If youâre a tech and investing nerd, youâll love this conversation with Dan.
We talk through everything heâs learned reading dozens of academic research papers on startups and venture capital, debunking many popular industry narratives.
I guarantee youâll learn something new.
We talk about the dangers of pre-mature startup scaling, the importance of origination stage investing, the concept of startup catering and why so many startups look the same, plus why he thinks the VC mega funds are still an experiment, yet play a crucial role in the ecosystem.
We also discuss what the data says about concentration vs diversification in a venture fund, specialists vs generalists, what VCâs get wrong about pattern matching, and why pivoting is more valuable than you think.
Support this Episodeâs Sponsors
Ramp handles all your expense management, automatically. Its a corporate card and expense management platform that over 40,000 companies, like Shopify, CBRE and Stripe use to streamline their financial operations. Ramp gives you visibility and control without the paperwork.
Stop chasing receipts: check out ramp.com/ThePeel to get $250 and see what a corporate card can actually do for you.
Harmonic is venture capitalâs worst kept secret.
Itâs the startup discovery engine used by firms like First Round, DST, 8VC, Redpoint, and many others Iâm not allowed to say publicly.
Put your startup sourcing on autopilot with their AI agent, Scout. Try it here at https://harmonic.ai/turner.
To inquire about sponsoring future episodes, click here.
đ Stream on Spotify and Apple
Timestamps to jump in:
5:18 Equidam: helping investors value startups
6:43 Whatâs the required rate of return in VC?
9:29 Venture capital needs new definitions
18:23 Are we in an AI bubble?
24:07 Re-branding early and late stage venture
28:25 We need more origination stage capital
40:05 Survivorship bias in emerging manager outperformance
42:57 Incentives driving larger fund sizes
48:10 Raising overvalued rounds re-risks a startup
52:08 Startup catering: why all startups look alike
58:42 Are VC mega funds still an experiment?
1:08:06 Late stage VC is competing with PE
1:13:42 a16zâs Fund 1 strategy
1:18:18 How diversified should VC funds be?
1:25:06 Performance of Generalist vs Specialist firms
1:30:35 How to value a startup
1:40:58 Why VC firm location correlates to returns, but startup location does not
1:44:05 Founder background doesnât predict success
1:48:27 Startups with one pivot are most successful
1:50:24 Premature scaling kills 70% of startups
1:54:47 Does mega fund model work for origination investing?
1:56:15 Value of Twitter and writing online
Referenced Research Papers:
Referenced Books:
Other Referenced Items:
Find Dan on X / Twitter and his Blog
Related Episodes
What Investors Get Wrong About Asset Allocation with Dan Feder at the University of Michiganâs $18B endowment portfolio
The Past, Present, and Future of AI, Robotics, Venture Capital, and Crypto with Michael Dempsey, Managing Partner at Compound
How Solugen is Saving the $5 Trillion Chemicals Industry with Co-founders Gaurab Chakrabarti and Sean Hunt
đ Stream on YouTube, Spotify, and Apple
Transcript
Find transcripts of all prior episodes here.
Turner Novak:
Dan, howâs it going? Welcome to the show.
Dan Gray:
Thank you very much for having me. Itâs great to be here.
Turner Novak:
Yeah, I think thisâll be fun. Iâve been following you on Twitter for couple years. I donât know. I donât know when youâve really ramped up on there, but you share a lot of really interesting stuff and thatâs what I wanted to talk to you about, a lot of research around venture. I feel like thereâs not enough people talking about that kind of stuff, so Iâm glad that youâre here.
Dan Gray:
It does feel like I found a niche, an underexplored niche. I really enjoy shedding light on some great research around VC that has really important implications for investing, but doesnât get the attention it deserves.
Turner Novak:
Okay, and is it related to what you do? I know I just wanted to ask you super quick, so your title is head of insights. You work at a company called Equidam. What is it exactly?
Dan Gray:
Yeah, Equidam, our mission in a nutshell is to make really unique, innovative, idiosyncratic companies more legible for investors to better explore... Thereâs this two-sided equation of pricing and venture. One side is fundamental value. What is the future cash flow going to be of a company, and how does that discount back to today? Financial math side of valuation, and then thereâs the auction house mechanic, where you compete on pricing and bid stuff up.
Our perspective is the first part is poorly understood, partly because itâs complicated and itâs a chore to get all the data-
Turner Novak:
Pretty hard.
Dan Gray:
Yeah, for sure.
Turner Novak:
Yeah, I mean, how do you predict the cashflow of something thatâs not even a business yet? Thatâs hard.
Dan Gray:
Yeah, exactly. Our perspective is if we can make that easier and better understood, then itâs good for everybody.
Turner Novak:
And it seems like you probably just spend a lot of time reading and doing research around all this stuff.
Dan Gray:
Yeah, I would say so. I mean, itâs extracurricular for me in most of the VC research side, I look at it, I share, or I write about is personal interest. Probably about a quarter of it relates to Equidam. I did a lot of work quite recently on understanding the typical required rate of return for VC, so then we build that into our methodology for the VC method.
Turner Novak:
What is the required rate of return in venture?
Dan Gray:
It basically boils down to a pre-seed investor should be looking to hit at least a 200x in value. And then as you go up through the stages, it obviously goes down in the multiple that you need. But thereâs a ton of maths behind it: looking at the dilution you expect to occur over time, the success rate at each kind of milestone that a company goes through, a few other pieces in there that basically tell you what you need to achieve for an investment to be good.
Turner Novak:
Interesting. I swear Iâm not lying to you, that is my general kind of window I think about where you probably want 100x return on capital, you want 100x just generally return, increase in share price or increase in value or position.
But you usually get diluted. I usually just say you probably get to lose. Might be more, it might be less. I mean it depends, and so I just say usually you probably need 200x return in valuation in order to make the math work.
Dan Gray:
There we go.
Turner Novak:
Yeah, I mean, itâs super lazy math, but I feel like people overcomplicate a lot of this stuff, too. It is pretty simple. Itâs like you buy a piece of a company thatâs not worth very much that could be worth a lot in the future. Thatâs all it is, and getting that right once is very difficult. The more times you do it in a little basket of investments, the more likely one of them will happen, and I think thereâs a trade-off of, do you need to do it 1,000 times? You need to do it two times?
I think the data says 40 is somewhat optimal, highest chances of optimizing returns. Is that right? Or I guess it depends on the stage.
Dan Gray:
It depends. Yeah, I mean, thereâs obviously very much diminishing returns after a certain point. I think you can go up to about 80 and itâs still reasonable. Yeah, certainly 40 to 80 is a good space to aim for.
Turner Novak:
Yeah, I think public markets, they generally say past 20 to 25 you start to get diminishing returns to diversification, which is interesting. I guess itâs the level of liquidity that you have that could be it, or maybe itâs maturity stage of the businesses.
Dan Gray:
Yeah, predictability is a big factor.
Turner Novak:
Yeah.
How do you think about just venture today? What are your general thoughts on the market when youâre thinking about it from everything thatâs going on and what you know should be happening, shouldnât be happening? I donât know.
When I ask you about this, whatâs your response?
Dan Gray:
The hot take is what is venture, really? So much of the activity that weâre all looking at at the moment is technically venture because itâs like venture allocation. Itâs in the same bucket for LPs as venture capital.
But is it really? I would say maybe not. If youâd asked a lay person what venture capital was 15 years ago, their imagination wouldâve been like a small firm managing a ton of money going out, looking for people, building crazy stuff in their garage and making outsize returns of that kind of investing. Whereas over the last 15 years itâs increasingly become financial engineering and investing through a spreadsheet. Engineering markups with lazy processes and more oriented towards the fee collection part rather than the carry part. All these incentive problems that have been pretty well-hammered, but nevertheless it seems to be the case.
Turner Novak:
Yeah, itâs interesting when you just say you pull up a pitch book or any of these data or at least a report on the amount of capital thatâs invested in venture capital as an asset class over time. We always see the big takeaways and the top five numbers, and itâs like the past probably year itâs been like OpenAI, Anthropics, SpaceX, maybe Ramp. Maybe Iâm biased because they sponsored the podcast. Thereâs probably one other one thatâs raised hundreds of millions, billions of dollars. Thereâs going to be a deep tech one that Iâve never heard of thatâs raised a billion dollars.
Those are probably pretty mature businesses, all those companies. OpenAI is probably one of the top five most used consumer products in the developed world. Itâs not really a startup anymore. You can argue margins. How profitable is it, really? They do generate a lot of revenue. People pay for it, but itâs not a startup anymore. Thereâs not uncertainty on will people use the product, so itâs almost like is it a different type of financing? Is it fair to classify the OpenAI $40 billion round that theyâre putting together, whatever the number is... Theyâre building data centers, whatever with the amount. Versus a pre-seed, theyâre raising a million dollars to whatever, something completely different. Not AI, itâs like biotech or something. Weâre inventing cancer drugs.
Theyâre just different things, I feel like.
Dan Gray:
Youâve also got to think about the definition of a startup is surprisingly not simple, I think. Is OpenAI a startup? For all the reasons you mentioned, probably not. But also theyâre thinking about branching into wearables. Probably they have a ton of stuff on their product roadmap that we canât even imagine right now. Is the chat interface going to remain the primary way that we interact with their products? Maybe not.
Thereâs still a ton of risk left in the company or a ton of uncertainty left in the company, which maybe means it is a startup. Itâs just an unprecedentedly large startup.
Turner Novak:
Yeah. Well, I think one of the recent episodes that came out right before this was with my friend Dan Feder, the University of Michigan, and he did make the point. I think he buckets into two categories, where itâs capital for ventures and ed to venture capital. Itâs basically ed to venture capital is like weâre going on an adventure. Weâre trying to figure this thing out. And then capital for ventures is like, you have a venture. You have a business. Weâre going to put some money in because we know it works and you scale up this thing that works.
Yeah, and itâs an interesting point of thereâs probably some companies that are only raising a couple million dollars where itâs just a capital for their venture. Itâs like itâs not a lot of risk. Itâs like we made a SaaS tool that works. Itâs this AI agent thing. We have good margins. Everything looks good. We want to put the pedal down and go really fast, and we know this works. Versus to OpenAI, we donât know if people are going to use their wearable products. Who knows? They might not.
Dan Gray:
But unfortunately the fundraising for that SaaS company I think is unnecessarily difficult today because they donât fall into the camp of explosively fast-growing AI companies I think, which are absorbing obviously a huge amount of investor attention.
There was more experimentation with different models of venture, a few firms that tried to do more moderate risk investing, looking at maybe smaller companies with more stable growth with better margins, but not traditional venture levels of risk. But most of that seemed to die out in 2022. That reset took away whatever little innovation thereâs been in the venture market in the last couple of decades. It really started to emerge in 2020, 2021, but sadly was washed out the next year.
Turner Novak:
Iâve definitely seen people, I donât know who is actually doing it, I bet thereâs people that are doing it, but really itâs a dividends. Youâre actually paying out dividends because the companies have cash flows.
I think the debate there is always like, why are you giving people an 8% return with a dividend, or whatever the number is? They should be reinvesting in the business to grow faster. I think itâs the trade-off with a startup, which I think to keep it simple, you probably shouldnât do that. It is easier, I think, and more simple. Itâs easier to bucket it in. If youâre an institutional investor, youâre like, âOh, Dan, you have this fund. You just do early stage 50 companies.â Theyâre all going to fail except for one. This is a good strategy, we get this. But if youâre like, âOh, but some of them... We might be giving you cash flow back.â
Some of them are tokens because itâs crypto, and some of them are equity, but half of them are safes and even some of the safes are going to start paying dividends. We donât actually have the corporate documents formed. Thereâs income coming through and thereâs all these K1s we have to solve for. Itâs complicated. I think it almost becomes not worth it to dig in for the average LP.
Dan Gray:
For sure. Yeah, I think itâs a specific product maybe for a specific kind of LP that needs returns on a certain horizon.
I think an overlooked fact in a lot of VC commentaries, like the different categories of LP and the different preferences they have, it does shape what they look for a lot.
Turner Novak:
Yeah, because the thing that I think a lot of people forget about is QSBS, which is basically... I mean, I think they actually made changes to this rule recently, but you can get a certain number, I think itâs like 10 million in gains or maybe 50 million in your gains, in things that meet a certain requirement tax-free.
If youâre investing in smaller-ish funds where youâre, letâs say youâre 10% of the fund, you invest $5 million, you get 50 million back or 55 million back, that entire 50 million of gains is all tax-free. Iâll throw a link in the show notes QSBS. I forget how they changed it, but I think you need to hold it for five years. It needs to be a certain type of company. It canât be real estate related or something like that, and it canât be a few other things. Maybe canât be biotech or must be healthcare related or something like that.
Thereâs different classifications for it, but that can change how you make decisions, too.
Dan Gray:
For sure, definitely.
I think thereâs also a second less well understood QSBS provision that allows you to write down some of your losses, but itâs for some reason not generally applied in venture.
Turner Novak:
Yeah, we donât write things down in venture. Itâs not good. It doesnât make us look good.
I actually have a friend who has a pretty interesting strategy, where he donates to charity his overvalued startup stock. You get the benefit of the tax write-off of saying, âHey, Iâm donating this equity, Iâm donating this valuable thing to a charity,â and then it is not worth that much, obviously, because an overvalued startup stock. Itâs an interesting way to do tax optimization. I think thatâs one of those secrets of the âwealthyâ type strategies you might see a clickbait YouTube video for. Itâs an interesting way.
I mean, thatâs part of it too is, how do you optimize on taxes? Not a lot of people talk about, which is less fun. Itâs way more fun to just say weâre backing the next Elon Musk and this is the next SpaceX. Get in or else youâre going to miss out. Way easier to say.
Do you think are we in an AI bubble right now?
Dan Gray:
That is something Iâve spent a lot of time thinking about. If youâd asked me as little as a month ago, I wouldâve said absolutely, obviously, but actually maybe not.
Turner Novak:
This is a classic top of bubble type thing to say.
Dan Gray:
I know, I know. Well, yeah, it depends, right? Because I think when you think about a bubble, youâre thinking about investing in a speculative asset that is massively detached from the intrinsic value.
The problem with that definition as it relates to AI and VC today is I donât think VCs are generally acting like investors. I think theyâre acting like traders. If they can get out of their positions before the thing crashes, then itâs a good trade. I think they in a cynical way recognize that. They all see thereâs a crash going to happen, but they are just hoping they can get out before then. And if they do, itâs a good trade. If they donât, itâs a bad trade, but itâs not necessarily a bubble.
I think a bubble, if we get to a point where OpenAI has gone public, Anthropics gone public, maybe Cognition, a couple of others.
Turner Novak:
I think Mistral was supposed to or something, but I think they recently did a round with ASML, I think.
Dan Gray:
But once theyâre in public markets, then I think we could see a bubble because obviously more liquid market, itâs more responsive to sentiment. Itâll climb much faster and crash much harder.
But for the time being in venture, I think itâs like you have to recognize it as a trade on the wave of AI. If you invested in an OpenAI in 2017, that was an investment. If you started investing in AI companies in 2022 as a strategy, probably thatâs better characterized as a trade on the momentum of AI. You are hoping those companies will ride that trend, generate good markups, maybe get a good exit, but primarily youâre betting on the category.
Turner Novak:
Yeah.
Are you sure 2022? I feel like thatâs a little early.
Dan Gray:
Yeah, that could be a bit early. Letâs 2023.
Turner Novak:
â23, yeah. I mean, â23, it was like the entire market was just scorched. That was the only thing that was moving.
I think thereâs just so many incentives, I talked about with this Dan too and that other episode talking about relevance, where you have to remain relevant as a VC, as an investor. Itâs something I struggled with a little bit, especially during the Web3 mania. When I did my very first fund for Banana Capital, I invested in a little bit more consumer stuff than I do now. Iâd been going into 2019, 2020 thinking that like e-commerce just generally is a trend, and it got stretched pretty far during COVID as I deployed the fund and as I first invested the fund.
I think I messed up on not just rationalizing that. I think itâs so hard when youâre in those moments to really realize whatâs sustained. I personally still donât really like going to stores. I personally do a lot of things online. Iâd rather order ahead and pick something up. But I had a lot in my LPs that are like, âTurner, why are you not a Web3 investor now? We thought youâd be doing Web3. Why arenât you in Web3?â Itâs pretty easy today to look back and...
Dan Gray:
I bet theyâre happy about that now.
Turner Novak:
Yeah, I think that fund will do okay, my 2021 fund that was over-allocated to consumer. I donât know how these things are going to benchmark. Iâm going to make money. Iâll give people their money back. Iâll get a little bit of carry from it.
Dan Gray:
At least on a relative basis to the market, you probably do very well because the market as a whole I think is going to suffer that for that year.
Turner Novak:
Yeah, but thatâs another thing I think you need to remember in venture is thereâs relative and absolute.
In 2021 I was on a relative basis. I was like, âOh, Iâm not investing in Web3. Iâm not investing in these like pre-line of code things at 100 million plus money, Iâm doing it at 50.â Thatâs half off. That seems so reasonable. But then when you back out and look at it over a full economic cycle or a 20, 30-year period, thatâs still a pretty high entry point thatâs still... You might be 50% lower than everyone else, but youâre still two, three, 4x probably higher than you should be. Yeah, theyâre all 10x higher than they should be, but youâre still fucked up at the end of the day.
The point of this is make money. I think the point of venture, you have to basically outperform every asset class. That is the job. That is the role of venture. People arenât in venture to give me some buffer when bonds are down, I need my venture portfolio to do X thing. Itâs uncorrelated, but itâs uncorrelated because itâs like just find people starting things that will be worth a ton in 10 to 20 years.
Dan Gray:
If your definition of venture is what we think of as traditional-
Turner Novak:
True venture capital.
Dan Gray:
... venture capital. Yeah, exactly.
Turner Novak:
Yeah, traditional early stage. Yeah. It is a good point of, I think we just need to think we need a new name for later stage venture, late stage venture. I donât know. I like capital for ventures is a good one.
Dan Gray:
Capital for ventures is good. Leslie Feinzaig calls it consensus capital. Kyle Harrison calls them like the agglomerators, I call them the venture banks. It does need to be recognized as a different strategy. Maybe still within the same bucket, but definitely a different strategy because it changes so much about the expectations.
Turner Novak:
I think itâs hard too because if youâre any of these data providers, you donât always know what the nature of a round...
I mean, this is a classic thing in 2021. Someone would say, âWe raised $100 million for our FinTech company,â and it was five million in equity and 95 million in debt. And so my general, when I see a number of like, âOh, you raised X amount of money,â if it was around a seed series A, I usually assume seed you saw about 20% of the company if itâs a series A. I assumed you probably sold 25. You maybe got diluted about 30 because thereâs some options and stuff like that that went in there.
If I see that number, âOh, we raised $100 million and it was a series A,â Iâd be like, âOh wow, 400 million plus money. Whatâs an appropriate evaluation? They must be pretty far along,â but itâs like, no, they only raised five million. They have one customer. Or actually I guess thatâs the other side is, if you got a debt vehicle that you raised, you probably have a pretty robust business, too. Hopefully you probably passed certain levels of sophistication and majority and able to prove that.
Yeah, I donât know. Itâd be hard to really actually figure out how to do that. Itâs almost like itâs just easier to just not worry about it and just call it all venture.
Dan Gray:
Itâs an interesting point though about what that does to the perceptions of valuation because you have people making calculations, like you said, where...
Turner Novak:
Because itâs all comps. I just say, âDan did this. This is public, this is what he did. Iâm better than him or Iâm at least as good, so I want to get the exact same amount.â If itâs not, then itâs not fair.
Dan Gray:
You have the debt part of that which complicates it. You also have obviously the terms. Was it raised with a 2x liquidation preference? What are the more investor friendly terms were attached to make that valuation work? You never really know any of those things. It has a very cyclical effect on pricing.
Turner Novak:
And then I have some friends who they strategically release and drip certain information. Theyâll say, âWe just crossed 40 million in ARR,â but theyâre actually at 80 or something like that. They strategically will put things behind.
I donât know if you know that company bolt.new, the text to website creation. I kind of figured it out from talking to him and we talked about it on the podcast is heâs like, âYeah, when we announced where we were at, we were actually a little bit ahead of that.â Because he was in a case where they were just exploding and he was every day waiting for it to taper off and heâs like, âIt just never died. I was just afraid that if I set a number and the number was no longer there... So I was like, âOkay, letâs hedge a little bit.ââ
I think that goes into it too is... And then thereâs the other side of weâre like, oh, weâre at 100 million ARR. Itâs like a creative definition of the word ARR. Again, a lot of that stuff is so tricky.
Dan Gray:
Yeah, thatâs a huge problem. Itâs interesting to me as well the extent to which a company hoping for an exit for an IPO knows that eventually one day theyâre going to have to get to gap compliant financials. Theyâre going to be audited, so for how long can you keep that mirage growing early on of the creative bookkeeping? Itâs dangerous.
Again, there are incentives on both sides. Some VCs like it too because it makes it very easy to generate markups on multiples.
Turner Novak:
I donât want to miss this question because we talked about it a little bit earlier, but you mentioned that you think basically all of the returns in venture ultimately come from the first check. True venture capital, the traditional, how we defined it earlier is like youâre starting a company, extreme amounts of risk and uncertainty. Whatâs the importance of that kind of origination round that first round, whatever we want to call it?
Dan Gray:
Okay, itâs axiomatic a little bit because obviously if you donât have a first check, you have no checks. You donât exist.
Turner Novak:
You need that risk capital, the inception stage, origination stage.
Dan Gray:
Yeah, exactly. Investors that focus on that level, the real genuine first check investors, and Iâm not talking about the ones that need to follow someone else into the round and need traction or whatever, they determine what venture capital is in a way. They determine what are the exits going to be in 10 years.
Turner Novak:
How do they determine that?
Dan Gray:
If it never enters the funnel.
Turner Novak:
Got it, okay. Theyâre determining what is allowed to exist, what startups are able to have a shot at trying to do something.
Dan Gray:
Exactly.
Turner Novak:
If theyâre only funding certain categories...
Dan Gray:
Yeah, the quality of what of is in the whole venture fund or venture funding pipeline from origination to all the way to exit at any point is downstream of those investors. Their ability to provide good coverage, to be able to explore every corner of the economy, every category, every industry is super important.
Turner Novak:
Should there be more people doing that, more creative... What does the data show?
Dan Gray:
I mean, thereâs been a very good study done looking at VC in Europe. It showed essentially looking at programs like big programs by the EIF, the European Investment Fund, where they put a ton of capital into VC. You get the same uplift in productivity from that investment if you just move the current allocation earlier than if you double the amount of capital.
Expanding the early stage origination area of the market has just as good effect as doubling the total pool of capital.
Turner Novak:
If youâre investing 100 million in startups, tech, innovation, whatever Europeâs doing, and then they wanted to double their impact, probably the most consensus thing. Well, letâs just give them twice as much money. Theyâll have twice as much of an impact. But it actually have double the impact if they just went half as earlier or how earlier? Do you know how early you have to go?
Dan Gray:
It wasnât too specific, but I think essentially itâs ensuring thereâs enough money at the pre-seed C series A maybe kind of stages. Make sure those companies have a path to really prove the business model and become success stories.
Turner Novak:
Why do you think more focus is around this origination stage?
Itâs interesting, I feel like a lot of people say weâre investors, or we back founders from the earliest stages, or whatever they say.
Dan Gray:
Yeah, the biggest question there is not necessarily who classifies themselves as an origination investor or a first check investor. It is what is the strategy that they apply to origination? Where are they looking? How do they source companies? Is it through a network of peer analysts or firms or other GPs that they know, in which case itâs already kind of coming through a filter, or are they genuinely doing essentially the equivalent to a startup of go to market? Are they doing that? Do they have a unique defined strategy to go out and find their own deals? Thereâs a few examples of firms that I think do that famously well. My favorite is always 1517 fund because they invest $2,000 in a college kid with a dream who has nothing, but that kid is like, okay, now I have the money to test this idea a little bit, and if I then start a company, guess who Iâm going to call? Itâs going to be Danielle.
Turner Novak:
The one who gave me the 2,000 bucks And itâs a grant, right? I guess free.
Dan Gray:
Yeah, exactly. Itâs literally just like at that point itâs a Venmo.
Turner Novak:
Yeah, the corporate Venmo account of 1517.
Dan Gray:
Yeah.
Turner Novak:
I wonder if thatâs actually there. I should ask. And so what do you think about maybe a YC? Is that the example of a good one?
Dan Gray:
Thatâs a very good question and Iâm kind of torn on that because written at length about YC and YCâs track record and their performance, Iâve always kind of defended against accusations that YC companies are overpriced because if you look at the historical success rate, they should be priced higher. They have a higher success rate, higher rate of exits. The question is that involves looking at data from companies that are obviously at least 10 years old. So has YC changed in the last decade, I think is an interesting one to consider. If you look at the breakdown of the cohorts 10 years ago, 15 years ago, how much thematic consistency was there? How many companies were in a very specific category of technology? Essentially how diversified was it across different ideas compared to today, where do you see people saying thereâs 20 different windsurfs for X in YC currently?
Turner Novak:
I think the interesting thing that YC has too is they kind of have this built in price making mechanism. If you think of it, an auction market, thereâs price makers and price takers and YC basically gets... As valuations keep going up, they have been able to keep this lower price that they come in at. I guess to your point, the way the marketâs changed, theyâve changed where they now do put a bucket, a check in at that higher price, that kind of uncapped safe that they go alongside the first check. So I bet thereâs sort of a weird step... There might be a step function change. Maybe some of the YC funds was an 8X fund, one was a 12X whatever, but I bet you get a slight bump down by a full multiple that might be a 7X or an 11X.
But itâs probably a bigger pool of capital that you get to work with. So itâs probably okay. If Iâm an LP, itâs like, okay, instead of my LP exposure to YC being 5 million bucks, I had to put in 20 and Iâd actually rather get a 7X on 20 million than an 8X on five. I donât know, I have no idea if those are YCâs numbers. Those are pretty good numbers. I bet theyâre actually higher than that, honestly is my guess. But to your point, I think an interesting kind of topic that brings up though is looking at historical data, some of these studies are in completely different market structures, completely different market regimes, so thereâs one end of the spectrum where you could say things are different. You canât rely on this data. All the maybe non-data supported things that the venture industry does is actually correct because the old... All the dataâs wrong.
The other end of the spectrum, that is just a classic thing that you say when everything is about to be over. History doesnât matter. The data we donât follow. Itâs like the whole, do you want to make spreadsheets or do you want to make money argument?
Dan Gray:
Yeah, I share a lot of old papers. I get a few different forms of pushback to this. I shared one a couple of weeks ago and a GP replied to me and he was like, this is from 2017. Thereâs no way this is relevant today. Venture capital in the traditional sense that weâve been talking about has not really changed that much. Itâs not changed much at all for 30 years I would say. Thereâs the bigger end of the market which has been added, but the traditional part, the risk factors, the decision making process, the portfolio strategies, all of that is the same.
Turner Novak:
So itâs sort of like venture capital started as this pretty small pool, very beginning stages of the market, and it stayed there, but itâs also expanded to become further and further along and all that expansion is like 10X, 100X bigger than what the first initial inklings of it was. Itâs still there, but this whole expansion just takes up most of what we think of it today. So itâs just like itâs still the same, but itâs evolved I guess. Thereâs new extensions of it
Dan Gray:
And I think itâs evolved in a very expected way in that the larger itâs gotten, returns of commensurately fallen and you almost think of dilution of the strategy, itâs literally more capital which allows for the scale of returns to be increased but not the efficiency of returns to be increased.
Turner Novak:
What does that mean? Can you explain that?
Dan Gray:
Itâs like if youâre a big sovereign wealth fund and you have to write checks of a minimum a hundred million into funds that youâre investing in, you can put that into a venture bank, a 16Z or general catalyst type firm and they can put that to work and maybe youâll get a two to 3X on that money. You canât really invest it in the early stage of the market because it just doesnât fit. You canât squeeze the check that large into a smaller manager.
Turner Novak:
Youâd have to do 10 or 20 different funds to get that $100 million deployed.
Dan Gray:
Exactly. But if you could break it down like that, maybe you could get a higher total multiple, but itâs more work, more risk potentially.
Turner Novak:
Thatâs sort of the pitch for a fund of funds then where itâs like, hey, give us a hundred million. If we have a $500 million fund in the top whatever, 20, 30, 40, 50 best managers and all their breakouts. Itâs kind of like hiring a consultant to manage your venture portfolio for you, but they just kind of do it. Itâs one line item on your portfolio and one headache. You have to manage one update call, one relationship versus 30 or 40 or 50, which thatâs a lot of time. You need a whole staff to do that.
Dan Gray:
If thereâs one startup idea that I could be tempted by in future, it would be trying to design an algorithmic fund to funds, so it could be very low fee, very efficient, but identify... Basically index emerging managers to try and manage a lot of the uncertainty involved in that, but to build a process around managing the biases involved in selection, all the uncertainty involved in selecting managers with no track record. It would just make for that section of the market, which is like the origination layer too, which I obviously care quite a lot about. It would make fundraising so much easier if they had access to that kind of capital.
Turner Novak:
I donât know if itâs true, but Iâve kind of had this maybe hypothesis. Is there some sort of survivorship bias in... I think itâs fairly well known that people say emerging managers outperform, fund ones outperform and you should just... If you want true venture, if you want the best venture returns invested in fund ones or whatever, emerging managers. But Iâve always been curious, is there a survivorship bias of letâs say you and me both start funds, you crush it, youâre doing really well. I fucking suck and I just drop off and does my data leave? And so all that they see is you crushed it. I sucked. We offset each other to equal or maybe weâre worse, but I dropped off. So only your data shows up. Have you seen anything related to this before?
Dan Gray:
Thatâs interesting. Nothing specifically to that. I mean, I will say some of the perception about how emerging managers perform or outperform is based on a lot of charts that the pitch book have released where they show who have been the top 10 funds in each vintage by size.
Turner Novak:
Is this pitch book or Cambridge maybe?
Dan Gray:
Could be Cambridge, you might be right about that.
Turner Novak:
Theyâre biased. I mean, theyâre trying to sell institutions on new funds.
Dan Gray:
Thatâs true, but they consistently have shown small sub-fifty million funds, for example, dominating the top 10 each year. But actually if you look at the underlying fund data, those small funds are massively overrepresented in the data set, so itâs obvious that theyâre going to take more of those top 10 positions. It doesnât necessarily mean that there are a better bet because there are a lot more of those funds that also obviously donât do well.
Turner Novak:
Thatâs true. I mean, itâs extremely difficult to raise capital for a fund and especially when you have not done it before or youâre first getting it started, so thereâs probably not very many large early funds. And also if youâre good, you just keep getting bigger. So itâs kind of hard to be like a... Itâs really hard to just stay small.
Dan Gray:
Which is one of the problems with the origination part because if small firms focus on origination but success naturally inclines them to grow, then they slowly move out of being able to invest in that such an early stage and they start to invest larger checks later rounds so that whatever talent they had at origination is kind of washed out, which is a shame because thereâs already very high churn in the industry. Thereâs already relatively weak persistence, so itâs a difficult part to make talent stick.
Turner Novak:
Can you explain for somebody who may not understand this of why do you want to have a bigger fund, why does small funds that do well ultimately end up getting bigger?
Dan Gray:
I think thereâs probably two main forces. The first one is the obvious one, which is the financial incentive because you have as a part of your compensation as a GP, you have your management fees and you have your carry. And the carry is, you could consider it like a performance bonus if your investments do really well past a certain hurdle, you get to share in that profit, which is if youâre a great investor, itâs incredible and can be a huge part of your compensation.
Turner Novak:
You need to liquidate the asset though. It takes a while. In venture, itâs probably like, I donât know, eight to 10 years until you get any of it. Probably longer, honestly.
Dan Gray:
You have to first return all the capital to LPs. Itâs a headache. Itâs difficult. Whereas much easier is the other part, which is the management fees, which are guaranteed you get them pretty much no matter what and thatâs like 2% a year or probably these days more like 2.5 I think a lot of the time, probably over 10 years. But then thereâs extension periods and caveats around that as well. But if you consider it like 2% 10 years, itâs 20% of the fund.
Turner Novak:
Yeah, you raise $100 million, you get $20 million guaranteed. You think of that like an ACV as a SaaS business, thatâs an incredible business. Thatâs you raise a billion dollars, you get 200 million in revenue, itâs guaranteed over 10 years. Thatâs an incredible business and you just keep doing it again and you raise another fund. So you raise one fund, thatâs a billion, you get your 200 million in revenue, you raise a 2 billion fund, because you did well now you got 400 million stacked on top of that 200 million and just like letâs say you started, your first fund was a $10 million fund and you got a 10X return on that fund, you took that 10 million and you turned it into a hundred million. So you got 90 million in profits and you got 20% carry, you got 20% of that, so you get 18 million bucks in profit, you probably get it 10 years, 18 million in profit. That is like if you go back to that billion dollar fund that you raised, youâre getting 20 million a year.
Itâs like, why am I around waiting for all my carry? Just raise a big fund. It does make sense, but I think the issue then though you might trade off, and this is something Iâve thought a lot about, is you have to build that infrastructure, pay for it, which is usually a team, a lot of people. You have to manage a lot of things. You spend a lot of time on that operational institutional design. So yeah, you get a lot of management fees maybe, but depending on the circumstances, itâs not quite as lucrative as you might think, but you do get leverage depending on who you bring in. You may be investing in things that are defensible network builders or brand builders or something like that. So it gives you more capital and power to work with.
Dan Gray:
I think itâs kind of useful to look at the extremes. Consider a small 50 million fund, just a couple of GPs and very little else in the firm. Theyâre spending their management fees on living, travel expenses, events, whatever powers their origination strategy that they want to do there, but itâs basically covering the cost of operations and then they cross their fingers that the carry is great, whereas the far other end of the spectrum, like an A16Z, the range of services that they provide, startups now they have their own compute that they provide to AI companies that they invest in.
Turner Novak:
Oh, so that is a thing. I thought that was a meme.
Dan Gray:
No, no, thatâs a thing.
Turner Novak:
I thought it was a joke, itâs real? Okay.
Dan Gray:
Thatâs a thing. And they have other financial products that they offer. They have the whole platform team.
Turner Novak:
They have anything you might want. Any department that your company will have, whether itâs recruiting, marketing, capital raising, go to market, yeah, growth, brand, engineering. They have consultants that can help you come in. Itâs everything. Itâs like any role you might hire for, they have it. They can help you with it. Itâs an interesting value prop to some companies. And thereâs also just the brand halo of A16Z. All your friends know A16Z. Theyâve all heard of them, theyâve seen the podcasts. Your parents probably use the Mosaic browser back in the nineties. Why would you not take Mark Andreessenâs money versus some random dude with a podcast that no oneâs heard of? Itâs an interesting value prop.
Dan Gray:
It is. Thatâs a whole fascinating rabbit hole itself. What is the cost of taking money from these big firms?
Turner Novak:
What do you think it is?
Dan Gray:
Thereâs a great thread by a VC called Rob Go, and he reflects on this as re-risking because essentially their value prop, part of it at least is that they will be less price sensitive when they invest. So they will give you much more money on much more friendly terms potentially.
Turner Novak:
This is a larger agglomerator of capital type fund?
Dan Gray:
Correct. Yeah, yeah. Like an A16Z. Exactly.
Turner Novak:
Not Rob Go, next view. Is that next?
Dan Gray:
Yeah, thatâs right. Thatâs right.
Turner Novak:
Okay.
Dan Gray:
So the downside of that is the growth that you kind of obligated to deliver as a part of getting that huge check is way higher. The stakes of failure are higher. The risk youâre taking on as a company by accepting that much money is much greater. The fragility of your future vision, how you can adapt if things donât go to planned, thatâs much more difficult. So essentially what should happen through a startupâs life is as they develop, the risk slowly goes down. You end up raising these huge rounds and your risk shoots back up to the top again, and then you need to work it down again. For some companies in some categories, itâs the right move. Yeah. But for many itâs probably not.
Turner Novak:
How do you decide that? Is it how much capital do you need? Is it a competition? Is it a certain product type, like capital efficiency or something? Do you know if thereâs a way to gauge that?
Dan Gray:
To some extent, itâs a bit of a competitive force. I think if all of your peers are in the same category are raising similar amounts of money. To some extent, maybe you kind of feel like you have to as well to compete with them, and that ends up with an ugly situation where everyoneâs spending huge amounts of money bidding their own margins into the ground against each other.
Turner Novak:
Yeah. Some kind of a paradox. Iâm sure thereâs a scientific word for this, but some sort of losers prisoners dilemma type paradox or something.
Dan Gray:
Yeah, itâs basically venture predation. Thereâs a good paper on that.
Turner Novak:
Iâm going to throw Robâs thread and Iâll throw this one in the description too.
Dan Gray:
Yeah, that oneâs looking at, I think it uses specifically the example of Uber versus Lyft. So they both absorbed a ton of VC money and a lot of that money was put to work basically trying to suffocate the other company. Who can undercut the other companyâs pricing enough that you steal all their business and can you survive long enough doing that before you die yourself?
Turner Novak:
Interesting. And itâs crazy how I think Uber is... Uberâs this fascinating example where itâs super profitable now or something, right? They just basically 3X prices over the past couple of years and itâs all free cash flow. I havenât looked, I donât know the number, but itâs pretty much they just had a ton of pricing power that was kind of hidden and weâre all paying it and theyâre making a ton of money now.
Dan Gray:
They also nailed their timeline. I think the period of time from inception to exit that they had luckily was kind of perfect. They could have exited maybe a year or two later. I think it was 2018, their IPO, something like that.
Turner Novak:
Yeah, 18, 19 probably.
Dan Gray:
But if theyâd been intending to exit in like 2023 and then suddenly they wouldâve hit that speed bump when all the capital dried up, that could have been fatal.
Turner Novak:
And they kind of did hit some speed bumps right before the IPO also kind of with the whole management team turnover.
Dan Gray:
Yep. Yeah, thatâs another interesting rabbit hole.
Turner Novak:
And you turned me onto this interesting concept, I think itâs called startup catering. I think thereâs a research paper on this. What is startup catering for people who donât know? I think of food, I think of getting lunch for the company, but what is startup catering?
Dan Gray:
Itâs probably my favorite paper on VC, which is an incredibly nerdy thing to say, but whatever, Iâve said it.
Turner Novak:
Hey, weâre in an audience of nerds here. Weâre all on the edge of our seats.
Dan Gray:
Itâs a safe space. It kind of crosses over the valuation part, which is obviously what I spend a lot of time doing with the what gets funding question, which I think is really important. So essentially catering, itâs a consequence of the fact that investment decisions in venture capital, mostly relative judgments. So you look at if youâre going to invest in a company and youâre trying to understand is it a good investment and if so, what price should I come in at? You probably look at peer companies, you look at comps. So if you were investing in a B2B SaaS company in 2021, pricing was super easy for you. Thereâs thousands of them. You probably know a ton. Theyâre at every stage in every region, in every vertical. So you can find a ton of companies that are very similar to the one youâre looking at.
Turner Novak:
And itâs from all stages, from inception stage, all the way to mature public markets. I have a ton of comps and thereâs a history of how these things go.
Dan Gray:
So in terms of information friction, to understand the company you are looking at through the lens of all these comps, itâs super easy, makes investing very simple. It also means downstream funding is going to be very easy for that company. When they get to the next round, theyâll be able to raise very easily as well, probably at pretty good terms. Itâs just like life is so much easier if youâre in one of those categories. The downside obviously is that itâs like a blood-red ocean, super competitive, maybe overvalued if investors are all piling in there. Thereâs real downsides too from the investorâs perspective, but the problem from the founderâs perspective is if you are trying to do something that is really unique and innovative, you are like Valoratomics or Rainmaker or someone, you go talk to VCs and try and explain to them your company and why they should invest.
They have no comps. Thereâs no relative judgment to make. Thereâs no other set of peers they can look at to understand your company. They have to if they want to invest and probably they should because those companies are often incredible. They really have to understand first principles foundational value or fundamental value, and thatâs much more difficult. So the upside of all of this basically is that founders basically have such a hard time raising money for these really interesting, unique ideas that they tend to cater to VCâs preferences. They tend to cater to where thereâs less information friction, so theyâre more likely to go and build a SaaS, which I think we both agree is pretty sad.
Turner Novak:
Yeah, I mean, you can solve problems, but the easiest problems to solve are probably the most over catered to I would think. So then why are the VCs investing in it? Because you just said, I mean, based on the last, I donât know, 53 minutes of our conversation, I would lead to be like, yeah, those returns are probably going to be pretty low. Why do people do it then?
Dan Gray:
Gets into the question of what is the primary motive of VCs today?
Turner Novak:
What do you think it is?
Dan Gray:
Depends. I think thereâs examples of both. I think by far the most capital is moved by investors seeking markups, seeking proxy metrics that show their fund is doing well, so they can raise another fund and get those management fees. I think that moves way more money, but ultimately I think the most returns over time are generated by true investments, which is like you invest because you see this company as a generational outlier and you think the exit is going to be huge. Youâre focused on the end game DPI rather than the proxy metrics along the way.
Turner Novak:
And is there a good way to do that? If Iâm listening to this and Iâm trying to change my ways, Iâm like a markup addict and Iâm trying to get clean of the markups and purify myself with DPI. Is there anything youâve kind of seen that maybe you can kind of think of for cling to get through that?
Dan Gray:
I think thereâs something said by Michael Dempsey in your podcast with him, Michael Dempsey from Compound, he talked about, if I remember correctly, it was their 2020 fund and they intended to deploy it over four years, but they ended up taking five because they just didnât find the opportunities that they wanted to invest in. Thatâs such a good attitude, and I think you would find that is relatively rare in VC, whereas most people are just trying to put capital to work, get it out the door, get the next fund, but they said, weâre only going to invest in companies that pass a certain bar, and if it means we have to wait longer, which means we essentially have less income to our firm, then weâll wait. And I think thatâs a great attitude. They definitely suffered for it in terms of versus the income they maybe could have had if they deployed faster, but long-term, the right choice to make and those decisions I think are how you build a generational firm and end up being able to charge 3% management fees if you want to in the future.
Turner Novak:
Well, because sort of related to this whole raising the mega fund model, big pools of capital, from my perspective, I kind of see it as it kind of works, it kind of solves this problem in the market. I donât think itâs going away, but you mentioned before that you still think itâs actually an experiment, so why is it an experiment still? It kind of seems like it works. I donât know, and maybe thatâs not true, but...
Dan Gray:
Yeah, thatâs a good question. I mean, I agree with you that I think it definitely has a place in the market, and I went personally from two, three years ago, I was very critical about it because I thought itâs just basically grifting, itâs just collecting fees. But actually, if you look objectively at what the big firms have done, itâs pull a ton of capital into tech, which has got to be a good thing. Ultimately, the question is how does that strategy work out for their LPs?
And I think we kind of have to see to an extent just how that shakes out over time. For example, I think AI is kind of the backbone of their strategy at the moment. Itâs where they dominate. They have enough money to back the big foundation language model companies. They have the clout, the media presence, the marketing pull to not just promote the portfolio companies like normal, but to promote the whole category, to build momentum for it, to push it in Washington even, and certainly to pull in other investors to herd in more capital. So to an extent, are they manifesting a wave and then trying to ride that wave by indexing a ton of companies that kind of join onto it? Is that going to work? It could be incredible. It could be six, 7X on a huge pool of capital, which would just be amazing results. It could be a failure ultimately, or it could be something they canât repeat in future. Thatâs why it seems like an experiment to me. Is it going to persist beyond the current AI wave?
Turner Novak:
Yeah, that is a question. Itâs like it feels like weâve... If you just look at the history of venture over the past 15 years, itâs like we keep finding new bigger opportunities. Some people might call them new bubbles, horizon bubbles. Look at NVIDIA stock price. NVIDIA at the top of crypto, I was like, this is crazy. Insanely overpriced. Thereâs no fucking way NVIDIA goes any higher than this, and itâs like 20X higher than the top of the crypto bubble. People are using the NVIDIA chips to mine crypto, and with VC, it was ZIRP. It was low interest rates that were just fueling higher asset prices. We had the COVID to work from home. All the SaaS companies were overvalued. We had kind of the Web3 stuff. Itâs like everything is going to be on the blockchain, the economy is like $100 trillion.
Fintech is a massive market, itâs all gonna be on chain. And then with AI, itâs like labor. Literally the entire economy now is software. So we figured out, even though the Web3 thing popped and is gone, itâs like we somehow figured out how to find an even bigger thing that venture could eat up and itâs like, what else is there? Itâs like weâre interplanetary now. The number of earths and the number of economies is 10X bigger and theyâre all software, so weâre going to be colonizing Mars and the moon or something. The animals start using software. So your dog needs a like a CRM or whatever. I donât know. Obviously, I think one way is biology or something, new things that have not been impacted by software yet, like healthcare. Itâs probably an interesting example where they have not fully utilized the internet or something like that. So I donât know, itâs again, you could say, will AI, the impacts of it continue to be felt in 20 years? Probably. I mean, thereâs still markets now that are changing from cloud, changing from the internet.
Dan Gray:
Itâs also kind of a question about LLMs particularly. You could look at it a couple of ways. You could say using the example of NVIDIA, this doesnât resemble historical bubbles because the videoâs revenue has actually gone up so much. Whereas in the past, look at Broadcom and the dot com boom, the separation of the two is much greater, but then thereâs other signals which do feel very bubbly. The narrative around what an LLM can do has gone from in 2023 it was like, weâre on the verge of AGI. Itâs coming, everyoneâs going to be out of a job. And now itâs like Sam Altman saying the biggest risk is that maybe we use too many M dashes in our writing. Itâs calmed down so much and why? Has the steam gone out of it that much? How much of the capital thatâs gone into it was predicated on AGI and is that now not happening?
Turner Novak:
Yeah, I think itâs interesting from a technological perspective of the cost is coming down and also the effectiveness is going up. So when you just say, like the internet, I donât know what you used to pay to access the internet, like a AOL dial-up. It was maybe like 40 bucks a month, 60 bucks a month or whatever. It kind of sucked, but it was good at the time. Itâs gotten better, but itâs not like we pay less. I need to actually go and call Comcast. I paid $230 for my Comcast internet last month for my office, which is way higher than what I initially signed up for. Totally different tangent and Iâm like, âFuck.â Well, I hate how they always increase the price on you. Like, âOh, itâs 80 bucks for an intro deal,â and they increase the price, but they keep increasing the prices.
Or on the other hand, you think of something like the cloud or whatever, itâs like you connect to the cloud, the internet is on and you didnât get better cloud, maybe the price came down, they allowed new services where you keep paying higher prices but you donât get better cloud. Itâs like, âOh, itâs like a better optimized cloud.â Maybe marginally, but I think the interesting thing with AI is over the past couple of years itâs gotten way better and the cost to service it has come down magnitudes every, I donât know, nine months or I donât know, every couple months. So itâs one of those things thatâs getting way better and also the cost is coming way down. I think you can argue that if you want to keep using the bleeding edge models, the pricing is more expensive, but theyâre still getting way better. And there are cases where you donât have to use a newest model.
So itâs kind of one of those technologies where itâs getting better and also cheaper, which I donât know if itâs happened to that extent ever. Maybe if you go back to the Industrial Revolution or something like that with... I think some people made that argument of if you think that LLMs are automating manual software labor, like computer labor, the Industrial Revolution was automating physical labor and it was making it cheaper and faster. So that had a pretty profound impact on the world. And so I think thatâs maybe an interesting parallel to think about it with AI. I donât know, itâs probably not totally correct, but thatâs one framing I would use.
Dan Gray:
Thereâs the incremental improvements as costs come down and it becomes more accessible and more useful, which I think is definitely true. Thereâs also the productivity shocks, so when AI passes a certain threshold to really do something very well for the first time. So I think the recent MIT paper that looked at adoption in the enterprise was kind of showed that the pilots generally werenât working out, like 5% of pilots were successful. And I think thatâs essentially a consequence of itâs not quite reliable enough yet, hallucinations are still a problem.
On the other hand, if you look at Googleâs recent image model, Nano-Banana, fitting for this podcast, itâs like, itâs so good and so reliable that if youâre a fashion company you can actually use it to do AI photoshoots. If youâre an little independent fashion company, you take a picture of the outfit that youâve made, you describe a setting and a model and it gives you a picture that is so close to real, itâs actually usable. And thatâs a fairly small but meaningful productivity shock, like a real jump in value that it creates. The more of those accumulate, obviously the more important AI becomes. Some of them are going to be huge probably.
Turner Novak:
Yeah, thatâs a good point. Thinking about how do you quantify that productivity.
Dan Gray:
Yeah, I guess itâs cost replaced essentially.
Turner Novak:
Yeah. One other thing, when talking about the mega funds and the way that they fit, itâs not like an LP whoâs putting capital in these things is saying, âIâm going to take my origination dollars, stick them in this 5 billion vehicle.â Itâs usually like, youâre probably taking some public market exposure and saying, âI actually think Iâm getting higher returns.â So I think a common kind of dunk on these funds is like, âOh, youâre only going to get a three X, only going to get a four X.â Could actually be higher, could actually be shifting.
That shift into that three or four X might actually be higher returns than they were getting than what they shifted it from. Like shifting that allocation might actually improve their overall expected returns. Because you could say that, letâs just say A16z for example, weâre going to invest in automation and software in the internet and AI in factories and physical things and that might actually be a better investment than an industrial bio private equity firm or something. I donât know if thatâs true, but you could probably make that argument. Iâm sure thereâs a ton of data to support that of you might get a two x on that bio fund, but you get a three x on the mega fund software, industrial bio thing. I donât know, itâs probably a bad argument, but Iâm sure... That is the way, Iâm sure if thereâs more data you could support that if I knew what I was actually talking about. Iâm sure. Yeah.
Dan Gray:
Not only are you right about that, but I think you kind of make a deeper point, which is that if for all that those dollars are currently allocated in the VC bucket for these LPs, if itâs not actually competing with actual capital should go to VCs traditionally, and actually itâs probably displacing PD money. Maybe thatâs because the strategy is really PE and it be grouped in with PE. Itâs kind of a, not something that anyone at a big fund would want to hear that theyâre basically private equity, but really maybe they are.
Turner Novak:
Yeah. Well and itâs interesting, one of the prior guests of the show, actually two guests it I think one of my only, I might actually be my only episode with two guests. It was the founders of a company called Solugen. Have you ever come across that one?
Dan Gray:
I donât think so.
Turner Novak:
Theyâre basically just making chemicals in a different way. Theyâre making chemicals. Theyâre all the same compounds that go into petroleum waste and petroleum type products, theyâre creating it from plants essentially. And the chemicals industry, I think itâs 4 trillion of just headline kind of GDP economic activity in the world that is chemicals. So you use chemicals on clothing, in healthcare products, in your computer, in cars. Everything is coated with chemicals in some way and the entire industry is very old and itâs very asset heavy, high bearish entry, basically everything... The whole industry is revolving around getting extremely high volumes on really low margins. Itâs been competed away and super outdated and thereâs no real room to innovate and anything and change anything. And itâs kind of caused, you could probably argue weâre literally taking oil thatâs on my clothing, itâs touching my skin. Thatâs not good.
Iâm not dead because of it, but if you think about pouring gasoline on yourself or being exposed to nuclear radiation on the extreme end, thatâs not good. And itâs kind of interesting. Theyâre basically a venture backed company thatâs kind of competing in the chemicals industry against these people that have 10% gross margins and 2% profit margins that has bad customer service. Itâs actually interesting, when I was talking to them, I didnât believe them. Theyâre like, âWhen you order chemicals, they give you a four-week delivery window.â Imagine if I sent you a calendar invite for this and itâs like a four-week window of like, âHey, show up at some time in these four weeks and record this podcast with me.â
Youâre going to be like, âHow the fuck do I get anything done?â Thatâs so inefficient. But thatâs kind of how that industry works where you donât know when this thing can show up. And so you could say that using that kind of operational rigor, that kind of technology to say, âWeâll deliver this thing at this time and youâll get your product and itâll be made differently. Our margins are way higher. We make things in a way thatâs way faster and more reliable. The customer service is better.â You can actually compete with those industrial bio type private equity type model, that form of capital. So I donât know, I think itâs net good to your point, more forms of capital for entrepreneurs is good. Generally speaking.
Dan Gray:
For sure, thereâs a ton of big industries with high idiot index scores still, but are ripe for the disruption by that kind of capital.
Turner Novak:
So to your point about bringing capital forward, itâs like you take... That is a very late stage capitalism type of capital, late stage economic activity that youâre shifting the capital into an earlier section of the stack. Itâs not origination necessarily, but itâs still earlier than what it was before. So youâre in theory creating more economic profit of some kind, which I hope is good for the world. Generally speaking, I think thatâs generally a good thing. So you commented one time about A16z is kind of like a common punching bag of the mega fund model. You actually said their very first fund had a smart good strategy. What was the strategy in that first fund? Iâm kind of familiar with it, but I think people would probably be kind of interested because itâs probably different than what youâd think.
Dan Gray:
Yeah, itâs interesting. This is kind of the moment when I came across this when I started to realize maybe Iâm judging them too harshly because I had this sole perception of them as the mega fund, that archetype. But actually if you go back to the fund one, they were a traditional and very well architected VC firm. They were targeting a level of diversification in that fund that was uncommon, is still uncommon but I think is sensible. And they had a follow on strategy for subsequent rounds that was more sophisticated than many VCs would consider today. So on the whole it was like, âOh, these guys actually really do understand VC, so whatever theyâre doing today is an evolution.â
Turner Novak:
So what was the strategy?
Dan Gray:
If I remember correctly, I think they were targeting something like 60 initial investments in their first fund, whereas the typical today is probably like 20, 25.
Turner Novak:
Maybe 30. I guess it depends on the stage youâre at. Yeah.
Dan Gray:
Yeah, for sure. For sure. Could be a size as 30 definitely.
Turner Novak:
So they were more diversified on average.
Dan Gray:
Which means they were able to take a bit more risk. They were able to be a bit more idiosyncratic, choose companies that are more outliers, less consensus, which I think is good. This is all good stuff. And then they had the strategy for investing a little bit in the A and then the B maybe, being able to double down on winners which has been described... Thereâs a very good paper by Joe Millam called Process Alpha and he describes generating systematic alpha through how you allocate capital. And this is one of the things he talks about a lot, like, how you do reserves is actually quite important and they had a very good grasp of this in 2009 or oh eight I think maybe.
Turner Novak:
What is the general best practice in around reserves? Because I feel like thereâs a lot of people that say every end of the spectrum. What have you seen that youâve picked up on?
Dan Gray:
Most recently Iâve seen a few GPs make comments along the lines of they donât do reserves because the subsequent round in that company is so overvalued that they wouldnât want to, itâs not worth it for them for the risk, which is a crazy thing to say because is that increase in value not reflecting an increase in potential?
Turner Novak:
Yeah, youâre saying my performance numbers are fake and theyâre not real.
Dan Gray:
Exactly. Iâll take the markups, thank you very much. But I donât want to be involved.
Turner Novak:
Yeah, Iâm not selling any secondary either, which is maybe then also a question is like, âWhat?â
Dan Gray:
For sure. But I think the biggest thing, and this seems very obvious to say that in any subsequent round of a portfolio company, you need to be able to separate yourself emotionally, completely, look at it objectively. âIs this a good investment? If I had never met the founders before, if I had no track record with the company, if Iâd never posted about them, never written a Substack about them, is this a good investment?â You have to be able to do that to make good judgments. Iâm going to keep doing this to you, I can send you links afterwards. So it was a very good paper that came out last year.
Turner Novak:
Do you remember what it was called? Iâm taking notes right now. So weâre going to throw them all in the description.
Dan Gray:
It is called the Sunk Cost Fallacy in Venture Capital Staging. And it basically describes how VCs grow attached to portfolio companies and probably they grow more attached to the ones that struggle a bit, because theyâre the ones that need the most help and they donât want to admit theyâre wrong in the investment. So ultimately they pile in capital when they probably shouldnât and it goes badly. So for all that it seems obvious to say you should evaluate each investment independently as an objective judgment. It doesnât seem to be what is typically the case.
Turner Novak:
Interesting. Okay. And one comment, I think we had sort of a public... I think it was a friendly disagreement hopefully on portfolio construction. I think thereâs been some studies that show in theory optimal portfolio construction and venture is way more diversified probably than most people typically do. I think my stance on it is the point of venture is getting as high returns as possible. Thatâs the only point of being in this asset class. So that probably means you need a more concentrated portfolio than not. And I think your view on itâs a little bit different. Whatâs kind of your view? And I think this is what the papers show in theory too.
Dan Gray:
Yeah, the question is essentially, what do your LPs want? Do they want that you get an eight X fund and then your next is a three X and then your next is a 1.5 and then your next is a 10 X? Or do they want four X, four X, four X, four X? Which is preferential? So that is going to depend a bit on the VC. Sorry, on the LP of course. But I would say generally LPs would be happier if the GPs they invested in had good public market beating stable returns. I think thatâs pretty fair. If you accept thatâs the case, then more diversification is almost a given because itâs how you achieve that slightly narrow band of returns, that slightly narrower band of returns, but still above benchmarks, still good solid VC returns.
Turner Novak:
Does that beat the public markets, you think? A four X? Like guess it depends on the time to liquidity on that, but...
Dan Gray:
Yeah, it depends a bit on the era, but even then itâs like it gets very complicated with the maths because... And the theory. Itâs not that it makes a four X more likely, itâs that it makes four X being your minimum more likely. You can still outperform it if you are well diversified and you pick well, you can still do a 10 X fund in theory or it starts to get difficult, but you can still get great returns. Itâs more how it limits the chance of complete failure. And if you look at the number of VC firms that completely fail, that should seem to be a priority to me.
Turner Novak:
Yeah, I think itâs a little bit of too high of a number. The numbers that completely fail. I think the thing that I honestly, I hate about that chart from that study. It literally, you cannot get above a six X. I donât know where the cutoff is, but thereâs zero samples, zero of N that hit anywhere above six X. Iâm like, six X is pretty good for venture, but you want a shot at a 10 X, 20 X, 30 X, and if that chart should be... I think the chart that encapsulates that study should look differently. I think that chart biases my head to be like, âIf thereâs no chance of getting a 10 X, I donât think thatâs a good venture strategy.â But to your point of, yeah, if youâre an LP, you want a relationship with a manager who will give you a very consistent strategy over a long period of time. So yeah, thereâs so much nuance on all this stuff.
Dan Gray:
Thereâs a lot. Yeah, thereâs a ton. So the chart itself, just to touch on that, thatâs based on data that looks at what is the typical distribution of returns in VC. So itâs kind of looking at an average profile of returns and how do you use strategy to maximize that. So for sure, like I said, you can definitely outperform that. You can do way better than that. Itâs just trying to understand based on strategy, where is the average return likely to be? How does it influence the rate of complete failure versus the rate of success? So itâs almost a distraction from the point because it leads people to make a lot of incorrect conclusions. The other two little things I think are worth adding onto that is research shows generally speaking, a more diversified VC is more likely to take or is more comfortable taking more risk on a investment basis.
So theyâll go out and back things that might seem crazy to other people more often. And a lot of the time I think that can help accommodate more real outliers in the portfolio. The other thing is research on how... This idea of you should have super conviction, you should be concentrated in your winners and thatâs how you deliver outlier returns is true in theory, but research shows that overconfidence is more of a negative than a positive more or less. Managers tend to have a better perception of their own ability than is realistic. And when they therefore concentrate, theyâre more likely to do themselves harm than benefit essentially.
Turner Novak:
What else do people tend to get wrong about portfolio construction? Is there anything else? Because portfolio construction is basically risk management. Itâs like how do you control... Itâs like the thing you canât control. You canât control how well a company is going to do ultimately at the end of the day as an investor. Itâs on the team, itâs on the founders. You can have a 0.1% impact, but you can control your own portfolio construction and risk management. So what is another thing people tend to get wrong?
Dan Gray:
Probably the two weâve covered I think are the biggest ones. One is diversification and the other is reserves and follow-ons. Outside of that, I mean probably thereâs something to how VCs think about diversification, not just in number of portfolio companies, but in exposure to particular themes. Because if youâre entirely in crypto and then the bottom falls out of crypto, your whole fund is toast. But if youâre a generalist and youâre a little bit of crypto, a little bit of consumer, a little bit of enterprise SaaS...
Turner Novak:
Got some biotech, whatever, deep tech space.
Dan Gray:
For sure. Youâre more robust basically.
Turner Novak:
Have you seen those studies that specialist versus generalist? Which outperform? Whatâs the data show on those? I feel like Iâve seen some data, but Iâm wondering what your opinion is.
Dan Gray:
Thereâs been a few studies, some of them... This is an interesting question that gets into good data versus bad data, which is I know something we were going to talk about. This question of specialist versus generalist is one of those questions that has been in VC for decades and decades, similar to the question of persistence, does VC performance last over time? So both of these questions have a ton of data, a ton of studies. I tend to bias towards looking at the more recent data because to the extent that VC has changed a little bit, itâs probably worth looking at the most recent findings and maybe theyâre the most indicative.
So on the question of specialist versus generalist, the most recent research that I know of was PitchBook, I think a year or two ago. And they found basically broadly generalists outperform for basically the reason we just described. They can cover a ton of different themes. Theyâre not tied to a particular theme through that LP agreement. So if something really interesting happens in year two of deployment or year three, they can get into it. The exception is fields where thereâs real hard science, like life sciences.
Turner Novak:
Some kind of expertise thatâs needed.
Dan Gray:
If you need real deep scientific expertise to understand a company in a field at all, then yeah-
Turner Novak:
Itâs like a blood testing startup, itâs like, âDoes this blood testing actually work? Yes or no?â
Dan Gray:
Well, you say that, but that was diligent out by Google Ventures in a simple little test they did. But yeah, itâs a good point all the same.
Turner Novak:
I think another lens Iâve seen on that with, Iâm piling onto your, because you just supported my narrative of being more of a generalist investor, but it seems like specialist funds will outperform early in a sector or an area of specialization, and then that kind of gets competed out. So an example might be Crypto Web3, letâs say it was a crypto fund in 2019, 2020, youâre deploying, itâs less consensus. But then if youâre still doing crypto only in 2022 and all your entry prices are when that is a consensus thing, you kind of match the generalist returns, you no longer get that sort of out-of-favor return. So I think specialist funds make more sense in a sector or in an area of specialization that are not within the realm of a generalist necessarily.
Dan Gray:
Yeah, it kind of depends on, you have to be specialized in an area that has amazing and pretty immediate potential that few other firms have realized. And if you can make those judgments, youâre in the right professions.
Turner Novak:
But then itâs like, âOkay, that gets competed out and you donât have an advantage anymore as a specialist.â So you need to have some kind of truly defensible institutional advantage within that area of specialization that continues to give you an attractive entry price and reason to win against maybe the more generalists. So I donât know. I think it can work both ways, but...
Dan Gray:
I think to be honest, almost with the exception of the life sciences stuff, just be a generalist. And if you have a thesis on the side, âWeâre a generalist fund, but at this point in history, we think thereâs an outsized opportunity in biotech,â then you can just start investing a little bit more in biotech. You donât need to build a whole fund around it. And I think that is from a basic perspective of understanding outliers and venture capital and recognizing that the greatest companies are generally not from well-understood sectors, it just kind of makes sense to me that thatâs how you should be.
Turner Novak:
Yeah, I think my friend Rex Woodbury had an interesting post about this, and it was kind of a chart. Youâve maybe seen this table, it was like years and then it was most valuable company founded in that year, and then it was hottest sector at the time during that year, or maybe some of those rows and columns are switched, but it is basically just showing in every year the most valuable company that was started was not in whatever the current thing was. And so I donât think you should just say like, âOh, AI is super hot. I am not investing in AI just because of that.â But itâs like data that shows, it kind of goes back to that point of if you want to actually generate returns, you need to make sure entry prices are attractive, but then itâs like velocity of dollars deployed. How much capital could you move? You probably should just invest in the stuff thatâs moving the fastest, even if itâs overvalued.
Dan Gray:
It depends. Yeah. Are you an investor or a trader?
Turner Novak:
Yeah.
Dan Gray:
Thatâs the question. Are you chasing ultimate outcomes or the proxy metrics that raise another fund? Thatâs the core of all of this really.
Turner Novak:
Okay. So maybe moving to a slightly different topic, but what do you think is the correct way to value a start-up? And you do a lot of it through Equidam. How should I think about valuing a startup?
Dan Gray:
Thatâs a deep question. So going back to what we were talking about towards the beginning, a lot of VC users relative value as a way to understand how to price a company or how promising an investment is. And it has all those issues we discussed around catering and it channels capital towards areas of consensus, letâs say. So the flip side to that obviously is if you are in the camp of what I would describe as real VCs during origination, you have to have a lens on what is the fundamental value of a company. And that means you have to understand essentially what is the future cash generating potential of the company, which is obviously a ridiculously difficult question to ask it pre-seed.
Turner Novak:
I mean, itâs one of the things, you canât answer it. You canât say, âHey, thereâs these two people. Whatâs the free cash flow going to look like in year 12?â
Dan Gray:
Yeah, itâs true, but you also can answer it because... This gets really deep into the theory. But essentially you have to understand a few different things as the foundation, one of which is valuation is always based on the future. Itâs like a hurdle for future performance. Itâs not a reward for past performance. The other one is itâs always an opinion. There is no calculation in the world that tells you the real value of a company. Itâs always an opinion.
Turner Novak:
When you look at public markets, it changes every day. Every stock price changes pretty much every day. The price changes of every stock.
Dan Gray:
Based on what do people believe basically.
Turner Novak:
Yeah.
Dan Gray:
So then the question becomes, you look at a company at pre-seed, you can see the market theyâre going after. You can see the product. You can maybe see what existing solution are they displacing, what budgets does that allow this company to tap into? You can think about how much competition is there in the market, how much better is this solution than incumbents? And that tells you a little bit about what their moats might be, which allowed them to preserve the margins over time. And basically, you look at all of this and you can make a more or less a gut judgment of does this company... The way I always put it is, can it unlock a ton of economic energy? Can it tap into something that has a huge amount of cash waiting to be deployed into a better solution or a different way to do things? And is that going to sustain over time? Are they going to be able to keep an edge over the market to let them to create a monopoly more or less? And I think you can make those judgments a pre-seed. And not only can you, but thatâs kind of the job. If you care about the ultimate exit and not the incremental metrics to go back to that point again, thatâs kind of the job Iâd say.
Turner Novak:
Yeah, thatâs almost like the role of the founders is to figure out the execution of getting there. And you can help maybe. Most good investors probably help in some capacity, but really just gauging, is this even possible? Is this an opportunity that I want to allocate my capital to, I guess with my crystal ball in a way? I like that framing of market potential. I think a lot about just customers. Do I think that you can generate a lot of cash flow from these customers? Do I like a company servicing this customer base?
I have one them looking at right now, really interesting idea, but Iâm just like, âI just donât think you can make enough money on this. I just donât think this is a customer that I want to go after.â And itâs just tricky because they might be able to actually expand the scope of what it is, but I just donât think they will be able to. I donât know. But I think a lot of it, like customer base, do I like going after this customer and do I think you can generate a lot of economic profit serving this customer or serving this market to your point?
Dan Gray:
And every VC, or letâs say maybe like 99% of VCs, probably 99.9 would say that like a DCF, a discounted cash flow model evaluation where you model out the future revenue, the future costs, you get to the future free cash flow and you use that to determine the value today, they would all say that has no place in venture capital. Thereâs way too much uncertainty. Itâs not practical. You canât do that for a pre-revenue company. Itâs ridiculous. But actually what you do when you think about whether or not to invest in the company is basically like a simple mental DCF. The only benefit of putting it out in a spreadsheet if you need to is you can pick apart the assumptions. Are you really going to grow at this rate? Is this really going to be your margin? And itâs not to get to a precise calculation, but to explore what are the assumptions and the opinion.
Turner Novak:
My general super lazy just gut check initially is just you just assume somethingâs 10X revenue is your exit multiple or whatever. Sometimes itâs lower than that. Sometimes itâs higher, but itâs like, okay, if itâs a tech company growing at a relatively fast-paced, outsized margins like 10X revenue at exit is probably semi-reasonable. Hopefully itâs higher. I think you want that uncertainty of, ooh, 22X revenue multiple, whatever. But I feel you start to get in trouble though, where you go in and you say... Letâs say you invested at 100X revenue multiple and youâre going to exit at 10, you need to grow 10X to then keep your valuation flat at a 10X multiple. So I like to think, okay, if Iâm only investing at a 50X, that cut in half the return expectations. You only have to grow 5X, or if you come in at a 20X, you only have to double to hit that same exit valuation.
And if you 10X, then youâve got a 5X increase in valuation at the exit if all things hold true. To your point about almost re-risking the business, I think about that a little bit when Iâm just generally thinking about this is how high is that hurdle that Iâm coming in at that the founders are giving themselves? And then in the context of the full portfolio, a basket of all of these, how is it going to all culminate when it all gets said and done in five to 15 years when some of the stuff really starts to play out?
Dan Gray:
And a big part of that is the multiple drops over time as the growth of the company slows, as it matures for sure. But also, as you go from pre-seed to a public market company, the quality of your revenue matters a lot more. And that, if youâre investing on an ARR multiple basis, itâs the point I made earlier about GAAP accounting and adjusting to that life. For how long in a companyâs life can it be valued on ARR and invested on that basis and still be able to do a hard handbrake turn at the end have become a reasonably healthy company in a public market or in an acquisition scenario because those buyers do care. It obviously created a ton of problems with SaaS companies in â22 because they could not turn around that quick.
Turner Novak:
Yeah. One thing I think too, when you look at just generally public market, I feel like you need to use public markets for this because thereâs not private market data on what drives returns, but I think generally in public markets youâll see them, I forget the numbers, but the majority of returns are basically driven from multiple expansion and earnings growth. If you say, in venture youâre typically fighting against multiple expansion, youâre probably going to get multiple compression.
But the holy grail is can you actually get multiple expansion on top of earnings growth? Thatâs maybe the, when you look at a hundred baggers in the public markets, that typically happens. Itâs like you bought a company for 4X earnings and now it trades at 20X earnings and it grew a 100X, and now itâs become this compounding durable sustainable growth. And the market knows, and thatâs why it trades at 20X earnings, but it was trading at 4X before and you bought it at 20 million in earnings, so you bought it at $100 million market cap and now itâs worth 6 billion because you expand your multiple X amount and you also grew earnings by X amount.
I talked to a lot of my public market friends and be like, âOh, this venture is, itâs micro cap penny stock investing almost.â Where yeah, if you go look at a public company thatâs worth $12 million, itâs probably worth zero. But youâre trying to find that diamond in the rough of this is a 300 banker, or this is an 8 billion company that you invested in it, 12 million market cap or something like that, and itâs like treasure hunting. Youâre just going through, youâre trying to find undercover gems really at the end of the day, finding founders. But itâs cool because in venture, they come to you. If youâre a public market investor, youâre probably not getting these companies emailing you, and if they are, itâs probably pretty bad, right? But in adventure itâs like theyâre coming to you to seek out capital. So itâs like this weird treasure hunting.
Dan Gray:
Well, that again, it depends on your strategy, the whole origination question. Are you relying on inbound? How good is your process for screening inbound? Are you relying on referred deals? Are you going out hitting college campuses and hackathons? Itâs a pretty deep question.
Turner Novak:
Thereâs this one study that you brought up that I think you posted about this on Twitter recently, which sounds non-intuitive, but generally speaking, investment firms that are located in popular startup hubs typically outperform the investment firms that are outside of popular startup hubs, but not necessarily on a company by company basis, just generally company based outside of San Francisco might actually outperform in San Francisco. How is that even possible? How does that play out and why is that in the data?
Dan Gray:
Yeah, it is a fascinating bit of research, and the bottom line that I try and that I hope VCs take away from that is that they should be looking at opportunities in non-traditional hubs because theyâre great and theyâre out there, but essentially I think itâs driven by the friction for a Silicon Valley VC to go and invest in a company in Ohio is very high so theyâre only going to make the effort if they think itâs really an outstanding opportunity. And also, at the same time, because theyâre a Silicon Valley VC, maybe with a big name, they can win that deal over any local investor. And also probably, they have an easier time raising from LPs than the Ohio VC does so they have basically a lower cost of capital, they can invest on more favorable terms. They basically have all the advantages and the ability to pick opportunistically, whereas the local VC is a bit more limited
Turner Novak:
Because I feel like thatâs generally what I see. If youâre the best company in Cincinnati, Ohio, youâre going to want to raise money from the best investors. Youâre going to be seeking out though you want a16z involved, you want their great brand. They will talk about you on podcasts, youâll get the news mentions, youâll get Twitter threads about you. Youâll show up in all the reports. People will know that you exist and youâre not going to get that from, I donât know, Midwest Ohio venture capital fund, but I donât even know where, I think Cincinnatiâs in the South. Have you ever been to Ohio? Because you live in Wales, right?
Dan Gray:
Yes, I do, and Iâve never been to Ohio.
Turner Novak:
Okay. Have you ever been to the US?
Dan Gray:
Yeah, I spent about six months living in Miami and I visited other parts, mostly West Coast.
Turner Novak:
Oh, interesting. Okay. Because I think in that one paper specifically, it talks about how proximity to portfolio company has no correlation to return to. Maybe thereâs a negative correlation, I forgot which one it was, but itâs not intuitive because you think, oh, if you invest in the thing down the street, you can monitor it and check in and add value way better than if itâs 420 hour plane right away.
Dan Gray:
Yeah, I guess whatever that adds in terms of friction is priced in so that you therefore would only do a deal if it was really seen as worth it.
Turner Novak:
Interesting, okay. And thereâs another study that shows actually that itâs related about founding team and the quality of the founders is not actually a predictor of investment success. Itâs much more about the actual initial quality of the business, but a lot of times thereâs no business, I guess were theyâre investing in a team. So what is that study that shows the actual business quality predicts, and then how does that actually play out in this origination stage venture when thereâs not really a business yet?
Dan Gray:
Thereâs three papers I can think of that have all found basically the same thing, I think that the best of which is called Predictably Bad Investments in Venture Capital. They basically all look at what are the founder attributes that VCs pattern match for? And itâs stuff like where they went to school, where they worked before, maybe socioeconomic background or they go to the same squash club or something. A lot of very weird, slightly tangential attributes that end up influencing VC behavior and investment decisions, but probably shouldnât in theory. And when you look at the data, clearly shouldnât, have quite a negative influence. Not only do they lead to bad investments, but they lead to missed good investments, which is a huge problem. There was a second part to your question, I canât remember what it was now.
Turner Novak:
Well, that actually the quality of the business is what actually predicts the outcome, not the founders.
Dan Gray:
Yeah, and how you understand that at a very early stage.
Turner Novak:
But yeah, if youâre investing in, thereâs no customers yet, how do you get business quality? Is it potential business quality or something or?
Dan Gray:
Thereâs a quote from Peter Thiel, he was asked a very similar question, do you focus on founders entirely? What else can you look at early stage? And his answer was itâs a complicated package. You have to look at the whole thing. So in order, obviously pre-seed, pre-product, all you can really look at is the founders. But you can understand the quality of the founders and the competence through the lens of how they articulate the idea, how they think about the markets. Do they understand the financial side of things? Do they understand the growth, technical aspects? Itâs very easy for a lot of early stage VCs to say, âAll we care about is the best founders. We donât care about ideas or we donât care about markets.â But actually, that just means youâre ignoring information that could be useful, even if itâs the wrong market, even if they end up pivoting the idea, the way they talk about it and understand it and articulate it is still really useful information.
Turner Novak:
And itâs like how did you come to your idea? What kind of research did you do? What was your process like? I have one portfolio company that when I invested, they had one idea that I thought was really good. They ended up pivoting. And the reason I invested was I thought, I bet if they change concepts, they have a good framework for doing this. I trust them with the capital, I trust that theyâll be able to do something. And they just pinpointed this almost perfect storm, perfect category. I mean, itâs turning out to be a really interesting business, still pretty early, but itâs one of those things Iâm like, âDamn, that was incredible pick. How the fuck did they find this?â Itâs also right in front of you kind of a thing.
But yeah, I feel like thatâs definitely a good way to think about, itâs just the founder ideation, quality discovery, quality understanding of what it takes to go into almost like Porterâs five forces or the seven powers type equation of when youâre starting a company, what are the ideas to pursue related to margin structure and dynamics in an industry and how those could change over time based on new technology.
Dan Gray:
Yeah, and you have to reconcile that with the fact that other research Iâve shared shows that high-tech startups are actually more likely to succeed after theyâve had at least one pivot, or no, after theyâve had one pivot. It goes down a bit after that, but one pivot is actually a positive, which is crazy because it means everything you think about the business at the first pitch meeting is probably going to be wrong if itâs a good company.
Turner Novak:
What is considered a pivot in that time? Is it like we are a consumer social media app and now weâre a B2B AI, vertical SaaS thing? How do you define a pivot in that one good pivot thing?
Dan Gray:
Itâs a good question. I canât remember how the paper described it off the top of my head, but it was a pretty fundamental shift in the direction of the business so it was significant.
Turner Novak:
Yeah, because when I think about all, every single, probably half of my investments I made are just pre-customer investments sometimes. Sometimes theyâre working on getting customers or in the discovery phase or whatever, but pretty much every single one of those, the thing that was in the pitch deck is the main product that was going to be used to generate the business revenue is not true anymore. Thereâs a lot of cases where generally the customer space or the general idea, but the way they actually solved it just completely different than the pitch deck. The pitch deck was just wrong, but itâs this discovery process of meeting the team and youâre almost sussing out their ability to go execute on this and figure it out really, because itâs an experiment, itâs an adventure. Youâre figuring this all out.
Dan Gray:
And that ties into a few things weâve talked about. Also, maybe another paper I would recommend, which is Premature Scaling by Startup Genome. So this ties together a few of the topics quite nicely. They looked at, basically they were trying to study what are the main causes of failure in startups, and they were surprised by their own finding, which was that I think in 70% of the failures they looked at, a significant problem of the company was that theyâd raised too much money. And essentially what that meant was before the startup had properly validated what they were working on, that they had product market fit and it was the right direction for the company, they raised a ton of money and invested a ton of money in growth and in hiring and technology and development, and then when it was wrong, the company just explodes.
Whereas if you go piecemeal step-by-step, get the very small checks from a nice, small friendly originating VC fund, you can work that stuff out as you go and you have a partner to help you through that. The parallel there on the other side is maybe the mega fund VC who would give you the huge bag of cash in the beginning and then just stand back and maybe you take off and itâs amazing, but quite possibly you explode. And maybe theyâre okay with that trade-off. Maybe thatâs built into the maths of their model.
Turner Novak:
Yeah, I would think so. Did the explosion or the failure happen because you set a super high bar of what the company needs to hit in terms of the valuation of metrics to reach that valuation and youâre just unable to raise more capital and you run out of money, is that generally why the explosion or the implosion happens?
Dan Gray:
Yeah, basically you have a very fragile future. The thing that you pitch them on that they invested a ton of money in, it has to work and it has to scale quickly, and if it doesnât, youâre done because the hurdle for that next round is now so high, you just canât clear it.
Turner Novak:
Interesting. Yeah, and if you just think about a $10 million fund that writes a 100K check, thatâs 1% of their fund. A $100 million fund that writes a million dollar check, 1% of their fund. A billion dollar fund that writes a $10 million check is also 1% of their fund. So in all those cases, if I invest, $10 million fund, 100K, if it doesnât work, it doesnât work, that can drive my return. But I was like, âHey, thatâs okay. You guys tried,â doesnât screw, thereâs a reason I have a fund. Letâs say I wrote 100 checks of 100K each, itâs fine.
For a fund thatâs a billion dollars, wrote $10 million check, didnât work. Youâre like, âEh, thatâs fine. We tried to get to the Series A that we wanted to put 3% of the fund into, didnât work, whatever. Itâs not a big deal.â
So I mean, as a founder, if youâre trying to play that game, I would say youâve got to find more funds that you can find that very small percentage of the fund option check. Theyâre there. I have portfolio companies that have done that, that have been $10 million option check from a big pool of capital to keep it going. Sometimes it works, sometimes it doesnât work, but itâs an option there for founders.
Dan Gray:
The question that it leaves me with is the mega funds, when they start doing origination themselves, how well set up are they for that? Or do they even have slightly different incentives? If theyâre doing this strategy of trying to ride the waves and trying to produce essentially venture beta rather than alpha, do they necessarily care about finding the crazy little outliers? Perhaps thatâs not how theyâre oriented, and that shows in the data where they actually seem to be increasingly concentrated in San Francisco software companies rather than branching out as they get bigger. So yeah, I have lingering questions about that and their involvement at the very early stages.
Turner Novak:
I mean, ultimately their business model is put a bunch of capital into only a couple companies, and you canât put a bunch of capital at the origination stage, you need them to get further along. Youâre probably on a cost weighted basis. Letâs say one of these mega funds fully plays out and you just look at, weâre talking in 2040, weâre looking at what they were doing in 2025. I mean, their average entry point on dollars is probably a Series B or maybe even a Series C because itâs 5 million at the seed, 10 million at the A, 2050. They might put 100 million in right before an IPO. So your average cost basis was almost at the IPO, really at the end of the day. But you basically use that origination check as a right to get there, to get into that company.
Dan Gray:
And actually, you want it to whittle down. You want to start with whatever theyâre doing, 100 seed checks a year, and you want that to be two companies by the end to concentrate all the rest of the money into.
Turner Novak:
Yeah, because you want to make as much money as possible while doing as little work as possible. As an investor, you want to give someone want to give them cash a year later, they give you 100X of it back and you did nothing. Probably for an investor that is the most ideal. No one will admit that publicly, but I mean thatâs ultimately what every investorâs going for. We first connected on Twitter a couple years ago, I donât know, I just feel like Iâve been reading your stuff for a while. Weâve talked back and forth. How did you first get started with writing?
Dan Gray:
My blog, which has the very unusual name of credistick.com, which is a Shadowrun reference, unfortunately a misspelled Shadowrun reference for anyone else-
Turner Novak:
Shadowrun, is that a game?
Dan Gray:
Yeah. Itâs originally a tabletop RPG that became video games later on.
Turner Novak:
Yeah, did you play the video game?
Dan Gray:
Yes.
Turner Novak:
Oh, so this is the one where it was a gun game, it was shooting, but there was magic?
Dan Gray:
Kind of, yeah. It was an isometric turn-based setup, but basically itâs a connection to science fiction, which is what I originally intended to start writing about, the intersection of science fiction and technology today. That led into writing about Web3 Metaverse because I was super interested in that. Iâm a big fan of books like Otherland by Tad Williams, Permutation City by Greg Egan, Neuromancer of course, they all have connections to this. And then Web3 turned up and I was like, âOh, this is amazing,â so I started writing about it and I quickly realized the people investing in Web3 didnât have any familiarity with the science fiction, the literature about it with existing virtual worlds. They didnât know about Second Life, they didnât understand World of Warcraft. I need the research on these topics.
So much interesting stuff was out there to be a very useful resource that wasnât being applied to this wave of Metaverse and Web3 companies. And thatâs when my writing took a pretty hard turn into being like, âHey, venture capital is kind of broken.â
Turner Novak:
Okay. What were the name of those three books? Iâm going to throw a link into the show notes for people.
Dan Gray:
The first is a series of four books by Tad Williams called the Otherland Books, the Otherland Tetralogy. The second was Permutation City by Greg Egan, and the third was Neuromancer by William Gibson. And thereâs also, thereâs a ton of others you could mention in that same kind of genre, but those are the three I think about the most when Iâm near future sci-fi that is still relatable. So if youâre a VC and youâre thinking, whatâs the world going to be in 10, 15 years? I think those are good books to read and think about.
Turner Novak:
What are the general takeaways in terms of whatâs the world going to look like? Is it not Web3, everythingâs tokenized? Is it more of virtual economy? Whatâs a way to think about it?
Dan Gray:
Iâd say Neuromancer is interesting because itâs cyberpunk, so itâs dystopian. Itâs like what would happen if the FTC didnât exist and corporations are allowed to do whatever they want?
Turner Novak:
Oh, thatâs basically today, right?
Dan Gray:
Kind of.
Turner Novak:
Almost.
Dan Gray:
You can see the direction travel a little bit.
Turner Novak:
Did you see, well, Iâll throw the picture up on the screen. I think Trump posted a picture, it was an AI generated picture of him looking at a computer that just said, âIntel $30, buy Intel stock,â or something. Did you see this?
Dan Gray:
I didnât, I missed that.
Turner Novak:
Oh man. Iâll send it to you right after this but itâs so funny. I was like, âWhat are we doing? Why are we allowed to do this?â This should not be allowed.
Dan Gray:
Itâs crazy. Itâs crazy, Neuromancer gives you a lens of what the worst end state of that would be. Also with some fun stuff like human augmentation, biohacking sprinkled in. Otherland is really quite literally what the metaverse should or could be like, and this is with brain interfaces, AR, VR, all this future visions of current technologies, which is really interesting and it is not set too far into the future so itâs quite attainable relatively. And then Permutation City I think is more what science could look like in the near future, biology particularly as well.
Turner Novak:
Interesting. Yeah, Iâll throw links to all those in the show notes if people want to check them out.
Dan Gray:
I should start a VC sci-fi book club because I could talk about that endlessly.
Turner Novak:
Yeah, you should. I feel like people appreciate that kind of stuff. Well, any advice for someone who is starting to think about writing online? Iâm assuming you probably did a lot of not online writing before this. Is there a transition that you went through or made or a change in mindset?
Dan Gray:
Itâs maybe tricky. Thereâs something, thereâs a leap you have to make when you go from just posting on X and being in the timeline with hot takes on topics to when you start being like, âOkay, Iâm going to have a blog or a Substack,â and the transition is what I write is no longer really ephemeral. Iâm carving it somewhere.
So you have to be comfortable being like, âThis piece of writing is a part of me or a representation of me, my view on the world in a certain way.â
You have to be willing to be wrong or to look dumb, to have people send you comments or DMs and be like, âHereâs 500 words about why you couldnât have a more idiotic take on this topic,â and maybe theyâre right, maybe theyâre wrong. Maybe it changes your opinion or you end up rewriting something. But that has to be part of the process because the alternative is you do what most VCs do, unfortunately, which is rewrite the same perspective on the market that everybody else is saying that adds zero insight.
Turner Novak:
Valuations are too high, mega funds are bad, whatever, AI is going to change the world.
Dan Gray:
Yeah, why whatever my portfolio is mostly invested in is the best opportunity right now.
Turner Novak:
Yeah, exactly. Thatâs good. You should write about that if youâre a VC, go for it, Iâm not going to stop you. What was the most bold cold email youâve ever written?
Dan Gray:
That would be by quite a wide margin. A few years ago I was working at a media company in Berlin. We were focused on data science, machine learning, more or less what you would call AI today, but people didnât back then, and itâs a media company. Itâs a grind, itâs horrible. Youâre trying to do anything you can to get page views, and I saw Mark Cuban had this new app, so I just sent Mark Cuban an email, and I was like, âHey Mark, can I interview you about this app launch?â
And he replied 30 minutes later and said, âSure.â I had a 0.1% chance in my head that he would even see the email. But it was pretty amazing. It taught me from then on, you got to take those shots because thereâs no cost, thereâs no downside. You might as well, and it can have big payoff.
Turner Novak:
Actually, dude, Iâve heard Mark Cuban responds to a lot of stuff. Iâve heard so many people say, theyâve emailed and he respond. He actually, he once retweeted. A couple months ago, he retweeted a clip from the podcast on the Twitter account, and I was like, âOh, was not expecting Mark Cuban to see this and retweet it.â So I DMed, Iâm just like, âHey, want to come on the podcast sometime?â
Weâre trying to figure it out. I probably got to have to fly to Dallas and deal with him in person, or maybe I can get him to do it virtually. He actually has a surprising amount of podcasts and itâs hilarious because if you look at his background, itâs the messiest office with just shit everywhere. Just YouTube Mark Cuban and just scroll and look at his YouTube Thumbnails on YouTube. You can tell he does not care.
Dan Gray:
Quite often with NBA trophies, if I remember correctly.
Turner Novak:
Thereâs just random stuff, you can tell that itâs just a guyâs office. Itâs a 55-year-old dudeâs or 60-year-old dudeâs office whoâs, heâs just got stuff going on. Thereâs stacks of papers, thereâs snowboards or football, basketballs and trophies, massive stacks of papers and binders and yeah. But anyways. Well, this was a lot of fun. Thanks for coming on the show.
Dan Gray:
I really enjoyed it. I would do this anytime, all the time. Youâd have a hard time getting me to shut up, but I really enjoyed it.
Turner Novak:
Actually, before we sign off, whatâs a good way for people to follow you? Twitter seems to be probably your main channel. You got a blog too, where you write sometimes. What are those?
Dan Gray:
Yeah, pretty much everything on Twitter, to be honest. My goal with all my writing is to put it in the most accessible place so that for the time being seems like Twitter.
Turner Novak:
Yeah. Okay, cool. Yeah, weâll throw links to those in the show notes and people can find them. Well, cool. This is a lot of fun. Thanks again for doing it.
Dan Gray:
Thank you for having me. It was great.
Stream the full episode on YouTube, Spotify, or Apple.
Find transcripts of all other episodes here.