Building a next-generation tech-enabled Venture Capital firm with Chris Farmer of SignalFire

0:00Welcome back to another episode of Venture Unlocked, the podcast that takes you behind the scenes of the business of venture capital. I'm your host, Samir Khaji, and on this week's show, we're excited to be joined by Chris Farmer, founder of SignalFire, a firm that was founded nearly a decade ago with the goal of disrupting the way venture capital is done. Unlike traditional VC firms, the firm has been built more like a company than a traditional asset manager, as it leverages a robust purpose-built data platform to augment its unique service model to entrepreneurs. Before starting SignalFire, Chris worked as a VP at Bessemer before moving on to focus on seed investing at General Catalyst, where he was involved with investments such as Coinbase, Stripe, and Venmo.

0:42Today, SignalFire manages nearly$2 billion in assets and has invested in companies such as Aura, Grammarly, and Rowe. During our conversation, we spent a lot of the time talking about his views on building a next-generation venture capital fund and the organizational design that goes behind it. I think you'll really enjoy hearing Chris's unique perspective on VC. Now let's get right into the show. Samir Kaji is the CEO and co-founder of Allocate. Allocate and Venture Unlocked are independent of each other. Any statements or references made by Samir or his guests regarding third parties, investments, or securities are solely their views and opinions and are not intended as investment advice or an endorsement of such parties or securities by Samir, his guests, or Allocate.

1:26Allocate or its clients may maintain relationships with or investment positions in guests, third parties, or securities mentioned in this podcast. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Hey, Chris, it's great to see you. Thanks for coming on. Thanks, Samir. I appreciate you having me. So I want to start in a slightly different place. I mean, you've run SignalFire now, I guess it's been 10 years. But I was looking at your website, something stood out to me. It said, we began our journey in 2013 as a bootstrap startup with a goal of disrupting the way venture capital is done.

2:01Now, there's a lot to unpack there. But ultimately, I want to know, what does that actually mean? What was that inspiration to start the firm, coming from more traditional firms like General Candles and Pessemer? The genesis of me applying data is I had run companies that had a data element to them as an operator early on in my career. And so when I got to Bessemer Venture Partners in the mid-2000s, I was covering wireless when the App Store came out. It seemed crazy to me to try to monitor the App Store completely visually. And this seemed like an awesome opportunity to apply data and algorithms to create a leaderboard, that type of thing.

2:46So that was sort of the initial wedge point, which is obviously a really narrow sort of scope. And then pretty quickly, it was clear that you'd end up with all social and ad supported gaming type applications, and you'd miss the transactional businesses or the SaaS businesses, etc. So it was very clear pretty quickly, you need to cut it with a lot more data sets in order to sort of filter through the noise, and that that needed to be approached sort of from scratch, as opposed to trying to append that into an existing venture fund. I started with a large research study of best practices and tried to take a playbook from every permutation of investing and operating.

3:30We interviewed about 170 different funds and fund types, everything from accelerators to incubators to venture funds to quant hedge funds to corporate venture groups, etc. and looked at how do you build systematic advantage through data, workflow, software, etc. No different than an operating company would benefit their sales team. And we actually took a lot more pages out of the corporate playbook than we did out of the venture playbook in the initial ideation. And I refined a bunch of these things with Jump Catalyst in the early 2010s before spinning out in 2013 with enough proof points that I was willing to sort of move from experiment into full commit mode.

4:13And the rest is history. What you're describing very much feels like a tech company, not a traditional venture. And we'll get into what it means to modernize venture and how you do it. You and I have talked about this. You always tell me I'm running a company that happens to invest. And that is the big unlock that I saw back in 2013. But what does that actually mean, that you're a tech company that happens to also invest in other tech companies? If you think of the venture industry in general, certainly from that era, you know, you basically hand someone a cell phone and an internet browser, email address and say, go hunt.

4:53I mean, there's no other industry that would ever run a sales process that way. And there's also no industry that would take basically the CEO equivalent and have them prospect, win deals, and then do a decade of customer success. And the odds that they're best in class at all of those things from sourcing, investment judgment, and being the best possible board member builder for the company are really low. And so you end up having to make a ton of compromises. And so we looked at it from a systems approach where, you know, if we have deal team members that are out hunting and trying to find the right entrepreneurs, why don't you give them qualified leads as a much more efficient way to focus their time rather than set up shop in this local Starbucks and say, I'm open for office hours, which just is insane to me.

5:43And we wanted to build a firm with structural advantage. My vision was to build the absolute best platform for the next generation of great investment talent to basically have the most leverage and maximize their odds of success. And if you could do that, you'd attract the best talent because why would you start somewhere else that just had brand, historical network, etc.? If you could build a true data advantage in a systematic way, you would attract the best talent. And then if you could attract the best talent, you'd ultimately build something institutional and much more scalable and much more persistent in producing outsized returns.

6:19That was sort of the philosophy going in. And so we broke the venture process into five core functions, sourcing, picking, winning, portfolio support, and portfolio construction, and then systematically applied first principles of how would you apply data workflow collaboration networks to each of those functions to do them sustainably in a better way than the venture industry had done historically. Let's go a little bit deeper. Obviously, you've mentioned all the things that do matter. It is sourcing, which used to be like, what network did you have? Who do you know? Picking at the early stages, largely founder-driven.

7:01Is the founder really interesting? Have they done it before? Are they really tackling a big problem and a big TAM? It's the winning, which is like, what is your value advantage? maybe you can break down each one of those pillars and how you've used data to really drive these more consistent outcomes in each one of those yeah so on the picking side of it it's what most people assume we're doing we're doing a lot more underneath the scenes to understand these things but on the sourcing side of it is we're tracking probably more data than anyone in the world on the startup ecosystem. So it's everything from consumer spend and behavior to app stores, to talent movements and clustering around companies' capital flows from smart money, chasing deals, et cetera, et cetera, as the same sort of lenses you would use as a human.

7:56You're just doing it at massive scale for 80 million entities, whether it's an open source project or a quote-unquote founder in stealth mode that hasn't even incorporated, right? You're tracking all of those even pre-formation. And we try to do things that humans do well. We're decidedly human in the loop. We're quantamental effectively. You know, humans are... Data is not going to be great at understanding the vision of a founder or the magnetism of a founder in the early days, etc. But humans are not great at tracking 80 million entities and monitoring them consistently versus episodically.

8:29And so we try to use the best of each technique and blend them together into something that is scalable, sustainable, institutionalized, and very systematic that brings a persistent edge to what we do. And so on the sourcing side, it's, you know, taking the lenses of these are my tastes, right? These are the things that I think are the characteristics of something that are important for an e-commerce company, you know, which is probably going to be more around frequency of transactions, repeat rates, you know, upsell, like all these types of things. Whereas, I don't know, the AI team may not even exist, let alone be important.

9:09Whereas if you're building an infrastructure company, it's going to start with team, team, team, right? And that's going to be, and maybe some smart money that rolls around it or an open source project or wherever it may be. The characteristics are very different for each type of company. And then we have put the team members into swim lanes and they're getting firehose of leads, alerts that are tracking from these millions of companies. So it's like a human loop cybersecurity system in a lot of ways. It's scanning all the startup activity out there and then highlighting, looking for outliers, clustering of exceptional talent, exceptional traction, or reviews or whatever it is that the community is, you know, that's out there in the public domain.

9:53And then we're giving highly qualified, basically SDR type leads to those deal team members to then go pursue the companies in their swim lanes. And what I love about it is it gets rid of the sort of paranoia of, oh my God, I've got to be at every event and do all this activity because I could run into the next Zuckerberg or Elon Musk at some party. And if I'm not there, I'm going to miss it. And now it allows you to sort of pull back, say, what are the characteristics? Let me scan this at scale. And it's as much about what meetings I don't take as what meetings I take. Because everyone is putting their best foot forward.

10:32And if you're building an AI company and have no AI talent on your team, that's probably not something that's going to get prioritized. and so on and so forth. If you have terrible retention rates with your customers, that's probably not something that's going to prioritize. But then you've got to bring the human in and understand the founder and the vision and where they're going because all you're seeing is a sampling of recent history of the company. And then you bring that human venture judgment overlaid on top of the data. But you also then, in the investment judgment piece, have the benefit of being able to see the peers and benchmark.

11:06work, right? Whereas, you know, if you're looking at company A, their competitor, company B, is not going to show you. They're just going to like, you know, open their books and show you all of their data because you're interested in investing in their competitor, right? Or if you're already invested in the competitor and then a subsequent follow-on round, they're certainly going to avoid you like the plague because they don't want to share any intelligence that you might use against them. And so having apples to apples comparisons that's independently generated from the companies gives you a sort of persistent advantage in just context and benchmarking for those investment decisions.

11:43And then ultimately winning is driven by, you know, founder MPS, et cetera. So we've built scalable systems. If you think of Y Cominers design like Stanford, you apply through an application, then run it through an algorithm. They then pull out GPA and test score equivalent characteristics of founders, put a human loop and interview them, buy a very cheap option, put them through a one-to-many support program, cherry-pick their best companies, bring in an optimized sort of fundraising environment, and then let, you know, hundreds of companies, you know, go at that point to sort of, you know, in a Darwinian-type fashion, all right, with the venture industry supporting the strongest ones over long periods of time.

12:26You know, they're designed like a university to be one-to-many. In many ways, we're designed, And if they're Stanford, we're Stanford Hospital, right? We're part teaching institution, part care institution. So we do hundreds of events a year for our portfolio companies, as much as the grad school version. We have a sales mastery series, a market mastery series, a talent mastery series, a people mastery series, which is more HR function and recruiting functions, et cetera. And then we have founder development programs and all sorts of things for scale. And then we have a network of hundred plus advisors that are specialists.

12:58So as a GP, I'm a general practitioner. I'm not, you know, I'm not the Swiss army knife to use to play that hospital analogy. But then I refer to a phlebotomist, podiatrist, psychocardiologist. We just have it all in network. Right. So you as a patient, as a founder, right, when something comes up, either preventative or acute care, you go to Stanford and know that they're going to have great care teams and all the different sub, you know, specializations that you might need. And we've done the same thing at a venture firm level. So it's a very different architecture than your typical venture firm.

13:30I'm a CEO, we have a CTO, we have a chief people officer, we have a CMO, all the different types of functions you'd expect in an operating company. And we use the same sort of technologies that a founder-facing developer tools or infrastructure company might use on everything from sourcing to support to onboarding to customer success, except for we're doing that as a venture fund. And the benefit, it turns out, we as VCs think we know everything, and the reality is we don't. We're good surface level at a lot of things, but we're not excellent typically at many things. And so this hybrid generalist specialist model actually produces much better results for the founders.

14:12Our MPS has been between something like 88 to 96 over the last seven, eight years. And over 85 % of our founders would say we're the most valuable investor on their cap table. And that's not because I'm the smartest guy in the room by any stretch of my imagination. It's because we have experts at the right time that we can introduce them to, to get the best possible judgment in that sort of acute problem. And that they have continuity of care from the deal team members and general partners, but they also get the specialist care. And we found that that's much more scalable. It's much more systematic.

14:47And frankly, you can't get orphaned as a company that happens. a lot of venture firms. If somebody transitions out, you have multiple touch points. And you're not just dealing with fire drills when something happens. You're able to much more systematically support companies preventatively before something happens versus like, here's something that aborted me and then overreact. It's just we look at it fundamentally differently in the way that we support portfolio companies. And that ultimately leads, that NPS leads to a closing playbook and reference ability because we don't have decades of brand to sort of rely on that really helps us close at a very high level over 90 % of our term sheets we win because it's just clearly night and day in the level of support and quality of support in a super tangible way.

15:32Then we use software as well to make these things scalable. We build a fellow SaaS recruiting platform, a full lead gen platform for our portfolio companies using the same data sets to take your ideal customer profile and then give you all the doppelgangers or a recruit candidate that you're doing and give you, you know, large scale lists of candidates that you can then pull into your ATS or, or generate lead lists that, that hands your recruiters. And we can support an unlimited number of recruiters at an unlimited number of companies at the same time, because it's built in data and software, right?

16:02And you can only get it through us if you're a portfolio company. I heard you say a few seconds ago that, you know, you have a CEO, you have a CTO, you have a CPO. I think in the early days, much of the team was actually engineers, not traditional investors. I thought we would have seen much more in terms of the use and application of data within venture firms. We haven't seen that much. In fact, I've seen it maybe pulled back a little bit in terms of people that are looking to build these models. And some of that has come because there's this natural tension point between does the data inform the decisions and the way the firm runs, or do the people then inform sort of the systems?

16:41How do you think about sort of these natural tension points? Because traditional investors at VC firms, it is, you know, I'm sitting across the founder, I'm looking into the future versus the data often is backward looking. So there's a bunch of reasons why. Well, some of them we were very fortunate to anticipate. Some of them have been surprises or learnings along this journey. If you look at the top quant hedge funds, because the way that they approach the business, the way that they recruit, the way that they construct a portfolio is radically different than a traditional analog hedge fund or mutual fund, right?

17:16There's only so many companies a human can analyze deeply in order to have conviction. And so they build much more concentrated portfolios and have a very different sort of approach. The reason the hedge funds that are quantitative tend to persist over much longer periods of time is they're taking many more micro risks and taking advantage of many more arbitrages over a much larger portfolio because they're rules and data-driven as opposed to sort of manual. So they may be in and out of a position in seconds, right, let alone by the end of the day, whereas many of the hedge funds are much longer duration because the opportunity cost of analyzing that company deeply takes a long time.

17:55So if you actually look at the quant hedge fund world, not a single one of the top quant hedge funds started as an analog fund and converted, right? They all started de novo as quant funds. Because Fidelity is not thinking about how to run fiber optics from Chicago to New York to shave milliseconds off a trade. Or how do I co-locate to the closest server to the exchange to shave another millisecond? Or how do I ingest that Goldman analyst report using NLP so I can trade off sentiment before you can read the first sentence as a human? It's not, you build different structural advantages. As a result of that, it's very hard for companies to sort of transform.

18:31We've seen the technology industry all the time to move from one paradigm to a modern technology stack or an AI-driven approach versus a sort of traditional SaaS approach, whatever it may be. It's very hard to reinvent yourself because in the quant hedge funds, the quants are the kings. In the traditional analog world, it's the portfolio managers or the analyst or whatever it is. And if you look at the venture world, it's the founder-facing deal partner that gets all the heat and light. right and the it the data guys are the it guys in the broom closet right and they're lucky they get any light you know let alone like free lunch this is where you know from day one we made everyone at signal fire on the investment team everyone on the entire firm gets carry we tell everyone on portfolio support you're a feedback luke on these companies and can help inform investment decisions they may not be deeply involved in the first investment but they're very deeply involved in follow-on investments in those decisions.

19:25How are they building their team? How are they converting customer leads? What is their learning trajectory? Because those are all leading indicators of ultimately how a company will thrive. And the engineers and data scientists are involved in the entire process. They source leads, they can help with diligence, and they actually do consulting advisory work to the portfolio to keep it interesting, but also to elevate them to their deserved stature as equals within these firms. From day one, we had no deal attribution. We did a lot of things structurally to make sure that they understood that this is a firm that deeply appreciated and cared about equally what they were doing on the engineering side as the quote-unquote deal team was doing on the founder-facing side.

20:09And all of that was critical to, we have the same CTO that we had for our founding. Our two top data scientists have been with us for close to a decade. You know, this is the consistency is very different than other venture firms who churn through whatever engineering efforts they have over and over and over again. And you also have to have a conviction and a tolerance to go through the mess of, I mean, we're seeing it in generative AI, right? The outputs on this stuff is noisy, right? And you've got to do all the annotations and the editing and the training to get it better and better and better.

20:41So we've been willing to go through that pain for a decade now. And that's very hard. No matter how much money you throw at this problem, it takes a long time in iteration to get it right. No different than the best LLM take tons of time and capital to sort of get them tuned to high fidelity. That's been something that's really hard to do in a traditional venture construct. As a senior partner, you're not incentivized to make a five, 10-year investment in technology that may not yield anything or may have negative yield in the short to medium term during your tenure. You're incentivized to throw bodies at the problem that will immediately pay dividends, even if it's not the right move long term strategically.

21:22When you look at all of those different sort of functions, I think about, we can go back to the earlier point of sourcing, picking, winning, portfolio support, portfolio construction. All of those things have what I see is a structural and systematic way of operating using both humans, as well as using sort of the technology that's purpose-built for what you do. When I think about those things, though, venture is a long feedback cycle. It takes a long time to know if you're doing well. At the same time, companies have OKRs. They have KPIs that they manage to. How do you think about sort of the success of this model, which is novel in terms of it takes a long time to get DPI.

22:04It could be 7, 10 years before you really know directionally where the portfolio is going to go. So how do you think about KPIs and OKRs? Yeah, I mean, the book on OKRs was literally written by venture capitalists, and yet almost no venture firms actually operate on that. We do, of course, and try to dog food our own thinking. But, you know, it's different for every function, right? On the sourcing side, you're looking at not necessarily the outcomes, but are these leads of companies that are high fidelity with companies that get invested by high quality investors upstream? If they are, I want to see that lead.

22:48Even if we choose not to invest, I want to see that lead in a window of time that I could have invested and led the deal. right it doesn't help me to get an alert after you know one of my peers has has led the investment that's a tombstone of a missed opportunity that's not that's not helpful signal on the portfolio support side are what's the engagement it's very much like a consumer enterprise app over 65 of our companies are on our platform on a weekly basis so you can look at utilization rates right this is product market fit the same things that we look at for a sas company apply to ourselves and what is the output of this right we use humans to train our founders and their teams on how to recruit and best practices and interviewing and calm studies and all this kind of stuff you know and techniques of sourcing but then we also give them sonar to make sure they're fishing in the right waters and give them leads on who's actionable right now who's in market who's showing signals of that who's high quality those types of things how do i give them additional information on how to contact them so they're not trying to do it through some channel that they ignore like, you know, LinkedIn Inmail or whatever it may be.

23:54All these things we're constantly doing. Then we do surveys and we do quantitative studies and we use the same analytics systems on usage that all of our portfolio companies would use, right, on their own products. Which pages are people visiting? What's the duration on it? How many people are using it? Are they the right people in the organization? Et cetera, et cetera, et cetera. We brought in product leaders who had run these types of products at Facebook and had founded companies. And so we have had a product who runs it no differently. We're selling and marketing to founders. So we brought in the CMO of Stripe, whose job on customer adoption was a freemium approach selling to founders as early as inception.

24:38It's the same sorts of techniques. We bring in domain experts that are going to use the same best practices. I mean, the irony is all these playbooks were written by our own founders many years ago. And the venture industry is just beginning to use any of them if they're using them at all. So there's very proven playbooks that we can use to make sure that we're making efficient, higher ROI investments. And we sunset products if it's not getting traction, no differently than a company would. And making sure that we then reallocate resources on things that are higher impact for our portfolio companies.

25:13And then the world changes. You know, it was all recruiting before the financial correction, but now it's on sales efficiency and other types of things as companies have to make better use of the resources that they have. You know, we have to adapt with the times and constantly be challenging ourselves and asking the hard questions. Are we getting the ROI? No different than the same sort of product review process would be at a tech company. Why do you think, I mean, if you kind of go back, I mean, venture has been around for 60 years, I think 70s and 80s, when it started to become a little bit larger and of meaningfulness in terms of some of the big firms coming to market.

25:52But it hasn't really changed too much in how it operates. There's been small changes around the edges, but fundamentally, what you're describing is a tech company and you have your clients, which are your portfolio founders. You have LPs that are your shareholders. You have your advisors that are acting as maybe the third part of this marketplace that are fundamentally adding value to this nucleus of overall value that you have as a company. Why do you think we haven't seen as much evolution in the venture market toward models like this? There have been lots of people who have tried. It's really hard to bootstrap.

26:32We had to do all sorts of unnatural things that causes all sorts of headaches and problems in order to get off the ground. For example, on our first fund, a$50 million fund, our AWS bill was half our management fee. Our credit card data set was more than the entire management fee, and we haven't hired anyone yet. And so we had to do corporate advisory and recruiting and all sorts of different things to have other sources of revenue outside of management fee, and we had to raise money from day one. And so we went out and found the most important angels and advisors and technologists, et cetera, in the ecosystem that had massive angel portfolios across them.