Auren Hoffman  0:01  

Welcome to World of DaaS a show for data enthusiast. I'm your host Auren Hoffman CEO of SafeGraph. For more conversations, videos, and transcripts, visit safegraph.com/podcasts.

Hello, fellow data nerds. My guest today is Johnathan Levin. Jonathan is co-founder and Chief Strategy Officer of Chainalysis, a blockchain analytics company which is valued at over $4.2 billion. Jonathan, welcome to World of DaaS.

Jonathan Levin  0:32  

Thanks so much. Great to be here.

Auren Hoffman  0:34  

Now you run a super interesting blockchain analysis company. You're sorting through like tons and tons of transactions helping businesses, financial institutions, governments understand like how people use cryptocurrency and other types of tokens. But blockchain, in some ways, it's like a public ledger. So did you just start by like doing some deep analysis on the publicly available data? Is that how it started?

Jonathan Levin  1:00  

Yeah, so maybe I'll rewind a little bit and give people a little bit of background about myself. So I came across cryptocurrencies over a decade ago while I was studying in Oxford looking at sort of the economics behind cryptocurrencies. And what I realized at that time was, yeah the blockchain, as you say, is this open record of transactions that anyone can look at. That was kind of an economist’s dream. So when I looked at it, I said well there needs to be someone that understands the most about how and why people are actually using this ledger. So the ledger contains all of these transactions, but it doesn't actually say which entity was putting those transactions into the blockchain. It doesn't say what the purpose was behind these transactions. So being able to label that information was really the challenge that everyone had. That was really the starting point of Chainalysis was to say, could we be the company that really understands the most about how and why people are using this technology? Can we label really the data that's publicly available that everyone is interested in with the context about which entities put those transactions? 

Auren Hoffman  2:21  

How important is to go all the way back to like t-zero—time equals zero—on a given cryptocurrency, whether it be Bitcoin or Ethereum or Salon, etc.? Or is it okay to just start at some period of time?

Jonathan Levin  2:34  

Yeah, so it's a really good question. I think it gets into sort of the data half-life discussion depending on what data business you're in. I think the thing to know about blockchain in particular is that the actual existence of the transactions that people could be interested in are going to be perfectly preserved for all time. That is the thing behind the technology that blockchain does really, really well is preserves all the records for all time so that someone 20 years from now could refer back to a transaction that happened 10 years previous to now.

Auren Hoffman  3:13  

Yeah, assuming that that particular coin doesn't go away or something like that, which sometimes they may.

Jonathan Levin  3:18  

Sometimes they may. Oftentimes that record, even then even if it even if it goes down to almost zero usage, there'll be a copy of that ledger somewhere existing in like an archive that someone will need to go dig out and sort of a sci-fi film. The point here is that information is always there, which means actually that it actually extends the half-life of the information that we collect because the information we collect is ephemeral in nature. Because knowing which entity was around at the time five years ago, say, transacting, that entity may no longer exist in today's world. So if we were there at the time making a collection from that entity at that point in time, that data is now no longer available to be collected.

Auren Hoffman  4:10  

You're somehow mapping these wallets to real life entities, or as much as you can. Maybe it's fragments of that real life entity, but you're trying to get some sort of data points about those wallets. Is that right?

Jonathan Levin  4:22  

Yeah, that's exactly right. I mean, I would say that, yeah. Until today, and just let's use Bitcoin as sort of the canonical example, is most of the transactions that exist in the Bitcoin ledger have been put there by services that individuals use to transact. 

Auren Hoffman 4:43

Like a Coinbase or something like that. 

Jonathan Levin 4:47

Like a Coinbase like, like a BitPay to pay for goods and services. So you could be trading, but all of those transactions therefore are being put there by services. Those services have come and gone over the history of Bitcoin. 

Auren Hoffman  5:00  

Like Mt. Gox is no longer involved, right? 

Jonathan Levin  5:03  

Yeah. So if you were trying to collect the Mt. Gox wallet information today, you can't because Mt. Gox is no longer around. And so we have a very large sort of effort for the last seven years really building that comprehensive collection mechanism to collect all of this ephemeral data. Building that database of which of those services have put which transactions in the blockchain. 

Auren Hoffman  5:31  

How do you inset those services to work with you because it's a bit of a data coop I imagine, right?

Jonathan Levin  5:37  

Yeah. So initially, like, we had no incentive for anyone to participate in the in the data coop, which is kind of the chicken in the egg problem that that all businesses face. The interesting thing though we had was we had the ability to go out and use these services ourselves, and start to create accounts at all of these different service providers. Then understand which of our transactions, how did they show up in the blockchain? Then start to run like pattern algorithms over that so that we could we could capture most of the activity ourselves, and then go to these customers and say, “Well, hey, we've already mapped out the wallet that you have, and a whole bunch of wallets that you're interested in. Do you want to use us to be able to manage?” 

Auren Hoffman  6:25  

It’s a bit like how Plaid got started out where Plaid is essentially starting out like scraping these banks. Then says hey to the bank, “It'd be much better if we just did an API between us.” Ah, okay.

Jonathan Levin  6:34  

Yeah. I think that, though we have something maybe additional there which is not just the convenience and the trust that gets built by an API integration. Like actually these businesses need us to identify money laundering risks on their platform. And so they have a real utility to stay in business if they use us for compliance purposes, which allows us to then get data rights on which transactions they're processing because they really need us across all the other entities that they're transacting with. They need to understand what their customers are doing with their crypto.

Auren Hoffman  7:20  

You have a lot of different types of customers. One type of customer is doing things for compliance. You have another type of customer that might be doing things for like law enforcement or something. In compliance, I would imagine the bar for being accurate is super high. Like you really want to make sure it's high. Whereas in law enforcement, it's like might be okay if you gave a lead with 20% accuracy. They could like run down the lead or something, right. So how do you balance those different types of needs for accuracy?

Jonathan Levin  7:48  

Yeah, so actually the interesting thing about the accuracy for the government—And you're right on the sort of lead generation side of law enforcement. But actually, we get used all the way end to end in these investigations where you're not just worried about show me some potential avenues of investigation. We're actually all the way used down to really proving beyond reasonable doubt that someone get in financially from a set of activity, and that's a very high bar.

Auren Hoffman  8:20  

Okay. That's beyond reasonable doubt. Okay. So super high. 

Jonathan Levin  8:23  

Yeah, so that’s a super high bar on the accuracy front on the law enforcement side. But you're right that like even within our law enforcement customer base, we have these different stages of an investigation where you want to dial that confidence level up or down depending on like whether you're looking for more leads or whether you're looking for being able to prove to a jury that someone has gotten ill-gotten gains. So I think that we tend to think about providing sort of a very—You know like the hype. I don't know how to say this. But instead of the lowest common denominator, the highest common factor. Like we think about, like making sure that our ground truth data and the picture that we give to all customers is at a very high bar of confidence. Then start to think about more behavioral analytics around the set of ground truth data that we can we can give people sort of the lower confidence, decision making potential for those types of use cases. You're right on the compliance side. That actually also we're starting to power decisions that the business is taking. So if I'm a company allowing users to buy and sell crypto and take it on and off my platform, I'm interested in what my competitive dynamics are. Are they moving their money off my platform to a platform that has a certain set of features or stuff like that? So those entities, the accuracy there is not important. Yeah. Completeness is much more important.

Auren Hoffman  10:04  

So okay, there's Bitcoin. There's Ethereum. There's Salon. Okay, but how do you know which one to go into? Is it just driven by your customer demand? Like customers are like I really care about the coin number 52 or something?

Jonathan Levin  10:16  

Yeah. So we have a bit of a range here. I think one of the things that we've done over time is really matured sort of our approach to this where actually, to your point about the different use cases and needing certain different things, we've actually managed to create a solution where we actually have a very low cost of supporting new coins for really just risk screening. A quick dirty check. Like we've had to really lower our cost of what we call coin acquisition down to virtually zero where we--

Auren Hoffman  10:52  

You're just like ingesting the ledger and doing some basic analysis of it.

Jonathan Levin  10:56  

Yeah. And even before we do that, we can even start to tag up addresses, even if we don't even have the underlying ledger.

Auren Hoffman  11:03  

Often it came from a Bitcoin wallet or an Ethereum wallet or something like that. Usually to start some of these coins, it's coming from some other wallet somewhere else, right?

Jonathan Levin  11:12  

Yeah. Or someone reports to us now. Like we have scam reports that people send to us and pieces of information that our partners give us where we say, “Okay, well, we'll flag all of those in case anyone else in our network is checking those transactions.” So we've had to really lower that for that use case. Now, to your point, when you get to the sort of level of sophistication where it needs to be in all of our products at full support, then we discern sort of  where customer demand is, and also where we think the market is heading. So we look at sort of where is the majority of economic activity happening in these blockchains. Like looking at the number of transactions, looking at the total outstanding value of these coins, looking at those types of metrics to guide sort of how we're thinking about that for the full support at the top level. 

Auren Hoffman  12:06  

Interesting. The thing about Bitcoin, really talking about a relatively small number of transactions with better high value transactions that are happening. Whereas you could imagine scenarios with other blockchain systems where you're going to see much, much higher volume of transactions with lower dollars because the transaction costs potentially are lower. I can imagine that could change the way you do analysis. Like how are you reacting to that?

Jonathan Levin  12:34  

It's a fascinating part of the industry, right. I think if you look at the last 10 years of crypto, we've sort of spent the time building what you could call like financial primitives and really creating an asset class. Which to your point has sort of seen a massive appreciation, and a lot of the activity is centered around stores of value and like very high value transactions. I think what you're seeing today in crypto is sort of an opposite trend. Which is using some of those financial primitives, you're now seeing people build marketplaces of art and music rights, distribution and ticketing, lots of consumer applications. Which to your point is definitely going to need sort of consumer internet scale behind it. We do think that over sort of a medium term, that's something that Chainalysis needs to be well equipped for. So thinking about how you deal with what people call like high throughput blockchains is something that we have really sort of been working on for a couple of years.

Auren Hoffman  13:47  

Especially if some of these things happen off chains. Like if you can imagine a scenario where you're going to be doing tons of transactions. You might do it off chain, and then you might settle it on a block or something like that.

Jonathan Levin  13:59  

Yeah, and I think for us we've been very focused on sort of the unique attributes of the blockchain, which allow us to build the business that we have, which is this—We're very interested in providing analysis on where you have the same set of transactions that everyone in the world is interested in, right? That's really the core value proposition. As soon as people start moving information off chain, which even happens today in crypto, but those are…. There are definitely interesting use cases there, but we tend to stick away from them because we're fully about how do we provide the best context for the set of transactions that the most people in the world are going to be interested in.

Auren Hoffman  14:45  

One of the things I love about Chainalysis is that you're this data coopt. Data coops have a lot of like winner take most properties to them. If you're going to give advice to another company starting a data coop, because they're really hard to get going, like what advice would you give to another company? What are some of the non-obvious lessons?

Jonathan Levin  15:02  

Yeah, I think from my perspective, the non-obvious thing is that you do need to have some sort of compelling reason why someone is going to enter this into you. Don't take this with rose tinted lenses and say if everyone did this, that this would be an amazing outcome.

Auren Hoffman  15:25  

For number one through 100, it's very hard for them to get started. They're not getting that much value back. Right?

Jonathan Levin  15:29  

Very hard. And I will tell you that like even as you scale a company, and like we look for new use cases today, I have to push back on that with our stuff all the time. Like really tell me why giving us this piece of information is a required capability that the customer could articulate to you?

Auren Hoffman  15:52  

Why are they getting a clear benefit from sharing this? Why is this really helping them? Okay, got it.

Jonathan Levin  15:57  

Right. And so that's one lesson. I would say the second lesson to me is that they're like very, very difficult to do this if you don't have any sort of external lever that you have control over that you're able to push into this data coop model, right? So the analogy that I would have is like if the business that you're starting really depends on you going around to 25/30 businesses and everyone needs to pour out what they've got in the data silo into the middle, you're going to have a really hard time doing that BD run. 

Auren Hoffman  16:30  

Historically like Verisk or something, the way they did that is they were like started by these companies or something, right? They actually—Or Visa or something. They owned a piece of the company.

Jonathan Levin  16:41  

Right. So being able to construct that model is like probably the hardest thing to actually get done. 

Auren Hoffman  16:46  

Yes. Just top down where you'd like the CEO or something, right?

Jonathan Levin  16:50  

Right, exactly. So instead, like look for angles where you definitely want to do that, but that you have some form of outside information that is valuable to the coop that you can that you can dial up and down depending on how much extra effort you need to have done. Like a good example would be like, okay in a new sort of cyber fraud coop that you want to start in a particular region of the world, how do you think about getting some datasets that maybe a lot of the people in that coop already have access to? How do you get it to them cheaper? How do you provide some form of extra analysis on top of what everyone's using today that makes the data more valuable that then allows you to then create some incentive and also some trust factor with those with those entities to be able to do that. With us, it was super straightforward. Like there was this resource that was in the middle that everyone was putting information in already, and we managed to feed it back to them with this extra analysis that we were doing ourselves.

Auren Hoffman  18:01  

Okay, so let's just rift. So let's say you and I, on the side we're going to start a new company. We're going to build a data coop or we're going to a data company to help companies understand days outstanding for all their new customers. So every time you get a new customer, we tell you the average days outstanding that everybody has for this. So let's say you get Ford, and turns out Ford pays their bills a lot faster than General Motors or something like that. We try to like connect to everyone's GL. Like what advice would you have for the founders of this company to get that movie? Because it's gonna be hard to get people to give you a connection to their GL.

Jonathan Levin  18:38  

Yeah, so I would say like this. Let’s figure out who are the big bill payment companies in in North America? 

Auren Hoffman  18:47  

Like QuickBooks or something, yeah. 

Jonathan Levin  18:49  

Let's go to bill.com. Because like bill.com’s business is becoming more ubiquitous and getting more customers. They're not really in the business of providing that as a data product to merchants who are needing to get paid. Although, I'm definitely going to give that back to the bill.com people. 

Auren Hoffman  19:13  

We just increased their value like 4x on this podcast. Everyone invest now. 

Jonathan Levin  19:17  

Yeah. Yeah, I'm gonna send it to the CRO. So the way that—And they're killing it, right. So. But the way I would think about it is are there some even anonymized and aggregated forms of information that you can go and uniquely collect from maybe a place that people weren't thinking about before? In order to then go and say, like, “Hey like can we do a benchmarking study on your GL versus what we're seeing in the general market?” Show like, “Hey, I've managed to uniquely connect this interesting benchmark, and you may have some room here to maybe open up some free cash flow or do something that that someone can do.” But that that would be my first approach is go to find like a unique source of information where possibly people aren't using that today, and run some analysis. Then walk around the market and see who sees interested in that set of information.

Auren Hoffman  20:24  

All right, cool. Now historically, like venture capitalists have been a little bit wary of data companies, and raising money has historically been difficult for data companies. I know for SafeGraph, when we raised our series B like we had a great outcome, but it wasn't easy. Like a lot of VCs said, “I'm not interested in data businesses.” Now, Chainalysis seems to be maybe proof that this is changing. You guys have raised, I think last time I checked, $365 million, $4.2 billion valuation, which is amazing. Like, how do you see that evolving?

Jonathan Levin  20:51  

Yeah, so I think that we we've been through sort of a journey and in sort of enterprise SAAS where VCs have been very focused on—particularly in the VCs that would be focused on these type of companies—very focused on sort of enterprise SAAS multiples. You've seen sort of amazing outcomes left and right in the enterprise software space. The thing that everyone in the VC world has to wrap their heads around is how proprietary is the access to information that a data company has? Because if that is the primary asset, like there needs to be a really compelling story on not just…Well, firstly, it needs to be proprietary. Like there has to be a reason why you as a company are the only the only company that can go out and collect all of this information. And then second, they need to understand sort of like your distribution strategy and how that further compounds your data advantage. And they'll want to ask like those…Like they're just not used to--

Auren Hoffman  22:02  

Like a data coop type of thing.

Jonathan Levin  22:05  

Like a data coop type thing because they're really trying to answer well, if it is so proprietary and you've got really great distribution, do you also have a long enough half-life to monetize this incredible advantage that you've got today? And those types of questions are just not very natural to investors that have been very focused on net dollar retention and churn and sort of the mechanics and measurement of that. I think if you look at even SAAS, right. I mean SAAS really born somewhere in the mid-2000s, and now like there were no metrics then to measure like the s ones that were being made. There were just like completely new metrics. 

Auren Hoffman  22:54  

Right. Even the ARR is relatively new or the CAC, LTV, or all the all the traditional things, efficiency scores, etc.

Jonathan Levin  23:00  

Yeah. And so now like 15 years in, you're very conversant in those metrics, and everyone is expected to be able to reel them off.

Auren Hoffman  23:08  

And we have things like SaaStr and all these other great resources that have been really helpful to all of us, and great bloggers on it.

Jonathan Levin  23:14  

Totally. So like, how do you measure coverage, right? Like, how do you measure…

Auren Hoffman 23:20

Really hard. 

Johnathan Levin 23:22

Like, really hard. How do you measure sort of like what degree of distribution someone has in a market? How do you measure data half-life? Like, objectively from the outside?

Auren Hoffman  23:34  

I mean generally VCs are not good at looking at products, right? If you think of a typical Series B investor, they're gonna look at your metrics. Even in the SaaS software, they maybe not even use the software. They may even not know what it looks like, right? So they're probably not looking at these types of things. Obviously, data is even harder to see. You can't get like a demo necessarily of data.

Jonathan Levin  23:55  

Yeah. I think like you know this very well from your go to market, but like structuring data bakeoff between two products is oftentimes how someone's making a decision in the in the DaaS world. That's not really something that a VC is particularly well equipped to do today. So what I think about is… Firstly, each business does need to have very clear North Star metrics on their own data advantage, and be able to tell the story eloquently about why they have proprietary access, what is their distribution strategy? What is the half-life look like? Like if that is very clear and you can paint a really strong story, I think you can get someone over the hump today and like people will start to understand. I think that the thing that we've managed to do very successfully is combine it with then enterprise software metrics that that we've owned some of the workflows associated to some of the decisions that people take based on our data. And this is where like, obviously, everyone's business is very different. But I said we need to really own also the workflow around investigations when it comes to crypto. Like no government agency in the world should want to investigate cryptocurrency related criminal activity without the use of our product reactor. No one in our compliance world should want to really be able to like monitor for money laundering risk without actually using our API's to screen their transactions. So those things have helped us paint the story of like, look, the real story is the data. We are a data company. We've amassed this incredible advantage. But ultimately—And we will choose which of these workflows we need to dominate over time as they become big enough markets. And as we can see people are super sticky in workflow when it comes with the addition of proprietary data they get used to having at their fingertips.

Auren Hoffman  26:19  

Now pushing back on that proprietary thing a bit. Like, if you think of ZoomInfo, which is probably the most successful data company that has been started the last 15 years or so. They sell business contact information that could just be seen on a—It's basically what you see on a business card. The name, the title, the phone number, the email address,. It doesn't seem like that proprietary per se. Now they say, “Well we have more accurate, and we have more breath than anybody.” That's probably how they get a lot of their customers. But there's probably other ways to piece it together. So how do you think of like that type of data business?

Jonathan Levin  26:55  

Yeah, I think that it's an interesting question, right? Because there is a level of proprietary, at least methodology. 

Auren Hoffman 27:04

Yeah cleaning the data. 

Jonathan Levin 27:07 

Then the analytics can. So they can be, but then then I think the point is I mean you are obviously a data business, but you're going up a level where there's some proprietary analytics that you're able to run that actually other people cannot. 

Auren Hoffman  27:24  

Or scale maybe matters. Like Walmart can buy toilet paper at a cheaper price. Like they might be able to make that marginal next business contact at a cheaper price than other people, and then they could sell it. So scale kind of matters potentially.

Jonathan Levin  27:38  

But that's super vulnerable, right? So I think that in today's world, and this is why like I think—And obviously ZoomInfo is an awesome company and has paved the way for like a lot of stuff. But if you look, if pure sort of engines of scale is sort of your advantage in the data world, you don't have—like from a VC perspective and from like a real long term sort of 10 year view—you don't have real advantage unless you're able to compound that somehow through distribution or methodology. 

Auren Hoffman  28:14  

ZoomInfo is now buying applications companies. So you see this move bit of some data companies also either creating or buying application companies. It's extremely rare to go the other way. I've almost never seen a SaaS or application company like create a data business. Like why is it so hard to go the other way?

Jonathan Levin  28:33  

I think the reason that it boils down to is that application companies are so nervous about sort of the trust that they have with their customers over the protection of information today that its super difficult to risk any of that and go the other way. Whereas if you've managed to amass some sort of data that lots of people need, then you can start to structure your application legal arrangements and data arrangements so that you can actually step into workflow. Like the other way around just becomes very tricky because you've sacrificed any of those data rights in order to achieve distribution in your application business in order to just like sell more. You're not thinking about it yet as what could we be aggregating? What could we be selling back as a data product?

Auren Hoffman  29:27  

Chainalysis has a very big business selling into government, law enforcement, national security, etc. A lot of “Silicon Valley based companies” don't often like to sell into law enforcement. To me, that's like kind of a competitive advantage for Chainalysis because you're maybe not going to have competitors from these like really big companies. How do you see that? How do you see that as part of your mission?

Jonathan Levin  29:54  

Yeah, I think that… Like the way I think about it as part of our mission is not actually about business defensibility at all. The way I think about it as part of our mission is that ultimately sort of I believe that more and more value is going to be stored and transferred over this technology, right. I see that as a total inevitability and just a matter of time.

Auren Hoffman  30:17  

Certainly that's been happening the last few years, right. We're seeing a massive increase. 

Jonathan Levin  30:22  

Yeah massive increase. So the thing that could slow it down is government's not understanding what the ultimate potential of this technology can be. 

Auren Hoffman  30:33  

So some bad regulator or something? Yeah. 

Jonathan Levin  30:37  

And really like you see the sort of in the national security space with ransomware and other things where there can be policymakers who, you know, don't understand that truly innovation is the biggest form of national security that we could have. So like we think we have to think about the way that we set up a regulatory regime for this industry, that both protects public safety, but also doesn't compromise on how fast paced and how much innovation can come from this and how much economic value, but also political power will come from a country like the US or any country frankly being able to foster innovation on this technology. So when I think about our government business I think about, we need to be in many markets around the world making sure that governments are able to really mitigate the risks associated with new technology. Because there are always risks. But also open up the opportunity for people to continue to innovate and build applications that everyone will use.

Auren Hoffman  31:49  

Now, selling into government is a different selling motion than even selling to a big bank or big financial services. Like what's kind of the same and what's different?

Jonathan Levin  31:59  

So I think that it is enterprise sales. You've got to have a robust sort of enterprise sales methodology. 

Auren Hoffman  32:07  

But that would be the same of selling into a big bank like JP Morgan or something?

Jonathan Levin  32:11  

Exactly the same, right? Like you’ve got to figure out their budget is. You've got to figure out why it's now. You’ve got to figure out the paper process. You really need to make sure that you've got a consistent methodology across your sales team to be able to forecast. The thing about government that's very different is, well, you've got budget cycles in different countries are going to make your business very seasonal. So you've got to be able to sustain, and you've got to be able to understand that that's really just how things work. And so the nice thing for us at the moment is we're selling in I think about 65 different markets around the world. So we're keeping track of those different budget cycles, and luckily they're not all on the same fiscal year. 

Auren Hoffman 32:56

They don’t all start on October 1st or something. 

Jonathan Levin 32:56

Yeah, yeah, yeah. They don’t all do that. So that's sort of helpful. But the point is that it's going to be a seasonal business. It's definitely, for the most part I would say at least for us, it's a fairly lumpy business. We tend to move in through like one unit or one division that sees the future in some way, but now are building up to sort of broader enterprise capabilities. I think that to that extent the ability to do land and expand is actually very similar in the government as it is in sort of financial services. In fact, oftentimes it's even easier because these things are more connected, and it is very hierarchical. It's pretty structured.

Auren Hoffman  33:50  

Now, the stereotype of law enforcement is that they're not as data savvy. Is that true? How has that changed over like the last five years?

Jonathan Levin  34:00  

Yeah, I think like there's clearly a massive push on data strategy more broadly within government agencies around the world. I have seen it firsthand sort of where we started, and people are obsessed with a workflow product that sits on someone's desk that someone…It’s like…

Auren Hoffman  34:20  

Perceived or, yeah. 

Jonathan Levin  34:23  

Yeah, yeah. Should we buy a car, or should we buy like a tool that sits on someone's desk? Like, yeah. They’ve got there to be able to think about bringing in software like that. I think the tricky part, and I think that the government has gone through at least one, maybe two cycles internationally of what I would call like data integration. It’s like they've got a lot of data that they need to make sense of as an organization. If you think about tax agencies around the world, they collect a lot of information. They need to make sense of it. So there's been a lot of, maybe even a decade or so at this point, like people really trying to make sense of that. I think that the latest trend, the one that you're pointing out is, they're starting to think about, like buying commercial data to think about data outside of their walls, to bring in and enrich sort of some of the things that they've already collected. That's like a whole new frontier for a lot of government agencies because typically, and this applies across the board, they've thought about, “Well, we have a lot of this information ourselves. Why do we need to go and ask a private company about it?” The more and more this is across different communities, there's a realization that the same like ZoomInfo. We need that type of information is a business to run. And so the government's also waking up to the idea that there's high value in being able to buy data to enrich even their own sort of government applications as well.

Auren Hoffman  36:07  

At SafeGraph we sell into governments, what state or local or federal. I have found personally that the data capabilities of the people who are working in these things has grown dramatically in the last couple of years. The people themselves have leveled up. That they've gotten much more data savvy. That they realize the importance. I don't know if you're seeing something similar.

Jonathan Levin  36:33  

Yeah, definitely. And I think it's also like top down as well. 

Auren Hoffman  36:37  

Even the decision makers are more data oriented today. 

Jonathan Levin  36:41  

Totally. So like if I think about sort of… We think a lot about trying to get our customers to give us what are the required capabilities that they're looking at when they come to cryptocurrency investigations writ large, right. Like I'm not interested in someone saying I’ll have another seat, thank you.  I really want to understand sort of what are all of the required capabilities that you need? More and more it's I need not just an application, but I actually need a data feed. I need an API endpoint that I can hear. I need something that we're going to be able to play with and integrate into some sort of government product. So we really have seen that transition, and I think it's very top down. So I think that it starts with, literally, I've seen sort of heads of agencies actually now get that there needs to be a data first strategy. That’s something that used to be something over in the technology division or over in like a particular shop. You've also seen the emergence of titles of people. I mean, the US Treasury Department has a chief data officer. Yeah, right. And so that really makes sense in a world where this is at the core of them being able to achieve their mission.

Auren Hoffman  38:00  

Now, a couple of questions about just starting companies. So you're co-founder of Chainalysis, and your other co-founder’s the CEO. Like, how do co-founders work best together to build an enduring long term partnership?

Jonathan Levin  38:13  

Yeah, great question. I think that one of the things about being…They say that, and I know you know this firsthand, but like being a CEO is very lonely. The way that I think about like a non-CEO founder is you get put in sort of a lot of utility positions that really served the need for founders to have good relationships. You need to sort of leave a lot of things on the side and particularly check your ego and be able to step in to sort of some of those roles and play that utility player and run sales for nine months like I've done this year. Done some of those tasks to be able to help the company just scale and grow as you build out the executive team and operational functions. I think the other thing that I'll say is that it's important for founders to realize how much they need each other. It’s the mutual need for like thought partnership and really the sort of, as you scale a company and you bring in those executives that have seen the scale of the company before and know how to build the operational rigor to get you to the next level. Again, the thing that you lose the most is the ability to innovate. As founders, you have an opportunity to not sort of distance yourself from the rest of the team, but really think about together how do you continue to innovate as a company? Really like there will be something unique that founders will share in a language that no one else in the company will ever understand.

Auren Hoffman  40:00  

Plus you have the credibility to implement it or something, right,

Jonathan Levin  40:03  

Plus you have the credibility to implement it. So that's something where, I think that over time, just being sort of vocal about that to each other and realizing like what those challenges as an organization will be brings you closer together as founders. Myself and Michael have definitely over the last seven years sort of your relationship changes a lot. But as you scale, we just crossed 400 people at Chainalysis. We are starting to think about well, how do you continue to innovate at the pace that you want? And that's a conversation that Michael and I have on a very frequent basis as well as with executives. But I find particularly as its founder that you can try and solve some of those problems.

Auren Hoffman  40:48  

There are companies like HubSpot, Google, Dropbox, Airbnb that have these long term enduring partnerships between cofounders. There are other companies where the non-CEO cofounders seem to leave once the company gets to a certain level of success. Like, have you learned anything from studying these other companies?

Jonathan Levin  41:09  

Yeah, so I actually--and I haven't done this yet. So I'm going to announce it here. But if anyone does want to reach out to me as a non-CEO founder, I'm actually going to start a group.

Auren Hoffman  41:19  

We get like you a Dharmesh Shah from HubSpot. Yeah. Okay. 

Jonathan Levin  41:23  

Yeah. So I'm going to start a group on it. Because I think that I think it is a unique experience that isn't talked about enough. So I've definitely had some mentorship from founders in that position before. Oftentimes, it is very challenging, but there's lots of CEO groups and help for the lonely person at the top. Like, the person who's sort of helping that person stay above the water and paddling furiously below doesn't oftentimes get the same level of attention. So I want to I want to make sure that we do that.

Auren Hoffman  42:03  

Okay. Yeah, that makes sense. Alright, last question we ask all of our guests. What is the conventional wisdom or advice that is generally bad advice?

Jonathan Levin  42:11  

I think this applies mainly in crypto right now, but I think one of the things that is happening with capital raising in general—and particularly like VCs talk about this—raise more than you have to. I think that one of the things that I've seen, particularly in very fast paced markets, is that there's such a momentum building activity when it comes to fundraising where you don't actually… You can value your equity more, create some scarcity around it, that actually allows you to raise sequentially with less dilution. The actual act of these funding rounds today, and particularly in the capital rich environment is that's a much easier thing to accomplish than most founders understand. So I think that it’s kind of that like classic sort of sales technique that the VCs will, will try and play, but actually I think founders can be a little bit more conservative today and can sort of get lower dilution over time.

Auren Hoffman  43:25  

Okay, interesting. I love it. That's great. Well, thank you, Jonathan.  Really appreciate it. Where can people find out about you on the interwebs?

Jonathan Levin  43:34  

Well, if you want to learn more about Chainalysis, you can go to chainalysis.com. Being able to spell that as a little bit difficult.

Auren Hoffman  43:43  

We have a very smart audience. So I think they can do that, yeah. 

Jonathan Levin  43:47  

Yeah.

Auren Hoffman  43:49  

I think we have the highest IQ audience of any podcast. So yeah. 

Jonathan Levin  43:53  

Yeah, no. I can only imagine so. And you can find me on Twitter at @jonylevin on Twitter.

Auren Hoffman  44:02  

And I follow you on Twitter. You're a great tweeter. So yeah.

Jonathan Levin  44:06  

Now that the end is in sight for me as far as leaving the revenue organization on an interim basis, I'm gonna get back on Twitter. 

Auren Hoffman 44:17

All right. I love it. 

Johnathan Levin 44:19

So start tweeting more. 

Auren Hoffman  44:20  

Yeah, the hiatus has been too long. Yeah.

Jonathan Levin  44:22  

The hiatus has been too long. So that's really where to find me. And yeah, no, really appreciate the conversation. I think that the future of DaaS is extremely bright. I think what you're doing here, in terms of establishing, what are the ways that people should understand these businesses is critical for not just fundraising, but actually entrepreneurs building great companies.

Auren Hoffman  44:43  

Awesome. Well, thank you John for joining us a World of DaaS. 

Johnathan Levin 44:47

Thanks so much. 

Auren Hoffman 44:48

Thanks for listening. If you enjoyed the show, consider rating this podcast and leaving a review. For more World of DaaS. You can subscribe on Spotify or Apple podcasts or anywhere you get your podcasts. Also check out YouTube for videos. You can find me at Twitter at @auren, and we'd love to hear from you. 

World of DaaS is brought to you by SafeGraph.