Don't work in venture capital

Separate the “job to be done” from the the job you think you want

I get a lot of people asking me how to get into venture capital. How did I get my job? What’s my story? Who’s hiring? What are funds looking for? Is Tusk hiring? Most weeks, I speak to one or two people looking to work in VC, usually for analyst/associate roles.

Most or all of them seem well intentioned, smart, and eager to contribute. I want to help them but we’re not hiring so the best I can do is offer whatever insight I have.

I tend to give them all some version of the same advice/feedback and thought it might be good to write this down for anyone else looking to work in venture capital.

The short answer is that, under generic terms, you shouldn’t. It’s not a good goal, won’t make you happy, and isn’t the thing you think you’re applying for. Here’s why.

1) It’s very hard to get a job in VC.

Capital scales very efficiently against labor. The number of people it takes to invest $100 million is not meaningfully different than the number of people it takes to invest $500 million. Consequently, venture capital funds don’t need to hire at the same pace as the startups they invest in and they don’t generally hire on any particular cycle (at least not one that will be familiar to bankers and consultants).

The VC hiring process is opaque and takes a long time. The job openings are rarely publicized and the competition is intense. Hiring is usually fairly ad hoc and opportunistic. There might not be a job opening at all until someone meets the right person and decides to hire them without a process.

Even when a process seems more accessible and has an open listing, VC is, for better and for worse, a connections-driven industry and the recruiting is no exception. As an applicant you need to maximize your surface area: constantly scouring the earth to get meetings and intel and contacts until you manage to be in the right place at the right time to even find someone hiring.

Taken together, it can often take 6+ months of work for even the most credentialed/qualified-seeming people to land a job in VC.

But even if you can wait that long and do land a job…

2) Most of the jobs are not good.

From the outside, people tend to think that being an associate at a venture fund is mostly about running up an expense account, tweeting, and going to parties. You get to play real life Shark Tank and build companies hand in hand with founders! But all of that is fairly incidental to actually doing the job. In many if not most firms, associates and analysts are glorified assistants. Their primary job is volume sourcing to get meetings on partners’ calendars and then doing diligence on partners’ deals (read: papering those investment decisions ex post facto). This is especially true of some of the larger growth-equity style funds that might have hiring processes that look more similar to a bank’s (cyclical, multiple openings at once, etc.).

Because of the aforementioned scaling relationship between capital and labor, venture capitals firms don’t need to add new partners very often. Most junior roles tend to be 2-3 year appointments with limited room for upward mobility. Given that you’re not going to be sticking around, what incentive does anyone have to mentor/invest in you? That’s compounded by the economic structure of venture firms, where carried interest (keeping a portion of the profits from investments) is a finite, rivalrous good. The more carry you have, the less I can have.

But even if you can wait a long time, do land a job, that job gives you room to grow, and you manage to wring some carry out of the partners…

3) You won’t make as much money as you think you will.

I’ve said this before but it bears repeating: almost no one makes money in venture capital, at least not off of performance (getting fat on management fees doesn’t count).  Benchmark returns are really pretty shit. The generic venture fund is worse than public markets but with no liquidity.

Venture capital is famous for following power laws whereby a very small fraction of investments (and by extension a small fraction of funds) produce most of the profits for the whole asset class. So if you’re getting carry in a generic/random fund, the returns just aren’t very good.

If it’s the salary you care about, you can make the same or more in banking or consulting, or in a business role at a later stage/pre-IPO company.

“But, Yoni,” you say, “you work in venture. Is your job terrible?”

I am very lucky.

I have a great job that does not conform to most of what I’ve described above. I came into a very unique situation, mostly through sheer luck. My experience has been so predicated on being in the right place at the right time that it’s really not repeatable/replicable. It’s almost not worth going into. Most of my analyst/associate friends in good situations at other firms also got their jobs through similarly non-replicable paths.

And if/when I leave Tusk, it won’t be to go work at another venture fund.

So don’t fall victim to survivorship bias. It’s only the people who are lucky enough to have figured out/landed in a really good situation that are left around to be asked advice.

So what should you do?

Look, if you’ve made it this far, you must have a pretty high tolerance for my sage wisdom. Why should I tone it down now?

Take a step back and ask yourself why you want to work in VC. Then think seriously about where/how else you might be able to scratch that same itch in another job.

Maybe you want to do strategic thinking and problem solving in biz ops, growth, or strategy at a late stage company. Maybe you should spend time as a really analytical thinker in private equity. Maybe you want broad experience juggling many balls in an ops role or as a chief of staff at an early stage company. Maybe you should go to business school for the network/community you’re trying to build (or maybe you should just start tweeting more - seriously). Maybe you love jet-setting around and doing meetings and making DEALS and sales could be the right place for you.

The list goes on and on. And maybe you’re ok with doing VC for a couple years and don’t want/need to have a shot at a partner-track position. That’s fine too, so long as you can be clear-eyed going into it.

No matter what the answer is, you have to separate out the “job to be done” from the the job you think you want.

IF you are going to work in venture, try to optimize for partner/fund rather than generic “venture capital.” As I’ve said, most jobs and funds are bad. The biggest brand name funds are functionally impossible to get hired at and working there is unlikely to be any better than what I’ve described above. I’d suggest going to a new fund where you stand a better shot at having upward mobility, economic upside (carry), and the ability to drive outcomes.

Now I know that this is somewhat of a controversial stance. Smart people that I respect have said it’s better go for a brand name and use that to pivot elsewhere where you can have that better job eventually; it’s easier to move downstream that upstream. My preference/bias is to bet on myself and new funds are extremely high risk/reward. Knowing that most venture funds, like most startups, won’t succeed/produce great outcomes, I’d at least rather have the potential for upside. So you need to have a clear sense of where you stand on these questions and accept the tradeoffs that come either way.

I don’t say any of this to discourage you or make you feel shitty. I’m saying this because I’ve seen enough people go through really long, painful process before they come to see what I’m saying here and I love you.

And if you take my advice and I’m completely wrong about everything, at least you’ll be more sure of what you already knew. A little bit of introspection and self-knowledge never hurt anyone.

Get poor quick schemes

Bitcoin to $50k

Couple of things before I get into this. First, none of this is investment advice. I’m an idiot and you really, really shouldn’t listen to my opinions. Second, if you hold me accountable to any of this then the joke’s really on you, bro.

Now let me spit some facts.

The Federal Reserve estimates that 31% of all the US dollar bills in circulation today are $100 bills. This is a 20% from 15 years ago. (US Fed)

80% of all US $100 bills are held overseas, according to the Chicago Fed Board, up from 15-30% in 1980. (Chicago Fed)

Looking at the total US dollar bill circulation, that means that $100 bills stored overseas accounts for ≈64% of the total value of printed US currency. This represents a hair (few hundred million dollars) over one trillion dollars in value (that’s $1,000,000,000,000 or one thousand billions of dollars) sitting in $100-denominated bills parked overseas.

Rich Dave Chappelle GIF

Why are people keeping hordes of $100 bills overseas? Lots of reasons. I’d guess mostly to do with crime on one hand and local economic instability on the other. $100 bills are obviously the most efficient way to store, transport, and transact in cash that you want to keep out of the regular financial system. $100 bills take up the least space, are universally accepted, and are well made enough to keep their integrity for years.

But crypto-assets are better than physical bills on each and every count.

So that $1 trillion in $100-denominated bills stored overseas? That’s the baseline price target for Bitcoin’s market cap. The max number of Bitcoins that can be mined is 21 million. So assuming none get lost (1/4-1/2 probably have been), that sets a price target for Bitcoin at ≈$51,000 before accounting for any other use cases beyond displacing $100 bills as a store of value overseas.

I think Bitcoin will be very price volatile but generally trend upwards until reaching something within spitting distance of that price point and then become relatively price stable for the future. And I think it’ll be Bitcoin specifically (rather than say Monero) because it has the biggest headstart and most “brand recognition” among all crypto-assets, not unlike the US dollar relative to other fiat currencies. To put it in perspective, Bitcoin makes up about 65% of the total crypto-asset market cap, depending on the day. 

I’m gonna be buying and HODLing all the way to $51k.

My DMs are open for any hedge funds that want me to come in and run things from now on.

You’re welcome.

Dunking on VCs

Understanding and enjoying the meme

Justin Gage has has an ongoing twitter thread exploring one question he keeps coming back to, month after month and IPO after IPO: Why do people take such particular pleasure in making fun of and/or criticizing venture investors?

Justin is right. VCs do seem to attract a particular kind of disdain. At the very least mocking them (us?) has become a whole category of memes. No, we’re not an endangered species and yes we’re all fine. It’s a small violin that plays for the woes of the venture capitalist.

Here’s what I think is going on.

First off, I think there’s a perception that venture capital is a house-always-wins scenario. Place enough bets and you make money even if most or nearly all founders get screwed, right?  So it makes sense that people who make fun of VCs. But not only is that perception not true, it also isn’t singular to venture capital and therefore isn’t enough to explain what’s going on. [1]

Private equity investors, for example, don’t get the same level or ire and shade. Venture capital is really just a tiny sub-category of that asset class. Hollywood agents and executives operate under similar power dynamics as VCs but don’t inspire the same feelings. What gives?

I think the answer is that venture capital has so few measurable outcomes and the timespans over which they occur is so long that assessing quality is nearly impossible. A firm might have a great collection of logos on their website but you don’t know how much they paid, when they got in, how involved they are, or even if those companies will ultimately be successful. Just think of all those investors who made their names as big time Blue Apron backers.

No, in venture capital you can’t really build a brand or reputation on performance. You can’t advertise competence via returns. It just takes too long. VCs measure performance in 8-10 year fund lifecycles, not quarterly returns or annual revenue. Instead, VCs have to cultivate brands for themselves and their firms through omnipresence: they and their takes have to be everywhere always. 

That brand is how VCs get access to deals and founders. Moreover, they have to seem like allies, friends, and advisors to founders, not mere financiers.[2] This leads to a kind of new-sincerity particular to this weird little asset class. It’s all precious agreement, and mutual backslapping, and virtue signaling, and affected humility. Everyone is humbled and grateful for their own magnificent, towering achievements.

The incentives are there for VCs to talk and write and podcast and tweet and generally have an opinion on everything all the time. Cutting through the noise means having not just opinions but takes. Preferably steaming hot ones.

Venture capitalists need to look smart because it’s impossible to prove they’re smart. Perception, not returns, is reality.

The result is an insider-y industry dominated by middle aged white guys named Hunter and Travis and Jeff (and all the various spellings thereof) pontificating and self aggrandizing shamelessly, constantly, and at ever-increasing decibels (he said via his blog that no one asked him to write). [3]

Making fun of VCs has become a meme unto itself and has now begun transcending into the kind meta-meme or hyper-take the internet so often produces. Now the cool thing for VCs to do is to make fun of other VCs to show you’re in the joke (the VC associate wrote in his blog post about making fun of VCs).

The best we can hope for is to die young and never tweet.

[1] Almost no one makes money in venture, at least not off of performance (getting fat on management fees doesn’t count).  Benchmark returns are really pretty shit. Benchmark’s (note the apostrophe) returns on the other hand…

[2] Note my use of “they.” I’m one of the good ones and all my takes are great.

[3] Sorry to any readers named Hunter or Travis or Geoff. You’re reading this which already means you’re great. Love you.

Role Modeling

Beware, what follows is a post on startup financial modeling/ analysis. Bear in mind, I have no formal finance education. Proceed at your own risk. You asked for this.

When a startup gives you numbers, how do you assess what’s real, what’s good, what’s bad, and what’s possible? How should you think about not just how the business looks like today, but what could it look like in the future? It takes critical thinking, good judgement, and some pattern matching/experience. But to put that thinking to work, you have to first be able to structure your thoughts and see the picture clearly.

There’s tremendous value in model making as an analytical rather than predictive tool. 

When I’m modeling, my goal is usually not to determine what will happen, but rather to understand what would need to be true for something to happen. To my mind, the point of modeling is to ask and answer questions rigorously, and to be explicit about your assumptions. Putting things into numbers and breaking processes into discrete steps forces you to be specific in your thinking and with the story you’re telling, even if the numbers and steps are themselves unspecific.

Because startups are money-losing growth machines by design, lots of traditional financial modeling just doesn’t apply. Too often that means overcompensating and looking at top-line performance absent any more rigorous analysis of what I think of as “sustainability.” Is the growth healthy? People throw around all kinds of terms to asses the health and sustainability of startups. I think it’s mostly bullshit and doesn’t capture or describe anything meaningful.

I’ve found myself increasingly creating models (which again are thinking frameworks rather than predictive tools) to blend together all the various top-line figures into a more-startup oriented version of indicative health. I like to think about things in terms of payback in particular.

To understand why, I’ll use Harry’s as an example.

(Note: all the numbers here are waaaaay off and if this were a real example, there’d obviously be more detail. I’m illustrating a general concept, not proving a specific point.)

Harry’s baseline razor subscription costs $9 per shipment. Let’s assume that the average customer receives 6 shipments before cancelling and that it costs Harry’s $40 to acquire each customer. So each customer represents $54 in revenue and $40 in acquisition costs. Great news! For every $1 you spend on marketing, you get back $1.35 in revenue, a healthy LTV/CAC ratio (lifetime value/customer acquisition costs). You can check the glossary at the bottom for some definitions of terms:

But wait, you might ask, making, storing, and shipping all razors has to cost something too, right? Right. So let’s say all those various supply chain and logistics costs, eat 40% of the revenue (in reality for Harry’s it’s probably closer to 20%). That would mean that Harry’s only keeps 60% of the revenue from each sale, representing $32.40 over the life of each customer. Not as good as it could be but not too shabby…

This is where the math would break down. If each customer is only really worth $32.40 to Harry’s and Harry’s is spending $40 to acquire them, that’s pretty obviously unsustainable. And this is even assuming you know for sure what the customer lifetimes are. Really by the time you do know that, you’re probably not dealing with an early stage startup anymore. So the question becomes “what would have to change for this business to work?”

Here is the general framework I like to use to see all the levers in one place and begin thinking about that question:

To get that lifetime payback (think or this as CAC-burdened lifetime contribution margin or revenue minus acquisition costs minus the costs of the razors, shipping, etc.) into the black, Harry’s could:

  1. Increase order values, either by raising prices or introducing add-on products

  2. Improve contribution margins by lowering some supply chain cost (this is where the line items that make up CoGS matters a lot)

  3. Extend customer lifetimes/improve retention

  4. Lower CAC through better paid marketing or more organic growth

Obviously some of those are easier or more feasible than others. Some may tend to “naturally” (usually that really means a lot of hard work) improve over time and with scale while others may get harder or even worse over time. Where are the reasonable upper limits? What could this reasonably look like? What’s normal? What do we care about? Is this concerning?

There’s no “if X > Y then I must invest” shortcuts or hard and fast rules.

This is where it stops being “financial modeling” and becomes a more creative/intellectual/interesting exercise. But to get here and to be able to answer any of those questions with any semblance of legitimacy and analytical rigor, you have to first ground yourself in a framework. Otherwise it’s just hand-waving. That’s the ultimate point of the exercise.

What this framework also suggests is that to survive, a startup has to be great on one of those levers and to really thrive, it has to be great on two or three. So when I look at a startup, I want to understand where they are today and where they can get on each. Does the prospect of a super sticky product make me relatively indifferent for CAC? Can the margins be strong enough to support low AOVs? So on and so forth.

With the appropriate tweaks, this general framework works just as well for consumer startups well outside of subscription CPG. I’ve found it helpful for understanding startups with depreciating assets, lease obligations, software/virtual subscriptions, and one-off/irregular (non-subscription) purchases, etc. And it can work in reverse as well, not just on a per-user basis. Here’s the same type of unit economic model for Rent The Runway:

We’re not talking about top-line growth, and certainly not cash flow or EBITDA. Even once all the numbers in this type of model look good, startups still have to pay their employees, and rent, and lawyers, and consultants, and health benefits, etc. I’m not concerned about profitability as an early stage investor. I care about a path to profitability. Anyone can produce great looking growth by burning enough money. Getting to sustainability is another matter altogether.

To make sense of that and understand it in a meaningful way, you first need to do the work of modeling out both the facts as you them and your judgement as you perceive. Then you can start bending and breaking and hopefully get somewhere.

I’ve put the above models into this Google Sheet. Hopefully I can get around to adding more types of payback models. In the meantime, I would love to hear people’s alternative takes on this.

Am I dumb, wrong, or just pointing out something obvious?

Quick glossary:

  • AOV: Average order value. How much a customer spends on average each time they buy.

  • LTV: lifetime value (sometimes CLTV for customer lifetime value). The total amount of money a customer spends over all their purchases: AOV x number of purchases.

  • CAC/CPA: customer acquisition cost or cost per acquisition. How much it costs (in advertising spend, referral codes, etc.) to get a new customer. These numbers can mean slightly different things in practice but I’m using them interchangeably.

  • Payback period: the amount of time or number of purchases until you’ve paid back your CPA.

  • CoGS: cost of goods sold. The cost not just of buying/manufacturing some item but also all the associated variable costs of selling it like storing, packaging, and shipping it.

  • Contribution margin: the amount of revenue after accounting for CoGS.

  • Depreciation: how much does the value of something (like a car) decrease over time/with use.

  • Lifetime payback: this is the metric I’m making up, which works either on a lifetime or transactional basis. It measures (lifetime) revenue minus CoGS minus acquisition costs.

Going galaxy brain

Cryptocurrency and antitrust

A couple weeks I got into a vicious, heated argument with a dear friend. We were shouting at each other about antitrust, internet governance, and the centralization of the American economy for 30+ minutes. Yes, I know, I’m a great hang.

Over the course of that conversation, a couple of ideas really came into focus me - almost like a lightbulb going off. To get there, I want to first set the table of the debate:

I was my making the case to my friend (let’s call him David) that economic centralization is a bad thing because it forces us to rely on the unilateral judgement of a very small group of people. When firms get too powerful, they can squash competition and bend markets to their will. And make no mistake, I said, our economy has gotten more concentrated and less competitive.

Figure 1

This figures to get worse over time as the world moves toward a more technology-driven economy where network effects compound the returns to scale.

Figure 2. U.S. Market Share by Firm, Selected Markets

Moreover, I said, when we have so few and such powerful firms, if any one of them fails (either by going under or by making some choice with catastrophic social impacts) the consequences would be massive. Why should we expose ourselves like that? I’ve made this case before, here in 99D in Free Speech and Fuckery:

the internet collectively decided that Neo-Nazi site The Daily Stormer would be no more. It was de-listed from Google, kicked off its cloud provider, and blocked from most of the conventional/consumer internet. That time was easy. We all agreed that the Daily Stormer was awful and best erased from the world. It won’t always be this simple […] When a rogue twitter employee temporarily suspended Trump’s account as an act of protest on his or final day, it brought this brewing problem into sharp relief once again. This was an aberration caused by poor internal controls but it nevertheless illustrates a point: single private actor (be it an individual employee or the company itself) could silence the President’s primary method of communication with a keystroke. We’ve let individual judgement become the sole arbiter of free speech. We’re betting a lot that we’re making the right call and we haven’t done much to prove it.

The problem isn’t Twitter or Facebook or Google on their own.

The problem is that we have functionally ceded the public square to a few private companies.

Better, I said, to spread that risk around by having more competitive markets with less powerful individual actors. If you wanted to make a portfolio of stocks, you certainly wouldn’t buy just one, but buying 5 or 10 isn’t that much better. The same is true for decision making or economic power. To de-risk the system, you need to spread that risk around.

David, on the other hand, argued that too much competition is not only inefficient because it creates waste on fixed costs and spurs firm-killing price wars but also because it increases systemic risk by allowing anyone other than the best and brightest minds to wield power. This is similar to the concept of “Ruinous Competition” popular with 20th century monopolists like Carnegie, who’s not exactly famous for his benevolence and sense of fair play.

For David, we have to trust that our oligopolistic technology firms will not abuse their power as free speech gatekeepers and arbiters of right and wrong. Better them the government or the mob. Large groups move slowly and chaotically, he said. Government can’t be the one regulating speech and competition because it’s too incompetent, too dysfunctional, too damn slow. When it comes to big complex issues, we need decisive action not political grandstanding from the government or mob reaction from the whole population, right?

Finally, David made the case that even if you could unwind some large firms or prevent the big from getting bigger, it would diminish the power of the internet. You’d kill what you’re trying to save. “LinkedIn would be worthless if there were 10 or 20 different LinkedIns,” he said, “the value is in the centralized, dominant network.”

“What if it’s a false choice?” I asked. “But what about email?”

This is when the light bulb went off.

Email is the perfect system precisely because no one owns it. Email is not a company or a platform like Facebook or AWS. Rather, email is a protocol (called SMTP or Simple Mail Transfer Protocol). A protocol is like a rulebook that describes a series of standards and practices.

Everyone building an email system relies on the same rules so, unlike private messaging systems like Instagram Messenger or iMessage, email is fully interoperable across systems. You can send emails from a Google account to a Microsoft account to an AOL account because they are all built on the same open and publicly available protocol.

Anyone can build their own application (email service) with unique features on top of that protocol but the basic organizing principles are the same, ensuring decentralization does not come at the expense of interoperability or functionality. The protocol itself is maintained by a non-profit which makes decisions through consensus between its members. That’s probably not an ideal process but it sure is better than the voracious thirst for shareholder value by any means.

Why should it only be email that works this way? What if instead of 1) screaming into the void at Jack Dorsey to get Twitter to ban white nationalists 2) debating the relative merits of banning or not banning them or 3) trying to migrate Twitter’s users en masse to a separate platform that wouldn’t tolerate racism in the first place, Twitter users could simply use different, interoperable clients with their own rules, features, and moderation. What if Twitter were a protocol rather than a service?

If that sounds ridiculous, recall that that’s basically how Twitter functioned for the first several years of life before it began restricting developer access and shipped its own native mobile app. The same is true of Facebook, albeit in a different capacity.

In the very first post on the new 99% Derisible site, I wrote

Most dating apps are basically a thin application layer built on top of existing Facebook apps, data, and services. Tinder works so well because it has such wide reach/network effects (see: Metcalfe’s Law). It has such wide reach because the onboarding process is nearly seamless. The onboarding process is so good because Tinder has access to Facebook’s user data. Tinder takes your pictures, likes, demographic data, friends (social graph), and profile information from Facebook and Instagram for profile creation and matchmaking. 

So it doesn’t seem impossible or even far-fetched that the social graph and underlying user data/content for social networks could be made available and interoperable via a protocol rather than a gated, centrally-controlled developer platform. If Facebook, for example, were a protocol rather than a platform, social networking could become decentralized without losing the network effects of having everyone able to connect with everyone else.

Individual clients/apps built on top of that protocol could make content policing decisions and build new features on their own (just like email), minimizing the risk of any disastrous fuck ups by ultra-powerful and largely unaccountable oligopolists. But it’s so much bigger than Facebook or social media or even communications overall. What if everything, our whole economy could be similarly decentralized?

This is when I went full galaxy brain.

Wait a minute, I thought to myself, why does this suddenly sound so familiar?

Without meaning to or even realizing what I was doing, I had basically made a full-throated and impassioned political, economic, and moral case for cryptocurrency and blockchain.

I had previously been so limited in my thinking about crypto just as an asset class, rather than a governance mechanism or infrastructure design. What if we didn’t have to choose between regulatory wack-a-mole and internet governance by billionaire fiat? What if we could directly incentivize people to work on the protocols that would replace the centralized platforms rather than relying just on their goodwill?

It’s usually the crypto crazies spouting off on the necessity of trust-less institutions, decentralized authority, and distributed systems, not normal people like me.

And yet, once I began thinking in these terms, it suddenly seems ludicrous to entrust the management of our entire economy to the 12 governors of the Fed board. At least with a decentralized currency/store of value (like Bitcoin), we can minimize our risk and lesson our dependence on the decision making of a few people because every bitcoin holder has proportional say in monetary policy. There are not just 12 points of failure (functionally a single point of failure).

I truly have no idea whether prices will go up or down, whether cryptocurrencies today are a financially viable instrument, whether the underlying technologies are sound, or whether anyone will adopt enough of it to matter one way or another in 25 years. What I do know is that it has the theoretical/rhetorical potential to enable the kind paradigm shift towards a more competitive and less risky economy that I was advocating for.

That on its own is really exciting to me.

Special thanks to Mike Demarais for being my spiritual guide on my journey to galaxy brain bitcoin maximalism.

Loading more posts…