Should you do a startup? A tactical checklist

On getting on the other side of the "AI counterfeiting white collar work" divide

The argument for doing a startup:

1. When working for some company, even an elite company like a FAANG or finance company, you are replaceable cog #24601, your individual actions and talents barely matter, and your output and impact is easily replicable by many others.

2. Doing a startup uses your skills and talents to the fullest, as you literally create a new product or service, create new jobs that didn’t exist before, and drive new and incremental economic value in the world at a much greater scale than you ever can as an employee. Your positive impact is multiplied tens of thousands-fold, generally.

3. Creating a company, an economic engine that you’re a part owner in, puts you on the other side of the “AI counterfeiting white collar jobs” divide - as a business owner, you now stand to benefit from that dynamic in the future, vs as an employee it’s all risk and loss.

But doing a startup, as great as it may be in relation to being an employee, isn’t for everyone.

I’m here to go over the primary filters you need to get through, the things that are necessary but not sufficient, and the EV calculations you should be doing as you think about your market and about whether you should take the plunge or not. This is a tactical checklist of the important factors to consider about your personality, your network, and about your idea’s plausible market sizing and EV.

Throughout this post, I’ll be considering the “peak optimum” best case, then will take it down to more realistic levels, and talk about common pitfalls and mitigations.

Why should you listen to me?

I started my first company when I was still in college. I had a list of ideas and a couple of thousand dollars to my name, and ended up inventing a new type of flip flop with a hidden pocket, then bootstrapping that to ~5 employees and sales in two countries.

Since then, I’ve spent many years of my life burning furious 80-100 hour weeks doing a few more startups, each in different industries, and have invested in several others. I’ve also worked in finance and FAANGS and have led a bunch of high powered STEM-degree Ivy folk in tackling various high stakes problems that generally drove in the tens of millions of NPV annually, so I have a decent idea of the backgrounds and alternative options for most people who are talented enough to do a startup successfully.

Filters to get through

Who is the ideal startup founder, from the standpoint of an investor? Just like with credit and finance, the ideal candidate is somebody who doesn’t need you at all:

1. High human capital - the ideal candidate is smart AND has grit AND is social and charismatic AND is a good leader AND can get into the technical details AND is good at sales, etc.

2. Legible credentials and success signs - there’s a reason YC has a fetish for Ivy degrees - Ivy degree holders are usually legitimately higher human capital, have better social networks, and impress other people enough they’re more likely to get high quality founders, additional funding rounds, etc.

3. The ideal founder has similarly great and talented co-founders, but not too many, just 1 or 2 more

4. The ideal founder has built an MVP with their own funds and is showing early traction

Now, obviously most of that isn’t going to be true for most founders, I’m pointing to a global optimum, a peak. But this should give you an idea of what VC are looking for when they invest and what other founders are looking for in their business partners.

But back to the “unnecessary” point, the best startup candidate is somebody that can fairly easily get FAANG or finance jobs cranking $500k+ a year. They are *guaranteed* to be millionaires in their lifetimes, with much easier work weeks than a startup founder.

Then on top of that, they need to be motivated or crazy enough to be willing to work more than twice as hard as any FAANG job, for much less money, with less probability of success, to do the startup they’re talking to you about.

So let’s say you’re not the ideal candidate, but you want to know the minimum standard you need to bring to something to have a chance at succeeding as a founder.

The types of people who should consider doing a startup

As I’ve said, I’m very pro-startup - I think creating something new, bringing a new product or service into the world that solves a pain point, is tens of thousands of times more valuable than working in finance or at a FAANG.

After all, if you create something new, every incremental customer you touch was affected by your personal decisions, skills, and talents. More, if you create a company doing that, you literally created jobs, and entirely new categories of value, and they literally wouldn’t have existed without you.

So broadly:

1. If you’re multi talented and routinely do “hard things” AND

2. You have a good social network with similarly talented people AND

3. You have an idea of a pain point that you and your network are uniquely suited to tackling, and that pain point affects a lot of people, AND

4. You and your team are willing to absorb a lot of costs and burn furious 100 hour weeks for years

THEN you should consider doing a startup.

What is necessary but not sufficient?

1. An incredible amount of motivation - if you and the rest of your founders are not willing to put in 80-100 hour weeks for years, maybe a startup isn’t right for you

2. A great idea - startups are about finding a “pain point” that affects enough people and is motivating enough that people will happily pay for your solution - we will talk more about sizing this later

   

3. The right team to tackle that idea - lots of people identify an idea and basically have one or more “???” spots where a miracle is supposed to happen, and then a clear road to success and plaudits past that point. This is usually non-technical people hand-waving things like “building the actual product,” or handwaving “then we get 1M engaged daily users,” or some similarly difficult core competency. Your founding team should cover those “???” places, you can’t just handwave them. As in, you should have a technical person who actually knows about building great products, and a marketing person who has some idea of the cost, channels, and expense of acquiring 1M engaged users, and so on.

4. Talented cofounders and a good social network - for some reason, “lone wolf” types always want to do a startup, probably because they have higher innate Disagreeableness on the Big 5 / OCEAN characteristics and hate having bosses. I’m not saying it’s impossible, but succeeding is way, way less likely as a lone wolf, versus as somebody with a robust social network and other talented founders. If you can’t convince other legibly talented people to join you, it’s a pretty serious red flag.

I recently ran across this illustration on Roberto Ferraro’s Substack, and broadly agree:

Oh, and it’s important to note that getting all three factors right isn’t a guarantee that you’ll succeed. You can still get hit by macro-economics, unknown unknowns, some gap in you and your team’s competencies, technological shift, being counterfeited by a big existing player, and way too much else to list.

There are no guarantees when creating something new, and that itself is probably a major factor to consider. If you have higher than average neuroticism or needs for safety and predictability, doing a startup is probably a mistake.

Problems and pitfalls (and mitigations)

Lack of credibility - if none of you have ever done a startup, if none of you have legible and impressive credentials, if none of you have a background and work history that’s relevant to the pain point and solution you’re aiming for, it’s a problem. What can you do to overcome this? The closer you can get to an actual MVP and actual traction, the better, and the more of those problems you can offset. Otherwise, reach out in your network and find people that fill those gaps.

Perfectionism and “ideas” focus - Ideas are cheap, so cheap they’re worthless. Successful people build and ship, and have a bias towards praxis vs theorem. Yes, you need an idea good enough to be relevant to many people, and to address a pain point bad enough that people will happily pay for a solution. But that’s a pretty low bar - we all see and live pain points every day. Figure one out that affects enough people and start building a solution.

Gaps in your team or understanding of the path to a solution - the infamous “???” gaps. What can you do to mitigate this? Make sure you have people on your team that know how to do those parts, or at the least, you need to demonstrate an understanding of roughly how complex and difficult the problem is, and a clear plan to hire the level of person who can solve that complexity and difficulty of a problem once you get the funding you’re asking for.

Creating something easily counterfeitable - everyone trying to create “Uber for dogs,” “Doordash, but for chimps,” and “Instagram, but for knees” and other things along those lines has two problems - first, will people pay your proposed price in sufficient numbers? But more importantly, if it’s that great of an idea and you prove it publicly, why isn’t Uber or Insta going to counterfeit you immediately, with their 10M times bigger user bases, marketing budgets, and lock-in?

Undercapitalization and lack of financial acumen - a lot of people who get into doing a startup have zero accounting or P&L experience. A common pitfall is starting with whatever amount they can come up with from themselves and friends and family, and hoping it will be enough - it’s basically never enough. These people are undercapitalized, and are going to struggle, and will likely fail. As a CEO, your primary job is making sure your company has enough money for the next year, recruiting and hiring top talent, originating and communicating the vision, sales and business development, and keeping everyone on track for hitting timelines and milestones. But doing that well requires at least a high-level understanding of how much money that will take, and how much money you have, and how much money you expect to make when the sales start cranking. Mitigation - demonstrate that understanding. Learn accounting or have somebody who understands it. Know what a P&L and what your burn rate is now and will be in the near future. You’re seeking this much now, and that will let you hire these people and hit milestones X,Y,Z, at which point you’ll raise your series B or C round from <blank>.

Identifying the right idea

We all see and live pain points every day, it’s just a matter of noticing them.

And on the general theme of this post, identifying the right pain point to tackle is another way to uniquely contribute value in the world. People’s lives are complex and high dimensional enough to uniquely represent them. You’re at a unique junction in terms of the people you’re connected to, the various different social worlds you connect via work and hobbies and personal life, and the problems you see in your job that affect both people and organizations. The trick, as always, is identifying something at the intersection of those things that is high value, and that you’re uniquely qualified to tackle, by dint of your experience, background, and connections.

High opportunity areas include pain points for companies and organizations, areas where you’re well connected with an outstanding talent that wouldn’t necessarily think to do a startup themself, and areas at the intersection of less frequently connected domains, like software dev and car racing, or data science and banana farming, or whatever.

But you should size the problem and the likely median and best case success scenarios to make sure it’s even worth it, because the costs are real and significant.

EV calculations

Lots of people who decide to work on a startup sell themselves short. If you’re giving up an easy $500k comp job, you’d better be shooting for something great. That’s $2.5M over 5 years. Let’s call it $4-5M fully loaded, because you get non-comp benefits, quality of life, and probably some advancement and raises in there.

If you have the standard ~10% chance of making it with your startup, your personal “success” EV needs to be at least $50M to make the gamble and 80-100 hour weeks worth it. So the "median plausible success case” you’re shooting for needs to be (your equity %) * company value * 10% success rate >= $50M. If you’re going to end at 33% equity, say, that means the company needs to have a plausible median success case of being worth at least $1.5B in 5 years. And 33% is an optimistic percentage! To be safe you probably want to double that, because you get diluted over multiple rounds and with multiple founding partners.

If it’s not that, you’re basically working for free, and should stay at your FAANG job.

The math works out similarly for below-FAANG job tiers. But you’ll notice you need some pretty aggressive values to be worth it. Even if you’re at half-FAANG, you need to be cranking on a company that can plausible be worth more than $750M in five years.

Probably the least anyone who can make six figures should consider is a company that has the potential to be worth $500M.

Let’s take it back to sizing your pain point and idea

A company value at a $500M size backs into the market size and price points you’ll need fairly easily.

Business values generally go for 5-8% cap rates depending on the industry, so just think like a private equity person. To hit a $500M valuation, you need at least a ~$40M EBITDA at an 8 cap. What can you do to plausibly hit a $40M EBITDA? This is simple math too - you need some top line revenue R minus COGS and operating expenses. As a rough rule of thumb, you’re probably gonna have to crank ~$100M in revenue to hit a $40M EBITDA. So what does that amount to? One hundred $1M dollar customers, or a hundred million $1 customers, or something in between. But now you have a rough idea of the size of the “pain point” market you need for your idea, because you’ll have an idea of your industry. If you’re in social media, your customers are worth $200-$300 a year, so you need to be able to plausibly have at least 300-500k annual users to hit your $100M. Sounds feasible! Banking or finance is generally the same depending on your segment, but $200-$1k is roughly right, so you need 100-500k customers. If you’re in enterprise software, your average license might be $200-$1k a seat, so you need that same 100-500k seats in your end state. See how easy this is?

This is the kind of napkin math you need to be doing when considering your overall pain point and potential market size, because you want to be working on a big enough problem to be worth it.

Far too many people either do way-too-optimistic napkin math at this stage¹ or don’t really consider the opportunity cost.

For god’s sake, consider the opportunity cost

Because I’ll be honest - I see a LOT of founders working on projects with a lot less headroom than this. Too many people are working on companies that might be worth a few million or tens of millions at an optimistic end state. I think they’re strictly self-owning - don’t be that founder. “Go big or go home,” as they say, or in this case, go big or stay in your easy six figure job.

The quality of life aspects of a startup are real - they’re hell on relationships, you’re gonna miss a lot of your kids’ key moments and milestones, and most people are burning those furious 100 hour weeks and absorbing all those costs for basically nothing, because even if they succeded the success case isn’t worth the opportunity cost! Even if you’re young and single, you could be out having fun and investing a good chunk of your six figures instead of burning 100 hour weeks, and that’s not a trivial thing to give up!

But okay, maybe not everyone is going to be able to crank on an idea worth at least $500M. I think you should seriously think twice and thrice before deciding on that, but it can be done in a sensible way.

When should you consider a company that’s only plausibly worth single to tens of millions?

I’m not saying “never do a company that will be worth under $500M,” I’m just urging you to use your head. Most small businessess are worth less than that, and many small businesses are worth it for their owners.

This isn’t insane, because small businesses generally don’t require the bone-deep commitment and crazy work weeks that startups require, you don’t get diluted, and you can generally de-risk things.

1. If you can self-fund with your other founders, or friends and family fund, because VC and investors aren’t going to be interested, generally. Other options are traditional bank loans or SBA if you have good income and credit.

2. If you can work on it as a side project alongside your “real” job and de-risk it sufficiently that you prove the model and traction and can know that it will work.

3. If you’re fine with creating yourself a “job,” as lifestyle or mom and pop businesses usually require your ongoing attention and time, and aren’t really as amenable to exits or setting them up with a good manager and forgetting about them.

Can it still be worth it to do that? Absolutely. There’s lots of lifestyle and mom and pop businesses out there that were worth creating, and it’s still better than working for somebody else. Also, you generally aren’t diluted, so even if it’s only making a few million a year, you and your partners get most of that.

If you’ve got an idea and an edge and know where to get some seed money, go for it. There’s little downside, and small business owners are still cooler than employees, are driving more value in the world, and generally have better quality of life.

Most importantly, it will put you on the other side of the “AI counterfeiting white collar jobs” divide.

It’s future-proofing

As AI ramps up, one thing we know is that more white collar jobs are counterfeitable. You know what’s a lot less counterfeitable? Being the boss and owner of a given company / economic engine. Even if you decide to ultimately replace some employees with AI, you’re the one on top there, and now you’re the one benefiting from these trends instead of worrying.

Who knows how inscrutable smarter-and-faster-than-human minds will change the economy? It certainly seems feasible that more entrepreneurial opportunities and pain points will be snaffled up by faster-than-human minds as things unfold. Certainly if large tranches of white collar jobs are counterfeited, the competitive pressures of starting businesses are going to be significantly higher, simply from the other humans out there looking to succeed - this is a chance to get in on the ground floor now, and create an economic engine that is exposed to more of the AI upside than downside going forward.

But wait, if AI is so good it can counterfeit while collar jobs, won’t it counterfeit these businesses?

I think a business typically has many more moving parts than just "white collar workers producing a product," and your business as a whole won't necessarily be counterfeitable by AI for a while, because you have a whole interlocking set of gears adding value to the various AI-counterfeitable inputs.

As an example, telling an AI "make Google maps for me, but for free and without Google" doesn't work. They spend zillions of dollars on people and satellites and cars with cameras and all sorts of value added stuff, only to give it away for free as part of a bigger strategy of supporting an ecosystem that is funded by being one of the online ad duopoly.

Most businesses should have similar non-counterfeitable layers of value-add within them.

Businesses are basically defined by "assembling a bunch of white-collar-and-other labor and processes in creative and value-adding ways that produce products or services other people pay for," and being able to counterfeit a few steps in that chain doesn't necessarily mean you can counterfeit the whole business easily.

Although I do agree that the competitive business landscape is probably going to change pretty significantly and get fiercer and faster - when a smaller company can use AI in nimbler and smarter and more efficient ways, they have a chance to take out bigger players who are moving slower and have less adoption and more internal resistance. But this is an additional opportunity that chalks yet another point up in the “do a startup now before the divide” column, in my own opinion.

1. The famous example of overly optimistic napkin math is along the lines of “clothing is a $555B annual industry, and I’ll be conservative and say I’ll capture 1% of that, for an easy $5.5B in revenue per year”

Clothing is impossibly large and diverse, and even the biggest companies you’ve definitely heard of struggle to capture 1%, and also your margins are shit because the market is so competitive marketing and operating costs are really high. That whole elision where somebody goes “I’ll say I’ll capture 1% of that” is the “???” in this particular example, because it would make anyone with domain expertise laugh.