“I don't think it's necessary to just get all the data all up front, no matter what. It's important to know exactly what questions that data's going to help answer.” – Mike Sall
Mike Sall (@sall) is Co-Founder of Goldfinch, a decentralized credit platform for crypto loans without collateral. Mike previously co-founded Unbox Research, was Head of Data Science at Medium, and served as Product Manager at Adobe.
To listen to Mike's quick interview on habits, routines, and inspirations, click here.
Chapters
- How Goldfinch is helping entrepreneurs
- Mike's background and interest in data science
- Mike's experience at Coinbase
- The origins of Goldfinch
- Why raising capital in emerging markets is so difficult
- Assessing comfort levels with borrowing in crypto
- Mike's previous businesses and learnings
- The Goldfinch Flight Academy
- On crypto regulations and compliance
Links
- Connect with Mike Sall: LinkedIn | Twitter | Goldfinch
- Google Analytics
- Medium
- Buzzfeed
- Evan Williams
- Coinbase
- Compound
- Blake West
- Kiva
- Bitcoin
- Ethereum
- Binance
- Kraken
- Almavest
- OYA
- Telegram
- Notion
- MetaMask
- Goldfinch Whitepaper
- a16z (Andreessen Horowitz)
- Goldfinch Flight Academy
Terminology
Books Mentioned in This Episode
- Principles by Ray Dalio
Learn More About This Topic
Emerging Markets: Introduction
Global Edge provides key economic stats for countries with emerging markets.
This playlist is a great way to start learning more about emerging markets, from commodities exchange in Ethiopia to the trend of African growth.
Goldfinch prepared a thorough overview of the protocol, borrowers, backers, auditors, liquidity providers, and leverage model.
This selection of TED Talks covers everything from how cryptocurrency can help startups get investment capital to the future of branded currency.
Top 5 Books to Learn About Blockchain
Investopedia offers this list of books to get you started on learning more about cryptocurrency.
Transcript
Daniel Scrivner:
Mike, thank you so much for coming on Infinite Games. I'm so excited to chat with you today.
Mike Sall:
Yeah, thanks for having me.
Daniel Scrivner:
We're going to spend all of today talking about Goldfinch, which is a new protocol that's going to launch in early 2022. And I wanted to, well, take a departure for a second, talk a little bit about your background, but I want to ask if you could just paint a picture for people of the big problem that Goldfinch is solving and maybe the elevator pitch around how you're approaching, solving that problem.
Mike Sall:
Yeah. We're going to talk about the problem, there's two angles to get at. One is from the crypto perspective, and one is from broader global perspective. Starting from the global perspective, we were talking to all of these businesses around the world, especially fintech and lending businesses, where they're in the business of making loans out to other folks and they need more capital to grow their business, to make more loans out. And they tend to need in the range of a few hundred thousand dollars up to 10, $20 million.
Mike Sall:
And this range is difficult for them to get the capital because they're too large for the local capital markets, but then they're too small for the really large institutional funds, they like to write like $50 million checks to make it worth it, to do like foreign investing. And so we talked to all these businesses and they really needed capital. And we realized, "Oh, crypto can help solve this." And then on the crypto side, there's technical problem, which is there are existing lending protocols like Ave and Compound.
Mike Sall:
And the way that they work is you put up say $150 of Eth in order to borrow about like $100 for that. And so you have to over collateralize all of the amount you borrow with other crypto. But these businesses around the world, and we're talking to folks in Brazil and Nigeria and Indonesia, all around, they don't have money because they're looking to borrow money. So the problem that we're solving from the crypto perspective is, how do you enable borrowing without having crypto collateral and being able to show your trustworthy in other ways? So that's the high level technical problem and the global problem that we're solving.
Daniel Scrivner:
Which is fascinating. I think everyone listening probably to grasp how big a problem and opportunity that is, but we'll get to that and talk about it in a little bit of nuance. And I encourage anyone that's listening to follow along, feel free to pull up GoldFinch.finance, just to read through and look a little bit at what Goldfinch is building. I want to ask you before we go any further just a couple of background questions, because you have a really interesting background. Before founding Goldfinch, if we go back a little bit, you were the head of data science at Medium for four years. That's a super interesting position to be in.
Daniel Scrivner:
It's become obviously a big prominent platform, especially in crypto for people to share all sorts of news and events and announcements. What was that role? And I guess, what did you learn in that role?
Mike Sall:
The high level role was like, "Okay, let's use the data that we have to make the product good." And so it was like, "Okay, how do we understand from the data, what we need to do?" And so a lot of the learnings were about what kinds of things are people looking to do and where do they spend their time? And also part of that, we started building out the initial recommendation engines for machine learning models that were figuring out what posts to show people in different places. And it was a little while ago at this point, but some of the learnings there were that we started finding like certain kinds of topics that people wanted to read about and making sure that we were showing those topics to them.
Mike Sall:
And then we also found things like there was a large correlation between how much time people spent into writing their posts and then how successful they were, an indication of the amount of effort. And we started to learn that there were no silver bullets when it comes to producing content that other people want to read, you've got to just work hard and do good work and that's what ends up getting people's attention.
Daniel Scrivner:
Yeah. Super interesting. One of the reasons I wanted to ask that question is, this is a topic that comes up all the time, talking with early stage founders, is, one, what data to be paying attention to, two, how to go about collecting that data. And then feel like a third one is just how much time should be spent focused on quantitative aspects of the business as opposed to trusting our gut, trusting our intuition about what to build? So I'm just going to lob that out as a not super easy open-ended question. If you were sitting down with a founder or talking with other founders, is there any advice you would give them in terms of just how to think about data? And that could be stage dependent, just any insights there, general things that are applicable.
Mike Sall:
A fun fact that you wouldn't expect prior head of data science to say, which is on the Goldfinch Finance site, we don't have any Google Analytics anywhere. We don't even bother with it. And part of it is privacy thing, we don't want to have Google Analytics tracking that information. But also, I've personally never found Google Analytics to be helpful for any of the questions that my teams were digging into or that I was digging into at these companies. And I hear a lot about folks being like, "Okay, we got to add the data in here because the data's going to be useful later on."
Mike Sall:
And this idea that like, "Oh, you just throw data at it and you'll get the answer that you need to something later on in the future." And I've never found that to be super helpful. Usually, it's combined with, "What are the specific questions that we have? What do we need to know to be able to make some good decisions here? And then what's the data that we need in order to help make better decisions there? And then let's implement that specific data to make sure we have that." And so, from my experience have come to this, like, "Okay, if we don't know how we're going to use the data, then just don't bother."
Mike Sall:
Usually it's more clear like, oh, we have to know how many people are using this particular feature or how much are people actually spending on this thing, then we have to make sure we have the data so we can answer that so that we can figure out what to do next. But I think I would say to other folks, this is probably counter to what other data professionals might say, but I don't think it's necessary to just get all the data all up front, no matter what. It's important to know exactly what questions that data's going to help answer.
Daniel Scrivner:
That makes me feel vindicated. And I'm so glad I asked that question, in part because I think just your note there, if I could try to restate it, because I've heard that as well too, it's just the "data" is some oracle that's intelligent that's going to give you insights. And it's like, no, it's not going to give you anything, it's just going to be a confusing mess of a bunch of different noisy data points. It sounds the advice is, one, generally, don't collect data unless you know exactly how you're going to use it. And then two, just think deeply about what you want to learn and then focus on just collecting data around those points. Am I getting that right?
Mike Sall:
I'll give actually an example there too, which is when I was at Medium, we did this whole thing to add certain attributes at the end of URLs so that when anyone went to URL, you could have a sense of where they came from, from a different URL. It was this whole way of implementing data. We saw that Buzzfeed was doing something similar, and it just sounded super cool, we were like, "We've got to do it." And it was like, "What are we going to use it for?" We weren't totally sure, but it was super cool we better implement it, and we spent all this time implementing it.
Mike Sall:
And then we just didn't need it for anything after that, and we ended up basically getting rid of it soon after. And so I would say, if we don't really know what we're going to use it for, it's not worth spending all the time to make sure all the data's in place.
Daniel Scrivner:
Yeah. I'll ask one final question. I've had a chance to spend a little bit of time with F. Williams who was the founder of Medium. I think he's just really interesting and amazing and very intelligent and thoughtful about what he builds. What was it like to work with him? Anything you learned from Medium from F?
Mike Sall:
I learned a lot about building products and how you think about prioritizing different things. I don't have any big pieces of wisdom or just the intuition that I think I built just by watching him make different decisions. And then he's very inspiring the way he talks about product and vision and what he's building. He was really good at framing things that we were always working on in terms of the broader picture and the longer term plan. And I think I learned that is also super important, just seeing how inspired I felt after he was sharing the different things that we were working on, the importance of framing things in that way too.
Daniel Scrivner:
Yeah. That makes a lot of sense. Seems he's really good at this concept that Ray Dalio talks about, Principles of Navigating Levels. And I feel like good founders, they can go to the high level, they can always frame up the vision and the broader picture. They can drop way down, they can give helpful insights there, and being able to go up and down that stack. I wanted to ask a little bit about, so before founding Goldfinch, you also spent time at Coinbase and you are the of product analytics there. And I want to ask a couple questions, because one, I haven't heard of that title, and two, it's just too fascinating to not ask questions.
Daniel Scrivner:
So I guess maybe the first one is, what were you responsible for in that role? And maybe if you could talk a little bit about when you joined, where Coinbase was in its trajectory and history? Had it been tracking product analytics? Were you there to kick that off?
Mike Sall:
Yeah, let's see. So when I joined, there were about, I would say four to 500 people at Coinbase. A lot of stuff was well-established, and it had an amazing data team already in terms of implementing data pipelines, having great ways of querying the data, and had a number of folks who were already analyzing it and understanding what was happening. And then they were in different teams, so building different part of the products had different analysts there. So then when I joined, we formed this data science team that was bringing all the analysts together to as all of the customer behavior across all of Coinbase's products.
Mike Sall:
At the time the focus was really on the coinbase.com app and Pro, the exchange, and understanding how were people using the product and how was that trending over time, and what opportunities did that show that we could be building or improving to make the product better for the customers?
Daniel Scrivner:
Was there anything novel that you did there because it's crypto? To make that a little bit less cryptic of a question, there's just things in crypto broadly, there's blockchain analytics, you can look at things on chain as well as obviously behavior in the product. Did you guys do anything there or was it really just on product?
Mike Sall:
Oh, you mean blockchain specific kinds of analyses?
Daniel Scrivner:
Yeah. Or if that factored in at all in any interesting ways?
Mike Sall:
Maybe a little bit less than you might expect because we treated it very similar to other Web2 products where people are going to a web interface to use the product and make trades and buy and sell different kinds of crypto. And so I think the extent that we was looking at, what are the different kinds of tokens people were buying? And does that signal maybe what other kinds of tokens people want to buy in the future? But that's the same as buying other consumer products as well, is, is it similar? So there wasn't as much, I guess, very specific different kinds of analyses that it only applied to crypto.
Mike Sall:
It was all broadly trying to understand, "What are people doing? What are their motivations?" One of the very first projects we did when I joined was try to understand the different personas of different users by their behaviors, not by their demographics, but by different kinds of activity profiles and say, "Okay, here's what one type of power user looks like. Here's what someone who is doing something just every once in a while looks like in terms of their activities." So it's similar to other consumer web products of just trying to understand... Once we have profiles of activity, then we can think about, "Okay, which ones do we want to really serve and focus on?" And then what we need to build to make those have even better experience with the product.
Mike Sall:
So I guess I would say it's similar to other products.
Daniel Scrivner:
I love that example you gave of using analytics to paint almost personas because typically the way that's approached is extremely qualitatively. It's almost like people are just interviewing customers and trying to invent this cookie cutter, I don't know, ensemble of a bunch of different customer types. And it just always seems really challenging. It seems really smart to do it from data. Did you end up feeling like that was a great approach and that was really helpful form factor or approach?
Mike Sall:
Yeah. I had also done that when I was at Medium and it was so helpful, that was like, "Okay, this is one of the first things we got to do," when I joined Coinbase. And I think the reason why it's helpful is first, it can surface different profiles of behavior you didn't expect because usually anyone at a company this, you're talking to your customers and you have a story, you're already telling about who your customers are, but then when you start from the data and just see what are the natural groupings of activities that surface, you find other kinds of folks you didn't even realize, or you find a group of folks who you thought were all different, but they're all using your product or service in the same way.
Mike Sall:
And it's helpful to first just know that that's what they're doing, and it just changes the way you think about what you're building. And then also it helps to prioritize who you're building for. So we would create different groups of people and then we could look at them all and we would say, "Oh, these folks are doing it this way, but we think the future is really for this other kind of a profile of user." And so we can spend a lot of our time with the design research team, talking to that other group of users. And then when we're trying to understand, are things performing well over time, we don't just look at the overall trend across all users, but we could look at what's the trend among this group of people who we are really hoping that we're building for.
Mike Sall:
And so even if it's successful or even if it's not working for one group of people, it actually really helps with this other group of people, then we're like, "Okay, this is successful because it's achieving what we want with this particular group of folks that we want to focus on." So I think that's where it can become the most powerful.
Daniel Scrivner:
That's fascinating. Well, thank you for indulging me. We can move on back to Goldfinch. You were talking about before we hit record, I think an interesting approach for this would be I've now been a part of multiple protocols from inception all the way through launch, but from the outside looking in, as an investor, as an advisor to that team, I thought it'd be really interesting knowing that Goldfinch was founded middle of 2020, you guys have been building for let's just say around 18 months, you guys are getting close to launch. That'd be interesting to just go chronologically.
Daniel Scrivner:
And so I wanted to ask just when did you or your co-founders, what was the initial seed of the idea? Because as we'll uncover, as we go a little bit deeper, this is a really big idea and it's very novel and I think it takes a lot of confidence and very high caliber of people to execute this. Did it start out as a big idea? Did it start as a small idea? What was that origin?
Mike Sall:
Let's see. The way I would say it started out with us seeing what was happening in crypto and in DeFi specifically and feeling like, "Oh, there's a lot of infrastructure in place here." DeFi can be having a much bigger impact out in the real world and having the sense that a lot of what was going on were different protocols, self-referential, all very contained within crypto. And my co-founder Blake and I, we just felt one of the powers of crypto in the future and Web3 in general is it can really help people around the world and expand access to capital.
Mike Sall:
And we felt the pieces are in place here. The industry could be doing things out in the real world and trying to think, "Okay, what's the key part now to actually get that impact happening beyond the central circular referential crypto space?" And we thought that DeFi lending was the key part of it, because we saw that Compound and Ave were doing so well, but even then no one else outside of crypto can use it. And we thought, "That is the key. If we can figure out how to make under collateralized, not collateralized by crypto, but collateralized by other things, if we can unlock that, that's a holy grail to getting the benefits of crypto out to real people all around the world."
Mike Sall:
And so we started there trying to think through, "Okay, how do we make that happen?" And we had rough ideas of how the actual design of the protocol could work and we were like, "That part, we could figure that, but the key thing is we need to know who's going to borrow, who's the initial go-to market." We actually just started talking to folks in crypto, early on we started talking to miners and traders and crypto enthusiasts. And we were like, "Do you want to borrow?" And what they would tell us in our early research calls was, "No, I don't really need to borrow, but is there a place I can lend, I'll provide money and lend capital."
Mike Sall:
And we started realizing that there was a lot of demand for the capital side, but not as much for the borrowing. Blake, my co-founder had been involved with Kiva for a while, super excited about what they do and he thought, "Oh, what if we start talking to the kinds of businesses that Kiva works with, these lending and FinTech companies around the world that are looking for capital?" And then we started just talking to dozens of them. And then those phone calls were totally different. We would be talking to them and before the end of the call, they'd be "Okay. So when can I get a loan? When can I borrow?" And we were like, "We haven't even started the company yet"
Mike Sall:
And then that was pretty clear that we knew exactly who we were going to build for, at the same time, we were doing all this work on the design of the protocol and how it could work. And there was a whole process there of trying to figure out, "Okay, what's the initial design that we can work with just to get things going? And then what are the holes we'll have to fill in later? And that's how the initial idea came about.
Daniel Scrivner:
For those calls, it sounds like very quickly, you just felt you had found a fit in terms of you hadn't built it yet, but you knew exactly who needed it, who wanted it, and who you could get access to it. How quickly did that click in place? Was that within minutes of you getting on the first call, did that build over time?
Mike Sall:
We reached out within folks that we already knew, just trying to get connected with anyone in the space. I think we talked to two different places and had 30 minute calls. And by the end of the 30 minutes, it was very clear that raising capital was their biggest pain point as a company. And by the end of those two calls, they were like, "When can we borrow money?" They just right away within 30 minutes, basically asking to be one of our first customers. And then they even started introducing us to other their folks. And so early on, we initially saw very clearly, these are the customers because they were introducing us to other potential customers.
Mike Sall:
And then we started building a large network of folks all within this emerging market space that all seemed to know each other. And that was where it was pretty clear that even before we started writing a single line of code, we had 10 customers lined up just waiting for the product to be ready to go. And so that was like, "We know that there's some fit here at least to start."
Daniel Scrivner:
Yeah. That's one of the clearest examples or stories I've heard of just finding a perfect fit where, one, you're hearing resoundingly back from customers, "Yes, we want this one. Can we get it now?" But two, knowing that you interview as a potential customer refers other potential customers to talk to you. So that's super novel. I think it would be helpful to go a little bit deeper for people. You and I have talked before a little bit about why the emerging markets is such a tough place for entrepreneurs to be able to raise capital. Help paint that picture for people listening.
Daniel Scrivner:
Because I think obviously for people listening here in the states or in any generally developed country where you just have raising capital, borrowing capital is relatively easy, I think they might not understand or grasp why that would be so difficult, whether it's sharing stories or just stats. Talk a little bit about why these founders and these companies have had such a challenge raising capital locally.
Mike Sall:
Well, I would say there's three components to it at a high level economic view. One is that in a lot of these markets, the governments offer high yield government bonds. And so the banks in these markets are perfectly happy to take bank deposits and go get these bonds, and they're not as interested in lending out locally. And so businesses that have great track records and great businesses still have trouble getting loans from the local banks because those banks are happy just to get the high yield government bonds. And then also just broadly speaking, the capital markets are relatively underdeveloped in these markets where there's just less capital, there aren't these vibrant VC or angel networks happening in these markets.
Mike Sall:
And so they usually need to look to foreign investors to get the capital that they need. And then talking to a number of other Western investors, for example, that are interested potentially investing in emerging markets, there's a couple areas. One is that they're just less familiar with the markets, so they're hesitant to jump in. Someone who's investing in the US is hesitant to jump right into Kenya or Nigeria or Indonesia. And then also there's a lot of hurdles. You have to really understand the local legal jurisdictions to alone and provide capital in these places, and how do you pursue legal recourse if something happens with the loan.
Mike Sall:
And so it requires a lot of work and that's why the kinds of folks that end up providing these loans are big institutions, really big funds. And they need to write 50 million plus checks because they have a whole group of lawyers who are understanding the legal systems in these other places to make it worth their time to go in there. And so you have smaller businesses in these countries that are looking for capital, but because the local capital markets are not as well developed and the banks are more interested in the government bonds, and then the institutional investors from abroad really need to write bigger checks to make it worth all the time that they spend doing those loans, it just becomes really difficult for them to get the capital.
Daniel Scrivner:
I wanted talk about some of the big ideas in Goldfinch, in the protocol and what you're building. And one of those is, we've danced around it a little bit, but this idea of importance of borrowing crypto without crypto collateral. So one the questions that I guess I want to ask is, in these calls you were initially having, did you bring up the prospect of, were people comfortable borrowing crypto? Was that a thing you explored and then you later realized, "Oh no, they just want local currency." What did that look like?
Mike Sall:
When we were talking to these businesses to see if they would be interested in using our protocol, we were trying to understand that. And one of the key parts was... There's different layers of it, which is, would they be willing to borrow Bitcoin and Ethereum? And absolutely not. They were not interested in that. But then it was like, "Would you be willing to borrow stablecoins like USDC, like USD pegged?" And they were open to doing that.
Daniel Scrivner:
Once you explained it.
Mike Sall:
Once we explained it. Some of them were familiar with it as well because stablecoins are becoming popular in a lot of these markets as well, but they were willing to do that because they are already used to receiving USD denominated loans, just like fiat. And going into crypto, they had some questions about, "Oh, what are the exchange rate risks on USDC? Is it definitely pegged to the dollar or not?" And so there was some education there to help them get comfortable with it. But one of the most positive signs we found was we were just concerned whether they would be willing to deal with crypto and set up MetaMask and do all of this crypto stuff.
Mike Sall:
And they were perfectly willing to do that as long as it meant that they could get the capital that they needed. So we definitely were trying to understand what was their willingness to jump through hoops in order to participate. And it was that they weren't going to take Bitcoin or Ethereum, but they were happy to take stablecoins like USDC.
Daniel Scrivner:
Yeah. And so I'm guessing that was the solution that you ended up with is basically your lending out or you're paying out USDC some stablecoin.
Mike Sall:
Yeah. The way that it works is the borrower goes to the protocol and they receive USDC, they need to borrow USDC and they pay back USDC into the protocol. But then all of the borrowers that we're talking to, they're getting account set up in exchanges like Binance and Coinbase and Kraken. All of these exchanges now are supporting a lot more of these markets around the world. And so they just hook up their bank account and they receive USDC into the exchange and they convert it to fiat or their local currency, and then they take that local currency out into their local bank.
Daniel Scrivner:
That's super interesting. The other piece of this, and this is a really big idea, and I want to try to flush this out for people because I think it may not be immediately obvious, but one of the things that's really novel about Goldfinch is you guys are trying to be the largest, truly decentralized lender in the history of the world. And today, if you look at how lending is traditionally done, it is never decentralized and it is always incredibly centralized where you have one central authority, typically a department of analysts, incredibly analysts and lawyers that are putting together these contracts. That seems really ambitious. How did that come up? And was there ever debate around how centralized or decentralized to make that?
Mike Sall:
Well, there was ever debate about whether to make it decentralized, the beauty of cryptos is all totally decentralized. And so the question was always, "Okay, we know it has to be decentralized, but how do you make this thing work in a decentralized way?" And I think of it as maybe individual banks are centralized, but we still have a broad banking system where many banks all around the world and they're all do doing their own things, and then even within those banks, there are different departments focused on different sectors and areas. And so how do we take this concept of different folks all around the world that have different expertise and different areas that they focus on and help organize that system into a protocol that can enable all of them to act?
Mike Sall:
And so that's the approach that we're taking with Goldfinch is this idea that you can have lots of different folks who are all spending their time, focus on different areas with different expertise and then create a place where they can bring that expertise, and they focus on the specific kinds of borrowers that they know, and they help get the right level of capital and the right credit to those folks, and then you bring them all together in one system. And so the net result is you have all these different folks in different places doing their own different types of credit analyses and it all comes together into one system. So that's how we approached it.
Daniel Scrivner:
Maybe to restate it, you guys weren't like, "How do we create a new lender?" You're literally bootstrapping a marketplace or a network or an ecosystem of basically people that are willing to lend, it could be different things, different types of businesses, which is fascinating. Talk a little bit about just the mechanics, and you can get as technical or as not technical as you want to because I know it can be pretty nuanced, but talk a little bit about the mechanics of how that works. Because I think when you hear it described, one, it makes a lot of sense. And it also this aha, why has this never happened before and yet can only be built in crypto? So that'll be interesting to explore that.
Mike Sall:
Yeah. I guess it'd be the mechanics of how the protocol works and how the different pieces fall into place.
Daniel Scrivner:
Yeah. And maybe just to set that up, give a real world scenario. Let's imagine this is an entrepreneur from Kenya that's going and trying to raise $15,000 for a local business. So they go, they submit this application on the platform, what happens next and how does that get green lit? And what does that look like? What does that process look like?
Mike Sall:
Well, what's interesting is there's multiple levels of borrowing that's happening. And I think we can hinge the story around a credit fund because that's the folks who are participating right now is take a credit fund, for example, Almavest. Almavest is this credit fund. And actually earlier this week, they just raised 10 million through the protocol. And just to describe who they are, they are a credit fund that is experienced with evaluating these different fintech businesses around the world. For example, one business they worked with is a company called Oya in Ghana. And what they do is they provide loans to small and medium-size businesses in Africa.
Mike Sall:
And so then what you have is a credit fund who is looking for capital and then they extend a loan to a business Oya in Ghana who is then working with loss of great businesses who are looking for capital to grow their businesses. So there's multiple layers of the borrowers here. But the great thing about working with a credit fund like Almavest is they have this great track record of finding amazing businesses around the world that are looking for credit lines and they can put a lot of capital to work. So they can raise $10 million and then extend multiple loans out to folks after that point.
Mike Sall:
So the way that the protocol works is Almavest goes to the community and they say, "We want $10 million. We want a $10 million loan," in their case, with a 12.5% interest rate. And they propose it to the community of their backers. So the backers is the main participant that understands who are these different potential borrowers to receive credit lines. And the backers, they sign an NDA with Almavest and they get access to a data room, and then they enter a chat room just with the folks who run Almavest. And they're asking them all these questions to understand, how does Almavest approach the loans that they give out, what's their philosophy?
Daniel Scrivner:
And this is all on platform, just to clarify really quickly?
Mike Sall:
Well, this is all through whatever Almavest wants to use. So they've used a telegram channel in the past. They set up a data room and a Notion doc. And this is the beauty of crypto, which is that they're doing all this stuff all on their own with their own infrastructure and they're talking directly to this community of backers. And then what Almavest then does is they create what we call a pool on the protocol. And this pool is a place for the backers to say, "Yep, we're going to put capital in here." And what the backers do is they provide capital, it goes directly to the pool which Almavest has access to.
Mike Sall:
And then in the process, they also sign an agreement directly with Almavest. They sign a loan agreement essentially that gives them legal recourse to the capital that Almavest is expected to pay back. And then the way that the protocol works is each of these pools has two tranches. And so there's a junior tranche that is first lost capital, which means if for whatever reason a borrower doesn't pay back the money, the first folks to lose their money are in the junior tranche, and then there's a senior tranche. So when this community of backers talking to Almavest and they provide capital, they're all providing capital into the junior tranche of that pool.
Mike Sall:
And then what the protocol has is a separate senior pool, which is just taking capital from folks that just want to have it be automatically diversified across different borrower pools. And that senior pool gets automatically allocated by the protocol to the senior tranches of these different pools. And the way it does is it looks at how many of the backers have participated in a particular pool, and if there's a lot of folks who have participated in it, the senior pool will provide more capital into that senior tranche. And it's an automatic algorithm that's based upon the number of backers.
Mike Sall:
And so the net result is that there's the senior tranche that has capital and that senior tranche will provide 20% of the interest it receives to the backers. So it increases the interest that the backers receive. And that means that the backers are receiving an outsized yield. And that is compensating them both for doing all the work of evaluating these different borrowers and providing the first capital into the pool as well as taking on higher risk by being the first loss. And so that's the key part of it is when the senior pool is automatically allocating capital, it makes the economics work for the backers to do all of this work of understanding folks like Almavest, and being willing to put in first lost capital.
Mike Sall:
And so the result that happened for example earlier this week is the backers provided a bunch of capital and then the senior pool provide additional capital and Almavest was able to get the full credit line that they were looking for.
Daniel Scrivner:
That's so cool. And I'm sure it's not easy for you to remember all those details and paint that picture, because there's quite a lot that's happening there, but it's also very elegant in that it's not all that complicated. I'm curious, how long did it take you guys to arrive at an elegant solution and align all those incentives? Is this something that you did in a sprint, you did leading up to the white paper or is this something that you've literally have been iterating on the entire time you've been building out the protocol?
Mike Sall:
It has been pretty iterative. What we started out doing when we first started the company was just, "Can we just build a pool that has money in it that we can get one of the early customers to try out and connect with MetaMask and just get capital from it." And so there wasn't any of this logic, basically at the very beginning, we were just doing test versions of things to see, can we even get capital in there and can folks draw down? And then as we were going, we started to get more clarity of exactly what are the problems here? What are the challenges? What incentivizes these different kinds of folks to participate? Which helped us clue in on exactly what are the pieces of logic that make it all fit together?
Mike Sall:
So I would say at the beginning we had broad strokes things figured out enough that we could build a minimum viable version of it. And then over time, getting more and more detailed until the white paper was published.
Daniel Scrivner:
I want to ask a separate question about the white paper in just a sec, but I want to ask a meta question, which is, that initial version obviously sounds very simplistic. You guys have iterated on that. How do you approach when you draw the line and say, "Okay, we're ready to launch," knowing that you're going to continue iterating after that. Have you used a framework? Is that just a discussion? Because that seems very challenging because it's something, especially here, it's very nuanced and intricate. How do you decide when it's ready to ship, it's done?
Mike Sall:
It's interesting with crypto versus like a traditional Web2 product, where crypto is a lot harder to be super iterative. You have to be really careful and you have to do things in larger blocks at a time because there's so much work that goes into it and there's so much more risk about having mistakes. But the general framework we would take is say, "In the next iteration of this, what's the key thing we need to learn? What's the main thing that we have to figure out?" And it's usually like, "Will this a participant be willing to do this?" Or, "Is this going to encourage these other folks to participate in this way?"
Mike Sall:
And then the question is, "Okay, what is the smallest possible version that is going to allow us to answer this very specific question?" And then that's what we would build. Build everything that we need to answer this question and nothing else. That would be how we would do it. And that's still how it's been going up until this point of fully launching.
Daniel Scrivner:
It's really smart. And that's obviously a huge parallel there between raising venture capital where the idea is, what is this fundraising unlock? What are you going to learn with this capital to then go and raise additional capital? I want to switch to that and ask a question around raising capital. So I want to just ask this in the broad sense because I know you founded a company before, was that company in crypto, not in crypto?
Mike Sall:
No. I have attempted to found multiple other companies in the past and none of them did very well at all. This is the fifth time I've tried starting a company and this is the only one that really, really got off the ground.
Daniel Scrivner:
Wow. I have to ask a question on that. What did you learn? That's staggering because I think that Goldfinch will be incredibly successful. Just knowing about you, knowing about the quality of the people on the team, I think you guys are going to have a lot of success. You've now done this five times, you've had this time take off, what did you learn from trying and failing four times.
Mike Sall:
I learned how many different things are just required. Say, what are this? All of the things that are super important is having great people to work with, with shared values and shared set of goals, so that we're for sure on the same page of what we're trying to build. And then I would say in the past I was too lazy about doing the market research early on. And I think we had like 100, 200 calls before this one, just to make sure we knew exactly who the customers were going to be, exactly who we're building for and doing just the hard work of research early on to know that there's really a market here.
Mike Sall:
And then also when it comes to fundraising, for example, it's also important to go into a large market that is going to be compelling for investors to invest into, or it's actually not that raising money is necessary, but then there has to be another path to starting a company if it's not going to be in a market that is compelling to VCs who need to see astronomical returns, like a coffee shop is a great business to start, but yet, you just have to make sure you have the capital, you need to start that in a different way. And I think in the past I was trying to start things hoping that VCs would want to invest in it, but they're just not big markets.
Mike Sall:
And so there are a bunch of those pieces, and then also knowing what we want to build and how it was going to work. And so I think the stuff that I learned was there were just these big core critical parts of the past times that I tried to start companies that they weren't in place. And so then this time I was like, "Okay, we got to make sure if we need to raise funding, which we needed to, it's going to be a big enough market that investors are going to be interested in it. And we know who the customers are going to be and we're aligned on our values and our goals for the company." And so all of those things have to be aligned for us to even get started.
Daniel Scrivner:
Yeah. I love that list you started off with of even just that note, because I've also found that extremely rare, just talking with founders and looking at a bunch of different companies to invest in, it is actually extremely rare to find people that have done even 100 customer calls, because that is a marathon in so many ways. And it's funny because actually now that you said that, it makes me immediately think anytime I've ever or we've ever, if it's a part of a fund I'm part of we've invested in a company that's done that amount of market research, that always tends to go really well. So it's surprisingly correlated, I think with success after that, of just putting in that groundwork.
Mike Sall:
Yeah. In the process of all those calls, we changed what we were planning to build and how we were going to build it and who we were going to focus on. So it totally had a big impact on what we ended up doing because we learned a lot through it. So it's a lot of work to spend all that time upfront, but it's critical I think.
Daniel Scrivner:
Yeah. I want to ask a question, what was it to raise capital in this market for a company? And I know one of the firms that you raised from is a16z, incredibly well respected. So just general thoughts on a16z, what has it been like to work with them and the team there?
Mike Sall:
They're great. It's a combination of really smart people to talk to. Ariana is the main person that we work with her and she's great. And there's just so many people on the a16z team to talk to who are great. And then they have all these resources as well, which is also good. And then they have this great reputation as well that it just lends and it's an immediate credibility with other folks. So it's been really great working with them. And I think it's also interesting because in the process of fundraising, both when we initially did it for our pre-seed and the seed and recent raises, it's a similar thing where investors that have high conviction really stand out like a16z.
Mike Sall:
And at the time, it feels like everybody else said no except for them. We only need one great investor who really has conviction. And so it means a lot when an investor demonstrates that. And we also have the same thing in our pre-seed round and it also comes after, I don't know, we get tons of nos from everybody else until you get the one yes from an investor with high conviction and really appreciate that too. They believe in what we're doing. I feel the same way about all of our investors and it makes a big difference for us.
Daniel Scrivner:
It's really well said. I think the best quote I've heard on that is raising around just requires a market of one, just requires one yes. But it can take a lot of nos to get to that one yes. And then there are different types of yeses, to your point. There's the yes of just, "Yeah, I have some capital." And there's the, "I really want to invest in what you're building and your team and be a part of this." I want to ask a naïve question around the white paper, which is, I've never gone through the process of putting together a white paper myself and you would think, or one would think maybe that that's just almost like an existential thought exercise.
Daniel Scrivner:
You just have this initial idea, you have all these customer calls, then you go and you spend a week writing out what this is. I imagine that's not it at all in that you're potentially building and parallel solving some of the technical things before you publish the white paper. Can you give a sense, even just at a really high level of what the sequence is and when you're ready to actually write that white paper and how that comes together?
Mike Sall:
Let's see. It was a gradual process with many folks. I don't want to make it sound like we just wrote it because it was community driven. We were getting feedback from all different kinds of folks who were contributing in different ways. And it was really starting out with... I guess the way it worked is early on when we were just doing prototypes, we weren't as worried about the white paper. We had internal docs of, "Here's a rough outline of how we think this could work, but we're just going to get a prototype in." But then once we wanted to start, bringing more people on, more the community to get involved, it becomes a lot more important to build it together, write the white paper together and have this be more of a community oriented thing.
Mike Sall:
So that was where it was like we had certain outlines of how it could work and then start getting multiple people to provide feedback and digging into different specific areas and then collating it together until we had an initial version of the white paper that we could share with everyone.
Daniel Scrivner:
And I know you guys are now, it's like.1. So I'm guessing there's also some dot-update rev that could happen on a white paper as well too.
Mike Sall:
Yeah. Now the white paper is also on the GitHub open source and people can propose different kinds of changes to it. And I think with 1.1, it was recognizing certain things that had to just be changed. I think there was even a typo, we had to change it, but it's a different inversion of the white paper. So it was from incorporating even more feedback. And then that will just keep happening, especially as we launch the full governance process, which is happening now in early 2022, is now there's going to be a whole system for submitting proposals and having broader votes on exactly how these changes get made. And so those of all get incorporated into future versions of the white paper.
Daniel Scrivner:
Super interesting. Thanks for indulging my curiosity there. I'll get out of the weeds now, I want to go way, way back up a little bit more meta and just ask you a couple of questions of what you've learned, what your advice would be for other people that are setting out to build a new protocol. The first question that I have is, knowing, especially now, I didn't know that coming into this, knowing that you've had four tries before at founding a company and you have this one now, and I want to say thank you for sharing that because that's something that I know a lot of people might hold close. I really appreciate, I think everyone listening really appreciate that you're open about that. So thank you.
Daniel Scrivner:
The question I wanted to ask was, what have you learned and what has this process been like for you? This is both the fifth attempt, you're the CEO, this is a really ambitious project. Just reflect a little bit on what this journey's been like for you and what you've changed or learned, or I don't know, with the last 18 months have taught you.
Mike Sall:
Let's see. I guess in the context of starting businesses and in the context of having really, really struggled in the past trying to do so, it feels different this time around. And I think the thing that feels the most different has been seeing the clear demand for it, having customers really want to use it and having people just reaching out, about using it and having other customers refer other customers. And that is something that feels a lot pretty different from before because on prior things that I'd worked on, it was the same like, "Oh, I to build things, and so I'm going to build things with people that I like working with." But we didn't really know exactly how it was going to end up going out into the market.
Mike Sall:
And then this time it feels different. And I think that is owed a lot to the early work we did into doing market research and actually just lining up specific customers. That was one of the things I told myself after one of the prior failures, is I was like, "You know what, next time I start a company, I need to know who all of the initial customers are going to be before I spend any time working on it." And that was a personal requirement. And then we did that. And then after we did that, just seeing more demand in certain areas and then adapting to where we're seeing the demand from the market.
Mike Sall:
So I think having in the past been in a place where it felt like a struggle to find who's going to going to use this, but it seems like it could be really cool, but also I don't know who's going to use it, that was a big difference this time around. It's very clear, the market is pulling us to build these things for them. And so I think that's what stands out to me,
Daniel Scrivner:
Which is really gratifying because I've definitely been in positions before where you just feel like you're excited about it, you're giving it your entire best effort to try to make this thing incredible and it's just like you're running into a wall, no one wants it. And it feels ultimately one of those unsolvable problems, I want to ask a two side of the coin question, which is, thinking back over the last 18 months, what's been the most exciting, thrilling part of it for you? And this could be a specific thing you worked on just a really memorable moment. And what's maybe a tough lesson that you've learned? And this could be something on the technical side, something just super meta in general.
Mike Sall:
One of the really exciting things that happened was we shared this program that we called Flight Academy, which was a way for backers, folks who are actively evaluating the different pools and which ne to participate in to make a program to help backers understand how to use the protocol and participate in it. And we were planning it out and we're like, "Okay, we're hoping that we'll have a few hundred people participate in this unless I create these educational modules for them. And then we should really design this in a way where we could see this getting to maybe 10,000 people over the next year or two a kind of thing."
Mike Sall:
And then we shared it and it resonated with a bunch of folks. And we had over 30,000 people come in and sign up to participate right in the first week. And it was gratifying both just to see how many people were showing up. But they also really cared about what we were talking about, and the story, and this goal of expanding access to capital around the world and helping real world businesses and emerging markets and all of this was really resonating. And so that was one of the coolest experience so far in the last 18 months was just seeing how excited people were to participate it, having it go away past what our expectations were in that moment.
Mike Sall:
And then they've been active, making memes and creating their own content and helping to really build out the community. So that has been the most gratifying part of it so far. And then I think the challenges, is that your question or?
Daniel Scrivner:
Yeah. I think just something I always try to reflect on, just thinking back in time is, I don't know, just making sure, obviously there's stuff that goes really well, but also just what did you not expect that hit you in the face or you encountered midway through solving the problem? Just anything there that was just a major lesson learned or a major obstacle you had to overcome.
Mike Sall:
Yeah. I would say being in crypto where things aren't regulated and it's unclear what's going to happen in terms of regulation, and we really want to be fully compliant, and it's hard to do that when the laws aren't in place exactly what's expected, but trying to understand where are we going to be and what do we need to do to be compliant? And we're taking a lot of efforts, for example, everyone who uses the protocol is going through a KYC process, which is not that common in crypto, but it's something that we're doing to make sure it's totally compliant. But one thing is we always have just underestimated how much work that is and what that process is working with counsel when they're trying to figure it out alongside us.
Mike Sall:
And so there have been a lot of, oh, shoot moments with lawyers when we're saying, "This is what we want to do," and they're just like, "We don't know how you get comfortable with that until we rearrange things a little bit to get it into a better place." And so we have just been consistently underestimating the time involved to go through with legal counsel, how to approach different things, and the amount of rethinking on different aspects. And that amount of work on its own is a significant part and one of our biggest challenges is how to get to a place where we feel like, "Okay, this is now a compliant way of going about this process."
Daniel Scrivner:
It makes so much sense. And obviously you guys are pioneering something. So I totally know what those calls are like when you're talking about something that's truly novel with lawyers or a legal team and they're like, "I don't even know where to start." It's like they have to take a second to sit down, wrap their heads around the problem. This has been a fascinating conversation. Thank you so much for sharing this. I love that you guys are building something that honestly feels, I think, what crypto in general has wanted for a long time, which is one, a massive impact on the actual real world, including people that may not be all that interested or all that crypto native, or all that just even part of the crypto community. The fact that the world at large can benefit from it. Any closing words, closing thoughts before we wrap up?
Mike Sall:
I'm super excited about it. I feel it's time for DeFi and crypto start having this impact. I really do think over the next year, over 2022, this is going to become more and more of a thing across Goldfinch, other protocols as well. And I hope that we can help show that this is the next wave of crypto that really has this kind of impact.
Daniel Scrivner:
And that lending can be done truly in a decentralized way as much as it can in a protocol. So for anyone listening, I know they can find Goldfinch online at goldfinch.finance. What's your handle on Twitter? How can people follow you on Twitter?
Mike Sall:
It's goldfinch_fi.
Daniel Scrivner:
Okay. goldfinch_fi. And how do people follow you as well too?
Mike Sall:
On Twitter, I'm @sall, S-A-L-L.
Daniel Scrivner:
Thank you so much, Mike. This has been a fantastic chat.
Mike Sall:
Thank you. Thanks for having me.
On Outlier Academy, Daniel Scrivner explores the tactics, routines, and habits of world-class performers working at the edge—in business, investing, entertainment, and more. In each episode, he decodes what they've mastered and what they've learned along the way. Start learning from the world’s best today.
Explore all episodes of Outlier Academy, be the first to hear about new episodes, and subscribe on your favorite podcast platform.
Daniel Scrivner and Mighty Publishing LLC own the copyright in and to all content in and transcripts of the Outlier Academy podcast, with all rights reserved, including Daniel’s right of publicity.