Henrique Dubugras and Pedro Franceschi, now the co-founders and co-CEOs of corporate card and expense management startup Brex got started early. While they were in middle school, Henrique started coding to build clones of games that his parents wouldn’t buy for him and Pedro became the first to jailbreak the iPhone. But when the patent infringement and cease and desists came, they needed to find new things to work on.
Fast forward to when Henrique and Pedro were in high school, they met on Twitter and later decided to work on a company together. That company was Pagar.me, a payments processing company serving customers in their home country of Brazil. They raised over $30 million in funding for their company, employed over 100 team members, and were acquired all before they stepped foot in a college classroom.
After selling Pagar.me, Henrique and Pedro decided to matriculate to Stanford, where they spent less than a year before dropping out to join the YC Winter 2017 batch. There, they hoped to work on a VR startup but after just a few weeks, they realized it wasn’t the right domain for them. As YC demo day neared, they pivoted to what became Brex after noticing that most of their batchmates couldn’t get corporate cards even though they were venture-backed.
Five years later, Brex has raised $1.5 billion from investors like Greenoaks Capital, Tiger Global, YC Continuity, and others; serves over 50,000 customers across its corporate card and expense management offerings; and employs a team of over 1,100.
I recently had a chance to sit down with Henrique to get a deeper look into how Brex was built from the ground up. In our conversation, we explore the granular details of Brex’s growth: areas like how Brex raised its first round of funding without a product or a pitch deck, what the team focused on during the 15 months it took to get to a public launch, the co-founders’ philosophy on hiring engineers and other functions in the early days versus now, and Brex’s recent decision (and rationale) to pivot towards serving larger customers. We also had time to zoom out a bit to discuss both the tailwinds and headwinds that Covid-19 brought to Brex’s business; the “new world we live in” and the role that Brex hopes to play in it; and the company’s biggest priorities.
Steven Li: You and Pedro co-founded Pagar.me together before starting Brex. What did you learn about the right separation of responsibilities and how did those learnings apply to Brex? And how has that model evolved over time?
Henrique Dubugras: Early on at our first company, we were both coders. But Pedro was a much better coder than I was, so he basically told me to go do the other stuff and the “other stuff” at that point was sales. At first, I didn’t think sales was a thing that anyone could do because I always thought of it as a function for really extroverted people who played sports in high school and really knew how to talk—and I was an introverted coder. But I read this book called “The Sales Bible” and it turns out that it works if you just do what’s in the book. That’s how it was initially but eventually, it evolved.
We realized that I was better at doing sales than managing salespeople. And not only sales in terms of selling customers, but also talking to investors, the press, and other kinds of external communications.
So we evolved to this kind of internal-external model—one that we also use at Brex today in the sense that I do a lot of external relations and he does a lot of the internal stuff. Brex was a repeat in the sense that when we started again, Pedro was focused on engineering and I was doing a lot of the other stuff like navigating the regulatory landscape, creating partnerships, getting the business off the ground, and fundraising.
Today, we operate on a co-CEO model where I lead the external stuff and Pedro does a lot of internal stuff. Especially now as Brex has pivoted more enterprise, my role has been also especially relevant because at a lot of bigger customer meetings and stuff like that, it’s helpful to be able to split those responsibilities between the two of us.
You started Brex because none of the companies in the W17 batch could get a corporate card, but the vision for Brex was a lot more than that. How did Brex approach customer discovery and what did you learn about what crucial features needed to exist starting from when the first customer was onboarded?
We knew that every company needs a corporate credit card from our days in YC so we didn’t have to do a lot of convincing; we just needed to make sure that ours is better than the solutions on the market.
And then the first version of it was like, okay, let’s go to the set of customers who really can’t even get one. And so anything that we do is better than nothing. So we were competing with nothing. And that was kind of how we prioritized. We thought about how we could make the minimum version that works, where you can swipe a card and a transaction gets processed.
Then, we went feature by feature. After we got a card working, we needed to build a dashboard where customers could see their transaction history. But we had customers using Brex even before they could see those transactions because we were competing with nothing.
Brex graduated YC in March 2017 and was able to raise a $6.5M Series A in the following month. My understanding is that you guys did this without any code (since Demo Day had been too soon after pivoting away from another VR idea you and Pedro were exploring) and without a deck. Can you walk me through how you made it happen?
I think that there were a couple of things happening and the first one was we had sold our payments company (Pagar.me) before and when Mickey (Malka) and others called the company that acquired us and our customers during reference checks, those came out positive. Being second-time entrepreneurs helped us build trust that we were going to build what we said we could build.
The second thing was that FinTech was hot at the time; we were in this phase where Stripe and Square were doing really well. In general, since we were early, I think it was a bet on both the team and space.
I remember in your fireside chat with Anu (Hariharan) at Techcrunch Disrupt that you and Pedro really wanted to get Max Levchin as an inventor and the plan to get in the door was to interview at Affirm and then ask to chat with Max before accepting an offer. Can you share how he ended up helping Brex in the early days?
In the early days, Max was very good at helping us figure out how the landscape worked. He told us about Marqeta, which was the first vendor we used to start issuing cards and explained to us how he got his first debt deal.
We knew things about the fintech ecosystem and about the payments domain from our previous company, but Max understood a ton about the U.S. environment and what was happening here. He also helped us figure out how we were going to launch and build Brex for the long term. As Brex grew, he continued to help us with mentorship around company-building.
How did you and Pedro define PMF in the early days and what were the early signs that Brex had it?
After our public launch, we began to grow really quickly. Customers were signing up and they were starting to spend and we were making revenue. We went from one to like 10 million in ARR over the period of five or six months after launch.
To us, that seemed like what people would describe as product-market fit in the sense that people were signing up, using our product, and liking it. We were also guided by revenue metrics and payment volume.
Brex took 15 months to launch after raising the Seed/Series A (combined round), which you mentioned was because you wanted to get a lot of things right and truly launch a differentiated product in the space. And in that time, Brex garnered over 100 customers and tested and iterated on the product. During that time, what did you guys learn, and, in your mind, what needed to happen for you to be comfortable with a public launch?
The way we thought of it was we really wanted to take advantage of the public launch. We wanted the signup to feel very instant, so we were trying to get like 70 or 80% of customers approved instantly without any manual verification. We also wanted to have feature completeness relative to what already existed in the market so that when we scaled up our customer base they were not going to feel like it was a terrible product. Even though our feature set was very incomplete looking back, we thought it was complete enough to launch at the time.
When Brex wasn’t publicly launched, I assume that the early team was deeply involved in onboarding each customer. But when Brex became publicly self-serve, I assume that the customer acquisition strategy needed to change a bit to keep up with scale.
I would say it’s pretty similar today vs. back then. In general, performance marketing didn’t work too well for us. Many startups had Google and Facebook as the main channels and just doubled down on them, but to be honest they never worked super well for us.
We were always more focused on community-oriented stuff like channel partnerships, referrals, branding, and work on getting more general awareness or bottom-of-the-funnel stuff. But as we go more mid-market and more enterprise, there’s a lot of stuff we should be doing that we aren’t and we’re working on it right now.
Brex grew headcount very quickly: from 12 to 100 to 400 to 600 to 1100 (numbers are year-over-year from launch). Back in the early days, you mentioned that your approach to hiring non-engineers was to pick the right people who you could hire as early as possible and grow into those roles if they weren’t already a perfect fit. And for engineers, you mentioned that you had a technical challenge to build an authorizer to discern between the more junior and senior folks. Can you share the rationale behind that hiring strategy? Curious about what you were looking for on the engineering side and what qualities you were looking for in those crucial early candidates.
Early on, we didn’t hire a lot of junior folks; our early strategy was to hire fewer engineers who understood architecture better and could build things the right way. As for why we ran the authorizer problem, we wanted to get a sense of how well someone could do something that they might actually work on at Brex versus things like how many tennis balls could fit in a bus.
A big differentiator on the engineering side was testing. Testing things is important to think about since scaling a payments company is different from scaling Facebook, for example. One scenario is if a million people are hitting the “like” button at a certain time on Facebook but it turns out you miss-count a couple of these “likes,” it’s not the end of the world. So a lot of the strategies you need to use to implement and grow Facebook are about a lot of users accessing a resource at the same time but that operation might not be reliable all the time. The question is how do you create algorithms and systems that can solve that kind of problem?
Whereas at Brex, every transaction is very important, and making sure we have the right balances and the right money movement is critical. In that sense, our authorizer question was pretty relevant and was actually relevant to how engineers would think about architecture, databases, and maintaining consistency and reliability.
From listening to a few other interviews, my understanding is that in the early days you always wanted to hire great engineers as you saw them. And then on the non-engineering side, you often took bets on people who might not have necessarily had the experience in whatever role they were applying for; you didn’t want to miss out on the chance to hire talented people even if they weren’t an exact fit for the role. Would you say that that philosophy has remained true as the company grew?
I would say that with most startups, as you are smaller you want more generalists and as you get larger you want more specialists. Today, we try to hire people to have like, very relevant experience to what we’re doing, because it’s worth optimizing every little piece of it.
Before, we needed someone who could learn and do multiple things. So like today at Brex, there are one or two people where their whole job is to manage social media. There are other examples but in general, we probably want to hire someone that has experience doing a specific function versus in the past we might’ve had one marketing hire that managed social media, email marketing, and other related functions.
Intuitively, it feels painful for companies to switch their expense management over from something like Concur to a new service because even if those services suck, they “work” and companies (especially ones that are not software savvy) might be resistant to change. How does Brex think about this challenge and can you walk me through the steps of a typical sales cycle that are taken to mitigate it?
For the longest time, Brex was very focused on getting startups that grew. So our whole model was hey, how do we sign a couple of companies that are small and then grow up with them? For example, Scale AI was our first customer and now they’re almost 1,000 people and are still on Brex Empower.
Now that we’re targeting larger customers—ones that might already be big—we need to come up with a value proposition that is good enough to be worth the switching cost. There are a couple of value propositions that can be valid reasons for a company to switch.
One, for example, is the shift to working from home, and many employees now getting stipends instead of lunch in the office. You need a better way to manage that process because if you have 1,000 employees and everyone is submitting expenses every month for some stipend that’s like a pain to manage for the finance team and for the whole company.
Another example is that a lot of business travel was focused on salespeople who are visiting customers or business people visiting partners. Now, with remote work, every single company is doing offsites all the time and traveling to see each other as teams. So a lot of the workflows that were built in Concur and others broke when you’re looking at these distributed companies and their travel patterns.
Then there’s the consideration of going global. A lot of companies are starting to hire people everywhere and to be able to do that well you need to have a system to reimburse team members in different countries. This paradigm shift is pushing companies to look at new solutions and reevaluate their current systems and we’re definitely thinking about that a lot.
How did the Covid-19 pandemic change the way that Brex operates? One thing I can think of is that when startups started to conserve cash they probably cut spending, which reduces TPV and therefore interchange rates. But at the same time, companies need their credit cards to work more than ever. If these assumptions are right, how did Brex adapt operationally in that environment?
If you look at payment processing in general it actually went up. For instance, if you look at Stripe and Adyen their numbers actually increased significantly because of the on-demand, “Covid-boosted” kind of companies like Doordash and Instacart.
For corporate cards, we actually did see processing volume go down because people weren’t traveling. And I will say during Covid there weren’t a lot of companies wanting new software because everyone is just focused on surviving.
Now, with the macro backdrop of the recession, we’re seeing decreased levels of spend. But there’s an enormous amount of focus now on doing more with less and software that makes your company more efficient. So a lot of times last year, if you weren’t helping drive revenue up you didn’t get the time of day of the CFO. But now every CFO is looking at new ways to help manage costs so we were focused on executing on that.
You’ve mentioned that your philosophy on fundraising is to build relationships with investors over time and only raise when you need to and see those fundraises as an opportunity to partner with amazing investors and re-price. Brex raised two rounds of funding in 2021 – once in April and once in October (Series D and Series E). Why were those raises critical at the time?
We definitely only fundraise when there’s an alignment where we find the right investor at the right price and in the right market. And we always like to be very over-capitalized. There are some companies that like to optimize to the bone so that they always get the minimum dilution and only get the capital that they really need at every moment in time. We like to have spare cash for moments like this. To this point, thankfully we have a lot of spare cash because we now have a lot of runway.
But we also don’t like to raise money from people we don’t enjoy spending time with or those we don’t think will be good long-term shareholders. We want people who can hold our stock for a long time because as a public company if there’s a huge shareholder that dumps a bunch of stock, the price goes down. We’re optimizing for long-term holders that not only won’t sell but would consider buying more as we grow. So we like to spend our time getting to know folks and understand them—and having them understand the vision. If we notice that an investor really understands our strategy and they “get” what we’re doing, that builds confidence for us to execute over the long term.
As for the timing, I think 2021 was a special year for fundraising. There was just so much money in the market that it was almost dumb not to fundraise last year. The cost of capital is really low and you could just raise—and we raised over a billion dollars last year. You could get insane amounts of money at relatively non-diluted valuations. We found the partners we wanted to do it with and did it.
Tiger led the first one and Green Oaks led the second one and I think if you ask them, they don’t regret doing those deals. They’re super excited about Brex and are excited that we got to partner together.
In June earlier this year, Brex made a decision to stop serving small businesses (defined as those who haven’t raised venture capital). And with the context of Brex empower’s launch, that direction definitely seems more geared towards enterprise. But to double-click into Pedro’s thread talking about not being able to provide “white-glove service” to these smaller companies when did that first become a problem?
We started with serving startups and had three expansion packs: eCommerce, large businesses, and small businesses. For the last one, we were thinking of serving traditional mom-and-pop shops and expanding horizontally. We tried to do and grew all three of these segments at the same time.
What we didn’t expect was that the small businesses consumed all of our resources; we went from onboarding 500-700 customers a month to onboarding 5000-7000 customers a month. Then, most of our efforts went to increasing support for these customers—initiatives like building more automation and tooling to combat fraud. At that scale, there were a lot of challenges and we allocated a lot of resources to serve this one customer segment, and we kind of neglected the other two.
When we looked at the business, we were like, okay, turns out we can’t do all these things at the same time because all these problems are very complicated. We need to pick a few that are not as complicated or pick a few that we believe are the best ones for the business to solve. What that meant for us was that we needed to choose between larger customers or small businesses and we decided to go with the larger customers because we actually had no alternative.
Our biggest customers like Scale AI told us, look, if you don’t build stuff for us, we’re going to have to figure something else out. So it was either doing that or losing many of our biggest customers. So we decided we had to pull the plug and focus on those customers—the startups that grow—and also the larger customers.
How did the team come to the conclusion to draw the sand at companies that are venture-funded? Why not something like revenue?
It’s hard to define it but basically, the way we thought about it is that if we look at our product roadmap and the things we’re building, we’re building things for a company that has reached some amount of scale. And by scale, we mean high headcount.
Because if you look at expense management, for example, if you have a lot of employees, there is often a lot of value in making that process super efficient. If you don’t have a lot of employees, there aren’t enough expenses to process so you don’t need to make anything super efficient.
Also, a startup that becomes big happens all the time but, for example, a retail shop that becomes Walmart is a bit more rare.
With Brex’s Empower launch and re-focusing of efforts into enterprise, what would you say are the three most important initiatives at Brex in the next year or so?
Number one is making sure that we support global companies. As companies go global, they have to figure out how to issue cards everywhere, reimburse people everywhere, and pay taxes everywhere through multiple entities.
Number two is to build spend management at scale—all the features that are needed by companies at the scale of DoorDash and Coinbase, which we’ve already signed up. We want to be able to say that if you’re a venture-backed company and you’re growing, you can be sure that you’re going to be able to scale with Brex indefinitely.
And then, number three is to build our go-to-market for this new world that we live in. Previously, all of our go-to-market and marketing was built for corporate cards—how we’d serve startups and our other segments. Now that we’re building for larger companies, a lot of our marketing, sales, and other functions need to evolve. That’s why we hired a new CRO, among other developments.
Since the corporate card and expense management categories are so big (even in the United States), what did the consideration of allocating resources to go internationally now versus later look like?
We believe that every company in our category is going to become a global company eventually and this is especially true given this macro backdrop. The cost of hiring an engineer in Brazil is so much lower than in the US and honestly, they’re just as good.
I think we were interviewing for an HR function and in Toronto, talent is literally half the price compared to San Francisco; and the person we ended up hiring in Toronto was better than the SF candidates. I think every company—especially tech companies that are efficient and profitable—is either already a global company or will look at hiring globally. The talent market is going to be the global market.
How would you describe “this new world we live in” and the role that Brex hopes to play in that moving forward?
If we look at what’s been happening over the last several decades, companies are becoming more distributed. And by “more distributed” I don’t necessarily mean remote work—just that businesses have grown to be multi-location.
If you look at Walmart back in the 60s, they started one store and it was really hard to open a second store but they opened it and they figured out how to do many stores. At the time, Sam Walton bought a plane to go between the stores just so he could manage them.
But then telephones started becoming mainstream. And then we had the internet, then email, and then Zoom. With all these new tools, companies could open more offices to either hire people at a lower price from a talent pool that wasn’t accessible before or have team members be closer to their customers. We’re continuing to go in this direction with Google Docs, Figma, and other tools that allow teams to work better while not being in the same place.
We’re building a lot of tools but no one has done that for spend management, yet money is kind of the resource allocation process of every company. And when we talk to a lot of companies and ask if they want to save money, they say “Yeah, we want to save money, but more importantly, we want to save money to allocate it to the places where it’s needed the most.” Companies want to make sure that there’s money flowing to the right places at the company—that it’s not being wasted in places that don’t matter. And I think with Brex Empower, our spend management solutions, we’re kind of enabling this next generation of companies that are distributed to allow for every employee to be empowered to make better financial decisions.
If we can empower people—the 99% who are doing the right thing—to not be bogged down by bureaucracy, we can help companies move a lot faster, regardless of where their teams are based. And that’s especially true on a global scale.