Building a Broker-Dealer: How Much Does it Cost?
This is the seventh and final post in our series on How To Build a Broker-Dealer. Please see the previous posts here: obtaining a broker-dealer license, the technology, choosing a clearing firm, vendors, the team, connecting to the stock market.
Every startup is different, and every broker-dealer is different, but in this post we share a breakdown of the costs we incurred building Proof.
It took us two years to go from starting this company to our first trade, and over that time we spent about $1 million. Since going live, our monthly run rate has been around $100k. We think this is a crazy low cost to build and run an institutional equities trading business, and we are proud that we have been able to work with such capital efficiency.
In comparison, our competitors regularly invest tens if not hundreds of millions of dollars building and rebuilding their equities electronic trading platforms. Granted, it is a far more daunting task to rebuild a huge legacy system with tons of pre-existing functionality (that non-technical business heads demand be recreated), and numerous mandatory integrations with legacy workflows. At Proof, we had zero baggage, and we had the luxury of starting with a narrower offering than what a “next gen” system would inherently entail. As a new entrant, we are nobody’s one-stop solution, and we were able to begin with just what we considered the most compelling algo offerings rather than needing to cover every customer electronic trading use case.
The other major capital efficiency advantage we have over large banks is the productivity of our small team. If you airlifted the six of us (back) into a major bank, and ignored the fact that half of us would probably quit immediately, it would be impossible for us to have accomplished what we’ve done at Proof for so cheap. First of all, our cash and liquid equity compensation would probably be 5–10 times higher. At Proof, like most startups, our cash salaries are well below market as each of us is primarily compensated with illiquid Proof equity. Second, at a large firm our individual output would drop dramatically as we would have to stay in our lanes and interface with rest of the bank. At RBC, I’d say two-thirds of my time and energy was spent figuring out ways to work around deficiencies elsewhere in our tech platform that were outside our team’s control. I would imagine for Prerak and Beau, this percentage was even higher. Plus everything is just more expensive inside a bank — technology must be requisitioned from the appropriate internal team; external contracts are bloated; major cost centers are shared across business lines. At Proof, we can pick and choose our vendors at will and have negotiated leaner contracts as needed.
Prior to our go-live, our entire team was quants and software engineers (we have since added Brian though!). There were no product managers or sales or business folks. Every single one of us directly wrote code for our trading platform, algos, and/or supporting technology systems. We did not, and we still don’t, have a sales and marketing machine, and that side of building the business is challenging. But no doubt the focus and concentration of our team is what has allowed us to remain so lean.
Our philosophy when it comes to spending money
It seems that most technology and fintech startups these days decide to hop on the VC treadmill and raise a boatload of money, and then they aggressively pursue growth above all else. At Proof, we have taken a lean startup approach, treating every expense as though we are spending our own money, and trying to operate in cockroach mode until we reach breakeven.
In-house vs. outsourced
In general, and especially when it comes to technology, we try to do as much in-house as possible. We feel our technology is our core strength. We want the freedom and flexibility to build and deploy new functionality at will, and we desperately do not want to be at the mercy of a vendor’s SLA when issues inevitably arise. Our whole team is well-rounded and self-sufficient, and we do not like contracting out to others any task we feel we can handle ourselves. That said, there are some significant aspects of our business that we have outsourced:
Key outsourced pieces:
- Infrastructure hosting (i.e. the cloud) — one of the most significant early decisions we made was not to build out our own data center presence and instead to build our trading platform in the cloud.
- Market data normalization and research environment — market data is a beast. Consumption, normalization, storage — these tasks are a significant undertaking, but fortunately there are solid off the shelf solutions that we felt comfortable going with. We use Exegy for real-time market data processing, and OneTick for working with historical market data.
- Clearing — unless we want to raise another $10–50mm, we need to partner with a well-capitalized external clearing firm. We chose to work with Apex.
- Market access — we have the ability to obtain exchange memberships and access the market directly, and we do so in the case of IEX and certain dark pools, but due to exchange access fees and pricing tiers, we have no choice but to access the majority of markets through a DMA partner.
- Regulatory Consultant and FINOP — we talked about Shannon Fitzgerald (Regulatory Ridge) and Anya Cross (Cartana) in our broker-dealer license blog post. Both of them bring substantial regulatory expertise that we were lacking on our team, and they have both been significant contributors to the firm.
Expenses from incorporation to launch:
And now for the juicy part, the actual breakdown of numbers. In the two years it took to go from incorporation to first trade, we spent just over $1M:
As you can see, even with our low cash compensation, salary + benefits were by far our largest expense. This isn’t too surprising because we did almost everything ourselves. The few areas we did rely on others was navigating the broker-dealer process (regulatory consultant + FINOP), setting up the company (legal), and a managed research environment (OneTick). It is also worth noting that we got a bunch of AWS credits from participating in YC, so our tech costs remained very low during our build out and even into our first year in live production. We also kept our costs down by waiting until just a few months before go-live to turn on and start testing with real-time market data. Even though it’s the “public market data feed,” SIP data is our single largest technology expense. It was a delicate balance deciding when to flip the switch on the SIP, because while we didn’t want to incur this expense unnecessarily, market data processing is one of the most critical and challenging technology pieces, and it did need to be extremely well-tested prior to live trading.
Ongoing monthly fixed expenses:
Since our first trade in January, our ongoing costs (ex variable trading costs) have averaged around $100k/month, although our monthly costs have been gradually increasing and are currently around $125k:
Once we went live, infrastructure costs increased as our credits ran out and as we needed more capacity. We have also needed to integrate with more and more third-parties that our pilot customers use as part of their trading workflows, and many of these vendors “double-dip” (i.e. extract fees from us in addition to their clients), and unfortunately we don’t really have recourse. We’re still running as lean as we can, but the costs do continue to creep up, and we currently anticipate our monthly run rate to grow to around $200k as we expand our cloud infrastructure, integrate with more vendors, increase salaries, and bring on a few more teammates.
Conclusion
All-in-all, there isn’t a whole lot we would change about how we deployed capital while building this company. We had the benefit of direct prior experience building new trading platforms at IEX and RBC, and as a result, we were able to do much more ourselves (and accordingly pay less to others). At Proof, from the beginning, we were extremely diligent about avoiding operational fat without cutting any corners on the tech platform. We are proud that we were able to build this business so efficiently, and we do think it is a huge testament to the quality of our team. It is our hope that by sharing our playbook, the prospect of building an institutional financial services firm may seem less daunting because what this industry desperately needs is an injection of more efficient, honest, transparent new players.
