottom-line benefits, but this means evolving from a procurement mindset to a partnership mindset.

Banks need to let go of the bias that it's too hard to change the core. It isn't. It just requires new approaches and mental models. No more big bang / rip and replace change. Look for the tipping points. Look for the right help outside in.


4 Fintech Companies 💸

1. Generation Home - A better mortgage lender (UK) 🏠

  • Generation Home is an authorized mortgage lender in the UK. That in itself is special.  They offer some neat pre-mortgage tools like income booster (adding a friend or family's income to the mortgage to boost what you can borrow) and deposit booster (add a friend or family member who can choose to be repaid or gift the deposit). 

  • Behind this Generation Home has live chat, 100% digital applications, and a smooth mortgage process UX. GH adds neat features like flexible over or underpayments (which historically carry a fee). Oh, and they automate remortgaging. 

  • GH is the most customer-centric mortgage product I've ever seen.  Having just exchanged on a house in the UK, the process took us nine months, and the level of paperwork and admin was excruciating. If they can build a balance sheet and fund their mortgage book, they will do well. Suppose you're in a bank reading this? GH is the benchmark for mortgage UX.

2. Penny - Pension (401k) aggregator (UK) 👴

  • Penny provides a beautifully simple UX for finding, aggregating, and managing your pensions into one app.  It works for roughly 92% of UK pension schemes and the homepage. 65% of Gen Y move jobs every two years, 84% expect to take a career break (for kids, charity, or travel). Gen Y will work long past the age of 65, with only 3% retiring before 50. 

  • If you change jobs, frequently your pension's capital is not compounding as efficiently as it could.  Bluntly, this means less money for you at retirement. Unlike in the US, where the self-directed 401k is the norm, most UK employers manage pensions on behalf of staff. The result is a hidden paperwork nightmare and worse outcomes at retirement.

  • With reports that Pensionbee is about to IPO with more than 65,000 customers and £1.2bn AUM, Penny is playing in an exciting space.  The design is stunningly simple, and the clarity of their marketing is excellent. It's a notion page as a design style.  The fees appear relatively high on this type of tool (e.g., Pensionbee and Penny), but if you're compounding, they could be worth it. 

3. MoneyMap - An API for human, financial advice (USA) 🤓

  • MoneyMap adds financial advice that can be embedded by fintech companies, banks, or wealth management firms into their user experience. The UX helps with goal setting (like many Robo platforms) but supports that with certified financial planners. 

  • For decades we have assumed financial literacy is the answer to money troubles, but the reality is it's often a better product.  But the role of humans in banking isn't doing what machines can. The part of the human is often empathy, problem-solving, and overcoming an emotional barrier to progress. A machine can help you find the right product, but a human can make you feel at ease in a way bots can't (at least until someone rolls out GPT-3 to chatbots).

  • I've been looking for the "stitch fix of banking" for about 12 months, and this is the closest thing to it so far. The big question is, like stitch fix, can they scale it? Unlike stitch fix. As they scale it, can they stay regulated?

4. Ecolytiq - Carbon impact API (Europe) ♻

  • Ecolytiq analyzes payment transactions in real-time for their environmental impact. It provides a C02 calculation, advice, content, and individual offsetting or compensation offers. There is something about just letting people see their impact that's powerful. Like, pile up all your single-use plastics for the week in one place, then take a photo. Seeing is insight.

  • ESG and sustainability may get a big shot in the arm from the Biden "build back better" approach, and it seems every bank strategy team is now thinking about how they do so in more than words. Ecolytiq is ready-made for that type of partnership (subject to impossible bank procurement processes and execution risk).


Things to Know 👀

1. Marqeta is to IPO and moves into Credit Card issuing and processing

  • Marqeta Inc filed its IPO eyeing a valuation of around $10bn. This would be more than double their last valuation in May 2020 of $4.3bn. This week Marqeta also announced it is offering credit cards via a partnership with Deserve.

  • 🤔 My Analysis: Marqeta plays in a space between Payments Processors (like i2C, FIS, and Fiserv) and tech partners (e.g., Synapse, Unit, and Bond). They're more modern and easier to work with than the incumbent processors (albeit more limited in what they do).  

  • 🤔 My Analysis: There's a trade-off in the supplier space of speed to market vs. economics vs. long-term flexibility and capability.  Marqeta is right in the middle of all of that, and that makes them interesting to watch. Marqeta is big enough for large banks but flexible enough for big tech.

  • 🤔 My Analysis: The partnership with Deserve is interesting. It suggests Marqeta could partner with BaaS / tech partners who are upstream from their core processing.

2. Brex applies for a bank charter

  • Brex, the credit card and cash management platform for startups and scale-ups, submitted an application with the FDIC and UDFI to establish a bank. Former SVB Exec Bruce Wallace will serve as CEO. The goal is to expand their offering to include more credit solutions and FDIC-insured deposit products.

  • 🤔 My Analysis: We're only just getting started in the B2B banking and fintech space. Brex is a current market leader. They've gone well beyond helping manage everyday expenses into the cash management space (e.g., payments, reporting, reporting, expense management, and built-in integrations to accounting platforms).  Brex offers far more than traditional bank solutions provide, but Brex still has a lot more they could do, and the charter would allow them to. 

  • 🤔 My Analysis: The sheer scale of B2B payments alone is staggering, but when you overlay the treasury management and working capital management (lending and deposits), FX, and other businesses, its TAM is unfathomably large.  

  • 🤔 My Analysis: The digital experience most banks offer is so bad it's criminal. They still expect businesses to send a flat-file via FTP (seriously) or use their awful UI. Tools like Modern Treasury help abstract this pain, but the deposits still ultimately sit at your bank. If Brex gets a charter, that begins to change. If you're in commercial or corporate banking and you're not obsessed with Brex, you're doing it wrong.

Quickfire thoughts

  • Robinhood's CEO appeared at the congressional hearing, and for me, AOC nailed it. She told Vlad that "Robinhood failed to manage its capital risks." 🎯  Payment Order Flow is the sideshow (a very useful sideshow). All brokers have the same rule book on what capital they need to put aside for what situations. Robinhood failed to follow that.

  • Stripe is valued at $115 in private a rumored private market funding round, not an IPO. If there was ever a market to IPO in, this is it.  So why not? My gut says the cash is for acquisitions and international expansion in a way public markets might struggle to support (although Amazon pulled it off).


It's been a busy week in fintech! If you missed any of it, check out Nik Milanović's This Week in Fintech for the headlines of the past week:

  • Goldman Sachs launched its digital retail investing tool, Marcus Invest, as Public.com raised $220 million.

  • Web browser Opera is launching a consumer fintech product strategy.

  • Stripe launched its business carbon removal tool.

  • Mastercard is partnering with the Bahamas to launch a CBDC credit card.

  • Neobanks Zeta and HMBradley both launched joint accounts.

  • Ikea bought a bank.

  • And Russell Westbrook invested in Varo Bank.


Good Reads 📚

1. Bitcoin is as a meme and a future

  • All money is a bet on the future; it is also summoning a future.  We only accept an asset or money today because we believe someone else will value it tomorrow. When people design new money forms, they usually intend to tell a new story about the future. The Bitcoin-curious aren't preppers who think the world is ending, but they are hedging on some level, just in case.

  • Trump is a meme who was elected president. QAnon is a meme made into a religion. Tesla is a meme that is on the S&P 500.  But the logic of crypto memes is also one of deferral. They instruct people who hold Bitcoin to continue doing so. No matter how high the price climbs, do not sell. When everything is crumbling, cryptocurrency feels like it stands beyond everything and is the center of everything.

  • 🤔 My Analysis: I loved how this essay contextualized humanity finding something to believe in; even if that thing is a scam, as a hedge against a world they see crumbling.  We all feel it; we're in the gap between one world and another and looking for things to anchor. What we trust is changing, and that's changing finance.

2. NFTs make the internet ownable

  • Ownership gives power. Historically music labels had power over artists and platforms over creators. Platforms ask you to trade ownership in return for distribution. NFTs allow creators and artists to retain ownership and reach their audiences directly. While the current hype and prices are more than likely a bubble, NFTs create a marketplace around actual digital ownership.

  • A common complaint about NFTs as digital art is that it's easy to create a duplicate. But consider a Warhol or Picasso, you can create a copy, but the original still has significantly more value. NFTs become media legos.  Creators make money every time the asset is resold; consumers become investors in content creation, and developers can mix and match all of these legos to create new worlds.

  • 🤔 My Analysis: If the whole NFT space is still baffling, try this. Download (or add to chrome) the Metamask wallet. Load that wallet with funds (the Wyre integration makes this super easy). Head to any of the NFT markets like Opensea or Rariable and find something you can afford. Buy it, then, if you like, put it up for sale again.  It feels a bit like using the internet with a dial-up modem; it's not ready for mass adoption, but it's a window into the next 20 years.


Tweets of the week 🕊


That's all, folks 👋