Transcribed by https://otter.ai
Scott McGrew
Are their legacy firms on Sand Hill Road? They won’t exist some years down the road because they’re not doing what you’re doing hundred percent. I’m Scott McGrew. Welcome to Sand Hill Road. If there’s something we’ve learned from theranos, the blood testing company, it’s that if somebody says they’ve invented something amazing, you would better see the damn thing work for yourself
Chris Farmer
misperception that
Scott McGrew
Chris farmer of signal fire has invited me to see his super secret future of venture capital, a decision making computer system called beacon You know, you’re smiling. You’re well you’re not quite smiling. You’re you’ve got a strange poker. Yes, as he pauses to take a sip of water. Okay, so you’re doing that. Then we’re going to talk about what Chris is doing and how but I’ve agreed not to say too much about What I’ve seen, but I can say, I have seen beacon work, a database of more than 40 million people and companies, including lots of details about you if you work in high tech, but it’s more than a database beacon, Chris says, makes decisions and beacon learns. When I walked in. I challenged you to show me beacon because I didn’t know what it would look like I said, was it you know, this giant blinking machine at a Star Trek behind a locked door? No, it’s on your laptop. And as you were showing me, you got very enthusiastic.
Chris Farmer
Yes, it’s a you’d have to be crazy to spend as much time building systems as I have, so you better be passionate about it
Scott McGrew
became a single laptop. Let’s signal fire become a data driven venture capital firm, from start to finish. full stack is Chris would say
Chris Farmer
we’re the first venture firm built full stack as a technology company. And so that means every process of the firm, from deal sourcing from, you know, is built with systems and data, but all of the workflow collaboration tools, task management, etc, is building a closed loop.
Scott McGrew
Every deal you make every hire that you suggest goes through beacon.
Chris Farmer
Yes, so one way or another, it may not always be sourced by the beacon platform, it may come from a network of advisors or traditional referral, but we use it as a core workflow of everything that we do to understand how it ranks compared to competitors, what the richness of the revenue is, how sustainable that is, what the quality of the team is, and all sorts of dimensions like that.
Scott McGrew
Some of those things, you know, seem like they’d be subjective the quality of the team How do you put a data point on the quality of the team? It’s it’s relatively
Chris Farmer
complicated depends on the first thing you have to do is categorize them by the type of you know, person they are. Are they a data scientists? Are they a software, you know, iOS developer, are they a designer, etc. And the characteristics no different than like an athlete would the stats would totally changed depending on what they are. But we look at sort of their work experience, how fast they’ve been promoted, what types of organizations they work for how hard it was to give them those organizations, that sort of academic achievements, they may have had sort of online contributions of various kinds, and how
Scott McGrew
many people will be talking about in this database? Overall, it’s
Chris Farmer
about 40 billion people about 40
Scott McGrew
million people. Yes. How could you possibly know how quickly somebody got promoted? If we’re talking about 40 million different people,
Chris Farmer
clearly, you can’t do it manually. So you have to build sort of like a mini Google in order to do that, which is large scale search processes and tons of data pipelines in order to canonicalize, disambiguate, normalize, etc, all the data that comes in unstructured and then but there’s tons of stuff on people’s websites. There’s tons of things and regulatory filings, there’s tons of things, some
Scott McGrew
of your some of your scraping some of it. You’re paying for
Chris Farmer
Around people, it’s all stuff that we collect in the public domain.
Scott McGrew
Okay. And then when it comes to the companies, you might you might be paying for some there are
Chris Farmer
some data sets that are not available in the public domain, you
Scott McGrew
think you spend about $10 million a year on in all
Chris Farmer
in systems and team and everything else, we we spend something in that ballpark, if not more,
Scott McGrew
it strikes me beacon is a lot like slack. slack started as a video game company that built its own internal communication software, only to discover it was the communication software and not the game. That was the real treasure. slack is now a $12 billion company, your venture capital firm, but it seems to me the most valuable thing in this firm is your laptop.
Chris Farmer
Yeah, so. So there’s a couple of interesting points to that. So one interesting point in that slack analogy, as it it’s not about a point in time. It’s about continuous monitoring. The the nature of the traditional venture approach is you meet with the company and you make a decision based on the data points. At that time, which if you were looking at, let’s say, the gaming company before it became Slack, you would have totally missed the opportunity. So first off, the system has to be built for continuous monitoring. You know, as far as the value the nature of competitive intelligence is, the more people that have it, the less it’s worth. And so you can either make a decision to be an arms dealer to all the other VCs, and then it would degrade in value, the more people have it, or you could use it as a competitive tool to build hopefully the next generation leading edge VC, which is what we chose to do. So that’s, that’s why we keep it internally.
Scott McGrew
Are there firms on Sand Hill Road or elsewhere that have something even vaguely similar?
Chris Farmer
There are some firms that are starting to take some more serious efforts, lots of people put up smoke and mirrors about the things that they’re doing here. You know, Google Ventures is probably one of the most advanced in some of these things. And then co to which is a Bitcoin hedge fund that’s entered the market is has a very serious data science and effort. As far as I’m aware, no venture firm goes Fully and the end the way that we do with all the workflow in the closed loop, that very definition of machine learning requires feedback loops for the machine to learn. And so if you’re taking a data science approach and using data, but then the loop breaks, because the whole firm is not built end to end like this, then you lose, I think, the vast majority of the value. And as far as I know, we’re unique in that sense. Yeah, the closed
Scott McGrew
loop, I was going to interrupt you. And then I realized you’re going exactly there that, that even if I build something that has the same number of data points, as you have, you’ve then leapfrogged ahead of me, just while I’m building it, because you are further reinforcing the data that’s inside yours, because it’s continuously moving through that closed loop. That’s certainly the goal.
Chris Farmer
So I mean, if you think about this from Google, you know, they created an algorithm early on, which has now been published and lots of people have folded what’s kept Google competitive, obviously, they haven’t stood still, but is also the massive amount of feedback loops they get from all the search traffic, because of the dominant market share. As a result, when Microsoft wanted to enter the market, they had to cut deals with Facebook and Yahoo and others in order to drive enough traffic in order to tune their own algorithm. And so, you know, I think it’s it’s a very important part of a system like this is to really build it out and Dad,
Scott McGrew
you invested in zoom pizza, and I would assume you put your investment questions through beacon when you did it. I understand that you don’t want to give away too much about beacon. But what what does beacon know about pizza?
Chris Farmer
Actually a remarkably large amount, I believe, at the time when we met was Alex Gordon was the sole founder with the PowerPoint. There was nothing else there was literally no data on the company. But that doesn’t mean that we couldn’t use the system to get context. And so we process trillion dollars, trillions of dollars of us spend, and we could use it to understand the dynamics and to verify a lot of his assumptions on what has happens at all the other incumbent competitors in the market, whether that be Domino’s, or Pizza Hut, or local, you know, companies, we also had tremendous amount of data and all the food delivery companies and could clearly see that that was a trend. We were actually quite bearish. On a lot of the space there was clearly a consumer phenomenon, but we were worried about the margin side of it. But they had some very interesting approaches that we thought and it was such a high margin product line, that was already the largest delivery category. You took that amazing founder with really deep insights into the structure of the industry that we could validate with the data. And that’s what led to the initial investment.
Scott McGrew
You funded. Green Park Chad Hurley’s startup is there. Can you talk about what that is?
Chris Farmer
I don’t know how much is out there publicly. So I’m going to I’ll keep this relatively narrow. I mean, obviously, it was a group of exceptional entrepreneurs that had founded you know, companies like Zappos and YouTube and they were very passionate about the sports world because You know, their, their, their sports fanatics, they’re actually part owners of the warriors. And they just saw there was a huge opportunity to create more community around those things and that they’re, you know, some of the gambling that was going on was was pretty damaging to a lot of people’s lives. And then a lot of the fantasy sports, sort of fragmented loyalties and whatnot. And so they had some very specific ideas of how to create a platform that that really enhanced the the sports experience for the community. And that’s the platform that they’re building and begin liked it. Yes, there there was, there was a lot to like in this particular company.
Scott McGrew
Have you ever had somebody come in and pitch and you’ve run their numbers through your systems and discovered that they’re full of it? You know, that that’s absolutely not what’s happening in the world?
Chris Farmer
Yes, many times. In fact, I’d take it a step further. We’ve definitely found a number of situations where they were Just to say, doing unsustainable things, in order to juice their numbers, and going into a financing, one of these companies actually became a public company that eventually crashed. But you know, it’s that level of multifaceted visibility into what’s going on. That’s critical. But all the time you you’ll meet with a company, and you’ll find out that one of the co founders blew out two months ago. And they’re, you know, they don’t mention that in the initial pitch. And so, you know, having that context so that you can, it may be nothing wrong, but at least you can put that on the table and be very transparent with the founder to really understand what’s going on on the company. Of course, every company thinks they have the best team in the world. Let’s just say that, statistically, that’s not possible or true. And therefore, this helps to really sort of cut through the noise and is is is as important for helping us to save time by not even taking the meeting and, and and using the entrepreneurs time as it is to knowing where to proactively reach out to an entrepreneur that’s doing something interesting.
Scott McGrew
In my notes, I have, you know, asked him, couldn’t LinkedIn Do the same thing that he’s doing. I’ve seen beacon with my own eyes, it is lightyears ahead of what LinkedIn is doing. Could they do it? Possibly. But it’s your so you’re way far ahead of that.
Chris Farmer
Yeah, I think that what LinkedIn is doing is is very powerful. I think they’re a dominant player. And it’s very different than what we’re doing. I mean, they’re, what they’re building out is a network, it’s a social network, what we’re doing is way more narrow. And in the narrow domain in which we compete. You know, we’re using it for an investment process, not for right,
Scott McGrew
but you have a database, it was at 40 million said 40 million people. You have a database of 40 million people, you know, where they went to school, you know, how many patents they’ve written, you know, how many scientific articles that they’ve, they’ve written, etc. Lincoln has some of that same data. I mean, it’s voluntary, you know, add it in at LinkedIn. But but it is nowhere near what you’re doing.
Chris Farmer
Yeah, I mean, I think the big differences is that we’re not coming Strain to any one source, we’re using 10s, you know, millions of sources in the longtail that were crawling and those types of things. Now some of those may be individual apps or web pages or those types of things. But there’s a tremendous amount of data out there and in various filings with the government, company web pages on people’s, you know, Twitter accounts, or fundraising websites or all sorts of different things. And, you know, no one site is complete. And it’s the aggregation and integration of all those different things, all the values in the join as they write, which is
Scott McGrew
going to be tremendously difficult because you’re talking about pulling data from an s one, and also some guys Twitter accounts. Yeah.
Chris Farmer
I would say that the data munging the data, janitorial work and preparing unstructured data to be structured is the 90% of the work at least
Scott McGrew
is their data out there that you can’t get or it’s too difficult to get that it’s not worth it that you wish you could.
Chris Farmer
There are many, many things that are out there that we wish we could be We have to be very cognizant of privacy rules. And
Scott McGrew
yeah, I’m not saying anything you’re doing is is not private, although I bet you know, if you could hook up with Planet Labs and somehow Just follow me around, that would probably be very valuable to your data set. But but there’s got to be stuff out there that is just difficult to integrate into what to what it is you’re doing.
Chris Farmer
Yeah. I mean, a lot of you know, a lot of companies have like interesting windows into data sets, like for example, stripe, you know, which is an amazing company and powers or AWS, which powers computing or stripe in the case of powers payments, but because their customer is the enterprise, that data is not out there, they have to protect that data. So I’m sure I’d love to have the pipeline of how fast every stripe customer AWS customer is growing and how much compute they’re using them. What kind, but unfortunately, that’s not in the public domain.
Scott McGrew
Where did this you’ve been working on beacon for 10 years and you said it’s been active for about five years learning as it goes, where did this interest in data come from?
Chris Farmer
So I used to cover Consumer wireless, going back, you know, 15 years and when the App Store came out, I was trying to monitor apps that were moving up in the App Store. And it worked when it was, you know, a few hundred apps. And then pretty quickly, it was thousands and then millions. And it was a joke that I was going to be able to monitor the biggest up and coming apps. And so I actually, we built some scrapers and things like that in order to sort of create a leaderboard, in essence of the App Store in the very early days of that, to try and get a leg up on finding those opportunities before other people did. So I started working on this in around 2007. Very quickly, it got so noisy that it was so one dimensional, with just the app because you know, the different business models, you would it would skew very heavily like a social networking app that was ad supported because the number of downloads was a huge number, but you’d completely miss like an Airbnb. Yeah, because it took far fewer apps downloads because the transaction size is very hard. So depending on the business model, like very much dictated sort of where you’d put thresholds of you know, Something crossing a line of being interesting. And so you had to add all sorts of other dimensionality in order to make sense of even the App Store itself. And that’s when it was became very clear, this is a much more complex problem, and that you really had to go all in if you were going to solve it.
Scott McGrew
Could you do the opposite with beacon instead of running the potential of a deal in front of you? And see if it made sense? Could you in some way have beacon find the deal?
Chris Farmer
We do that in very large scale? So we’re, you know, a big part is is the anomaly detection systems that that something is outperforming on some material dimension of its type of company, which is very different for like 2000 different types of companies because the database is going to look radically different than a social network, which is gonna be completely different than a fashion marketplace. And so, yes, we, we monitor something like 15 million companies in real time and when we see it something exceptional explosive growth, etc. Team quality happening at one of those companies it lets us know you’d be silly to tell me but I’d be silly not to ask. Right. Okay, can you give me an example? We’ve had a number of about a third of the companies that we invest in are sourced entirely by the system itself.
Scott McGrew
Okay, I’m going to break right in here and make sure you heard what Chris said, a third of signifiers, investments were discovered by a computer that would change venture capital
Chris Farmer
forever. And we’re improving constantly, month on month. So it’s a it’s a rapidly learning system.
Scott McGrew
I’m going to repeat that back one third of the deals, baited signifier. Were in some way initiated by a learning machine. That’s great. That’s astonishing.
Chris Farmer
The best part is it’s only getting better. So I mean, I think I think what ultimately differentiates Is everyone here has eyes and ears everywhere in the categories that they’re doing, not just what they could humanly monitor?
Scott McGrew
Is it possible that someday I’ll come to signal fire? And it’ll just be a laptop sitting on a coffee table?
Chris Farmer
No, no, I am not a believer in that we’re decidedly we believe that fundamentally that humans plus technology beat humans or technology, and that there’s lots of qualitative aspects. I mean, founders want to know that you’ve got a partner there that it’s not based on some data set, that you’re going to be with them for the lifecycle of the company through the highs and lows. All of that is true. And you can tell the grit and determination of a founder from from the data, you can’t see their vision. You can’t anticipate future products that haven’t launched yet, that type of thing, but it does give tremendous context and it does let us know where we should focus our time no differently than the analogy of an MRI using machine learning or something like that, that that detects potential anomalies that a human radiologist or whatever might look at, in order to make sure that they don’t miss something. You know, just from being tired or whatever it might be,
Scott McGrew
if you ever had a team member Come on, and you’ve trained them on beacon, and they’ve been resistant, you know, know, I’ve been doing this many years I have a gut. I have a gut for this, Chris, I know where when a deal is a good one. And you say, Well, no. beaconed says it’s not.
Chris Farmer
Yeah, so it’s not exactly how it works. But if effectively, anyone who joins here better be all in this approach. I mean, I think people pre select to work here, just by choices they’ve made in their career, it’s actually something we look for. They, they don’t have to just be great venture capitalists or technologists, they also have to be sort of predisposed to believing that the future, you know, leverages data leverages networks, you know, and that and for them to want to sort of bet their career on that. And so, you know, it’s obviously something we filter for, but also something that I think any candidate that be talking to us would have to inherently believe that that this is the future for them to to skip the the sort of maybe more more obvious payday of working for, you know, an incumbent firm and really betting on this sort of next generation.
Scott McGrew
But of course, choosing the right investment is just half of a venture capitalist job. The other half is supporting the entrepreneur, guiding a small company as it grows, the startup doesn’t care how or why you funded them.
Chris Farmer
They don’t care if you used a carrier pigeon, or you ran into them in Starbucks, or you use some fancy algorithm, they care about how you’re gonna help them build a company, right? And are they going to be better off working with you? Or are they going to be better off working with someone else at the end of the day, every venture from differentiates on their value add, and that’s what differential commodity capital and so, you know, we ourselves use systems to do that as well. And so, we run a microservices architecture sort of modeled on how the technology industry and we break all these different services industry components, we measure all of them meticulously. And so we run service much more how, you know, the Ritz Carlton would run service measuring everything and making sure that we have the highest service quality highest SLA of any venture firm, we run on okrs just like the tech companies, and we run ultimately services much more like like a Stanford hospital would run services. So rather than the GP trying to do everything, you’ve got a product question, let me whiteboard, you’ve got a pricing question. Sure, let’s pull out spreadsheet. Instead of trying to do it as a Swiss Army knife. We have a whole network of advisors that are world class and almost every some function you can imagine. And so I get the the opportunity to to refer our founders to bottom as a podiatrist and eco cardiologist to use that medical analogy. But in our case, someone around pricing or an engineering leader or those types of things, they’re going to give a plus advice in that subdomain, rather than sort of the the Swiss Army Knife approach that has been typical the venture industry for so many years as a result of that. This has resonated unbelievably well with our founders. And they’ve given us tremendous feedback on how they want us to tune the systems, what services to offer, how to scale that and we use a systems and data approach. Again, even building these Sort of full SAS platforms for our founders and for the network of advisors to help make them better at what they do so that the whole community wins together?
Scott McGrew
Is the network of advisors paid per deal, or they have they have skin in the game in the company and the fund itself.
Chris Farmer
So the network of advisors goes even beyond that, you know, like any modern tech company, that sort of deficit spends for a substantial period of time management fee doesn’t did not cover and still does not fully cover what we do. And so as a result of that, we raised money from a network of the most connected most well thought of angels and advisors in the valley after having interviewed over 500 founders on who was their favorite advisor, who would they first go to for capital. And so they own about a third of the firm itself. And collectively, they’re a very large investor in the funds. And so in addition to that, they get access to systems they get co investment opportunities and all sorts of things. There’s no amount of money we can pay these people because they’ve already invested successful and in many cases extremely wealthy. But we find ways to provide more leverage to them than they even provide to us. And when we do that, they want to keep showing up and supporting our portfolio because they themselves win at the same time.
Scott McGrew
When you think of all the industries in which there is a, you know, a young data driven industry that that supplants a old, an old system, I could see this happening and venture capital are their legacy firms on Sand Hill Road that won’t exist some years down the road, because they’re not doing what you’re doing hundred percent. I mean,
Chris Farmer
you’re already seen the implosion of a number of sort of very well regarded firms. There. There’s huge generational risk between things and the next generation of great you know, venture capitalists want to go to the platform where they’re going to be most successful, where they’re gonna have the best competitive advantage and get to work with the best, most inspiring, most world changing entrepreneurs. And if you are not that platform, here at risk, Being able to recruit the next generation of talent, which inherently puts the entire firm at risk. And so, you know, I do believe at the end of the day, we are that we our goal is to build that platform where people will be more successful on our systems than they would on any other. So imagine you’re an associate joining a new firm, you join with what an internet browser and a phone at most places, in our case, you start in a semi autonomous driving car, it’s already feeding you leads, it’s already, you know, helping wire you into those entrepreneurs with a totally differentiated story of how you can impact and support those entrepreneurs with a laundry list of benefits to the entrepreneur and references that support those claims. And so I mean, I think as somebody entering the industry, why would you want to start you know, with with no competitive advantage if you have a chance to sort of be on a platform that gives you every advantage and so any any other inherent skill or talent you have yourself is just amplified. Our goal is to turn strengths into superpowers through the platform.
Scott McGrew
Does it frustrate you that I’m not going to use names but that there are some names in venture capital that everyone knows even if they don’t go think about venture capital? And signifiers, not necessarily one that rises to the top.
Chris Farmer
I would say that my my personal bias and that of a number of people on the team is not to be overly promotional in that sense. I mean, certainly we sell one on one to an entrepreneur. But we were intent intentionally in stealth mode for years and years. We didn’t have a website until our second thought. And even then, the website beforehand was a splash page with satellites on it that had nothing to do with anything that we were doing by design. You know, I’d say at this point, our entrepreneurs are asking us to be a little more public because it’s important for them for recruiting purposes and for credibility in the market. And so we’re somewhat reluctantly, telling a little bit more about what we’re doing and hence, you know, interviews like this, but At the end of the day, you know, we’re really sort of heads down focused on building the best platform and supporting our founders as best we can. Their references are our lifeblood, because we don’t rely on brand and marketing and being super loud in the industry. We rely on the referral network of founder to founder advisor to advisor, you know, spreading the gospel of what we’re doing and why, you know, it’s so effective for an entrepreneur as a part of
Scott McGrew
the closed loop. That’s right.
Unknown Speaker
no different.
Scott McGrew
Chris farmer with Signalfire. Next week on Sand Hill Road Tess Hatch, with Bessemer Venture Partners, investing in space. Sand Hill Road is produced by Sean Meyers executive produced by Sarah Bueno, and Stephanie androoni.
For more interviews with Silicon Valley’s most influential entrepreneurs, Check me out on TV at press here. That’s Sunday mornings on NBC Bay Area and everywhere in the world on iTunes. At it press here TV. com
No Comments, Be The First!