I’m a crypto VC. That means I spend my days talking to crypto entrepreneurs, hearing pitches, and evaluating products. The first thing you realize working in this industry is that pretty much everyone is winging it. (That applies to me, but it especially applies to founders.)

Crypto moves at a lightning pace compared to most industries, but there’s no good blueprint for what we’re doing here. We don’t know how large the market will be, how to value companies, what metrics to focus on, or whether most of these products are even valuable. There are no IPOs or fully validated success stories beyond Bitcoin itself. This leaves founders and investors to figure it out as they go. The inevitable result of this is that there are a lot of weird companies in crypto; you’ve probably noticed this.

So I’ve decided to write this guide to building a crypto startup. I have worked in a crypto startup and co-founded one myself, but most of this advice is gleaned from observing many entrepreneurs more successful than myself building companies in this space. This will not be an exhaustive guide—it’s impossible for any document to be the final word on building a company—but consider this one investor’s outline of what I think about when looking at a startup’s trajectory.

Let’s take it from the top.

So you want to start a crypto startup.

Before you go any further, ask yourself: do you know enough about crypto to start a crypto company yet?

When I wanted to take a swing at my first crypto startup idea, I had embarrassingly little understanding of the blockchain ecosystem. It took me a while to appreciate just how little I actually knew—and how bad my initial ideas were!

If you are not deeply familiar with crypto, its culture, its products, and the history of the industry, then you should spend some time learning first. The fastest way to learn is by joining another crypto startup. Spending time in the trenches is the ultimate education in any industry. This is precisely what I did in my time at 21 (now known as Earn.com).

Read voraciously, go to meetups, read newsletters, play around with the tech, and immerse yourself in the community.

Let’s say you’re now pretty familiar with the industry—you get what it’s all about. There’s a second consideration before starting a crypto startup: are you technical?

If you aren’t technical, you will almost always need a technical cofounder if you want to build anything valuable. Wrestling with crypto requires deep technology chops, and a solo non-technical founder rarely gets funded. Of course, the best way to find cofounders is by working at another startup, so good thing you already did that.

As for assembling a team, the best teams are comprised of friends, or otherwise, people who have worked together before. Remember, the #1 cause of company failures is cofounder breakups. If you haven’t built up enough trust with your cofounders, you’re less likely to mesh well, stick together, and ultimately build a great business.

There’s also no need to jump directly into starting a company! Don’t incorporate before you’re ready to make that degree of commitment. It’s perfectly reasonable to start with a small project or consulting gig with your co-founders first. That will help you figure out how you work together and whether you’re a good fit before pursuing something larger.

The last question you should ask yourself is, why do you want to build this startup? Is it for money? Fame? Or because the world needs the thing you want to build, and you’re the only one who can build it?

A word to the wise: startups that are primarily motivated by making money seldom do. I don’t know why—it just doesn’t seem to bring out the best in people. On the other hand, startups that are motivated by an obsessive desire to change something in the world are the ones that tend to survive when the going gets tough.

Keep this in mind.

(Note: so far, this is mostly standard startup advice, so if this is new to you, you should definitely check out the YC startup school.)

But let’s say you’re convinced, you’re ready to build a crypto startup. Then it’s time to enter the ideation phase.

“The ideation phase”? But I already have my idea!

No. You need to workshop many ideas, because your first idea, sorry to break it to you, is probably bad. In fact, it’s almost certainly bad. It takes a long time to find good ideas in this industry.

Blockchains are complicated, confusing, and full of intellectual tar pits. Furthermore, blockchains are completely open platforms—that means pretty much anyone in the world can launch a competing crypto product. This means most of the straightforwardly good ideas have already been attempted. If you’ve come up with an idea, it’s probably either been tried before, or it’s such a bad idea it’s not even worth trying.

Genuinely good ideas are rare. Arriving at a one takes a long process of refinement and distillation.

Ideation

There’s a truism in Silicon Valley that “ideas don’t matter, execution is everything.” Like most startup advice, this is great advice not because it is true, but because it is useful.

In reality, ideas obviously matter. A bad idea, no matter how well-executed, will only end in wasted years of your life.

So how can you spot a good idea? The best way is by deeply studying the domain you’re trying to attack. Balaji Srinivasan describes this as “the idea maze.”

“A good founder is thus capable of anticipating which turns lead to treasure and which lead to certain death. A bad founder is just running to the entrance of (say) the “movies / music / filesharing / P2P” maze or the “photosharing” maze without any sense for the history of the industry, the players in the maze, the casualties of the past, and the technologies that are likely to move walls and change assumptions.”

First become an expert in your respective idea maze. Interested in building a DEX? A lending business? A market making firm? A new layer 2 mechanism? There is a rich history that precedes you, and you should study the other players in the maze, both living and dead, to understand how the maze is laid out. How do these businesses make money? Who are their customers? How did they differentiate? What features did they launch that drove their successes?

If you have no idea what maze to even pursue, I’d suggest engaging in this exercise: how do you believe the crypto landscape will change two years from now? Don’t build another layer 2 or another smart contract platform — people realized that was a problem a couple years ago, and by now they’re already very far along. You’re late to that race. You need to think about what will become more valuable in the future when your product finally reaches maturity. This requires vision and some conviction about how the future of crypto will evolve.

What will be different about the world by then, driving demand for your new product? Maybe it’s that we’ll have high-throughput L1s, maybe it’s that more real-world assets will get tokenized, maybe it’s mature interoperability—whatever it is, you’ll have to peer into the future and project what the market will need.

On the other hand, there are some idea mazes that we are pretty sure only contain dead ends; any entrepreneur that enters them is likely doomed to failure.

The following are some ideas that got funding in the ICO boom, but are now mostly un-fundable. For whatever reason, they still float around as cached ideas. Here are just a few of them from my own perspective (if your idea is in here, I would be happy to be proven wrong):

A new fiat-backed stablecoin. Unless you’re Facebook or JP Morgan, there are simply too many stablecoins at this point for yours to be competitive or differentiated. Even if Tether goes belly-up, there’s a lineup of fiat-backed stablecoins with strong distribution that are ready to take its place.

A decentralized stablecoin that is slightly different than another major design. (Crypto-collateralized, algorithmic, etc.) It’s too late to start from scratch here, and the decentralized stablecoins that exist have not demonstrated enough demand to warrant competitors. Retail investors have shown they’re unwilling to differentiate between stablecoin designs (for obvious reasons—they are all supposed to have the same price).

“Bitcoin-killers.” Sorry to break it to you, but Bitcoin has already killed all the Bitcoin killers. The reason why Bitcoin is king has little to do with technology, so competing with Bitcoin on the basis of tech is unwise.

The window for Ethereum-killers is fast-closing, and in general I don’t think they’re fundable anymore. There are already ~10 in the pipeline launching within the next two years; unless you have a profound distribution advantage, yours will probably be too late to be seriously competitive.

Blockchain for X. X can be advertising, oil and gas, bananas, whatever. It’s the same pattern as happened with “social networks for X,” or “Uber for X” — almost all of these are unnecessary and don’t successfully solve the problem they seek to solve. Especially if you try to monetize by adding a payment token for this mini-economy.

Let me develop this intuition for you, because so many would-be founders land on some variant of this.

Blockchain was first invented to solve the problem of creating decentralized money. Many people, myself included, initially assumed that blockchain technology could be generalized to solve many different coordination problems.

So far, that thesis has come out negative. We have not seen blockchains work in pretty much anything outside of money and property rights.

I don’t think this is an accident or just a matter of being too early. It seems, rather, that very few coordination problems have the particular shape such that technology alone can solve them.

It’s tempting to look at a fragmented ecosystem with players who are not cooperating and think “hey, I bet a blockchain could solve this.” I’ve looked at hundreds of startups pitching solutions that look like this. After digging deeply enough into them, the pattern is almost always that a blockchain alone will not solve the problem. Often these startups do not sufficiently understand why the players in their domain are not coordinating (usually, it’s because the dominant market participants are incentivized not to cooperate).

Alternatively, lots of “blockchain for X” plays are really about just getting everyone onto a common data standard. Unfortunately, this doesn’t actually have all that much to do with the unique properties of blockchain and is hard for other reasons.

Global, permissionless, programmable money and property is actually new. They were the first thing we created blockchains for, and up until now, they seem to be the only fundamentally novel thing we can do using this technology. I encourage you to search in that direction, as that’s what I believe most of the innovation in this industry will build on top of.

Building for customers who don’t exist yet. I see this all over the place. It’s really hard to build a product for people who aren’t here yet. It’s like trying to build iOS apps before the iPhone was invented.

This extends to building for the Fortune 100. They are not using blockchain. Yes, I know they say they are. They’re not; they’re running toy proof-of-concepts out of their innovation departments. This is analogous to cloud computing in 2004: for now, it’s a fancy buzzword that gets corporate innovation guys excited, but it’s way too early for decision makers to actually transition their infrastructure. Skip big corporates for now, and expect them to start using blockchain en masse 5–10 years from now. (Note: If you are a Fortune 100 company using blockchain in production in a meaningful way, I’d love to hear about it.)

You might argue: but crypto is so new! It’s going to change the world! It’s going to result in a totally new world with completely changed norms around money! That’s where my customers will come from!

Fine. But all of the great crypto companies that exist today built something that at least someone wanted at the time. Those people might have been a small group: cypherpunks, ETHheads, crypto traders, whatever. But don’t fall into the trap of building for a cohort that doesn’t exist at all.

The best way to bet on a future trend is by building for a small customer base that you believe will grow. Don’t posit a customer base that doesn’t yet exist and try to imagine their future preferences. You will wind up building for no one, with no feedback, and no idea whether you’re making any progress. For a startup, that’s pretty much purgatory.

Building for crypto influencers. This is another common trap for crypto startups. In most industries, if you build a product that influencers will love, millions of other customers will follow. But crypto is a weird space — the preferences of crypto influencers are very unrepresentative of crypto customers.

(These numbers are very loose approximations.)

If you went off crypto influencers, you’d assume most crypto consumers are paranoid cypherpunks who run their own full nodes and never hold their crypto on exchanges. In reality, this describes a vanishingly small slice of crypto users. Most crypto users hold their coins on exchanges, have no idea how to navigate a command line, and have never even heard of Austrian economics. They care a lot more about making money and a good UX than about decentralization.

Building in unfriendly regulatory jurisdictions. If you are building anything that interacts with thorny regulations (which at this point is almost everything in crypto), the jurisdiction in which you build your company matters a lot. If you’re building a global product, that means places like Hong Kong, Singapore, and Switzerland are great places to locate your crypto company. On the other hand, if you build a crypto company in New York City, you’re paying exorbitant salaries and rents only for regulators to repeatedly jab you in the eye.

That doesn’t mean that no successful companies are built in New York, but you should weigh the advantages against the costs. If you’re doing anything in the exchange space, DeFi, or are otherwise trying to build a global company (as much of crypto tries to be), consider your regulatory strategy carefully. Most entrepreneurs tend to gloss over this, but picking a jurisdiction from the outset is one of the most pivotal decisions for your company’s future.

That said, don’t be afraid of regulation! Regulation starts off as an impediment, but it can also serve as a powerful moat after you’ve established your business. On the other hand, you can’t really afford to spend all of your money talking to lawyers before you’ve got a business to speak of.

Once you feel your idea has been well-workshopped, you should flesh it out into a broader plan. Slava Akhmechet has a great 20-point checklist you should go through with your startup idea.

Validating your idea

You’ve now got the rough contour of what you’re doing. It’s time to go to validate it. Attend local meetups. Go to hackathons. Talk to other entrepreneurs. Tell everyone smart about your idea and get their feedback (safeguarding your idea is a huge antipattern).

Build a proof of concept. Show it off online or at a hackathon. Get people excited about your work. Many of the greatest crypto projects were built this way—InstaDapp was first built at a hackathon (it was originally named CryptoPay) and BitMEX launched their janky “alpha” exchange before they were able to raise external capital.

Spend as much time as possible talking to your users, figuring out their pain points, which products they’re using and where those products fall short. By now you should also know who your actual customers are going to be. Are they speculators? Traders? DeFi users? Exchanges?

If you’re working on something that’s more like basic science (think deep tech, a new blockchain, etc.), it’s going to be harder to get user validation. If that’s the case, you should be talking to lots of technologists and projects that might want to build on top of you. Design feedback is invaluable at this stage. But mostly, expect to be architecting and building on your own.

Raise money! (or don’t)

Let’s assume you’ve validated your idea and you are now confident it’s worth turning into a business.

The next question is: do I need to raise money? Often, the answer is: hell no! If your business is not capital-intensive, or cheap to prototype and get off the ground, then you might not need to raise money at all. CoinMarketCap, the largest crypto price aggregator, started as one engineer’s side project while he worked full time at an enterprise software company in New York—no traditional venture money. It’s now one of the largest advertising-driven businesses in crypto.

But many companies are capital-intensive (or just need to move faster to beat out competitors) and will need to raise venture capital. If that’s you, then let’s walk through what you’ll need to pull off a fundraise.

First, you need to prepare a deck. If you’re doing something technical, you’ll also need a white paper. (Why a white paper? Mostly because the crypto industry has cargo culted Bitcoin and Ethereum. It’s dumb, but best to follow industry norms here.)

Your deck is very important. It needs to convey what your project actually does (you’d be surprised how few do this clearly!). It also needs to convey how your project works. If as an investor I can point out any obvious weaknesses, attacks, or oversights that you don’t address, that does not give me confidence that you’ve fully thought through your system.

Your deck should also describe your economics and why your token / equity captures value. And given that it’s 2019, you need to include an at least plausible go-to-market strategy. “Build it and they will come” is no longer convincing given the thousands of crypto projects you’ll be competing against.

You should also start thinking about valuation. This doesn’t need to be in your deck, but you should consider how you will pitch pricing your company when the conversation arises. With early stage deals, this is usually done using comparables. Looking at projects that are in the same space, have similar quality, and raised in the same market cycle, how were they valued? (It won’t do you any good to comp against a project that fundraised in 2017; that was a very different market, and no one will price you the same.)

If you’ve never made a deck before, you can’t go too wrong following the Airbnb structure.

Now that you have that, you need to pitch an actual VC.

How do you get in front of VCs? The answer is pretty much always through their network, meaning you need an intro. This can be through another VC, or, ideally, through an entrepreneur they’ve worked with. (This is why it’s a good thing you’ve already done so much networking.)

Don’t count on cold approaching VCs at a conference, and don’t use LinkedIn. I honestly don’t know of a VC who’s made an investment in this space through a cold email unless the VC was already familiar with the project. Because crypto is such a global industry, we get so many random inbound pitches from all over the world, many of which are outright scams; there is simply no way we can diligence them all. Almost every VC firm in Silicon Valley will tell you: get a warm intro, and it’s doubly true for crypto VCs.

But if you come referred, you’ll usually cut through the noise. Of course, there’s an art to warm intros too.

Finally, once you’re fundraising, you’ll want to set an explicit deadline on your fundraise—both to build urgency and to signal your seriousness (also because VCs are busy and will often hit snooze on your fundraise if you let them). You should also set a deadline because fundraising is genuinely a waste of your time as a founder. You want to get fundraising done, out of the way, and then get back to building your business.

So who should you raise money from? First, you want to be mindful of what stage you’re raising for. You want to make sure the stage of money you’re raising aligns with the fund you’re talking to. a16z Crypto and Paradigm are amazing crypto funds, but they are both large, later-stage funds that in general can only write big checks—$500K pre-seed deals simply aren’t worth their time, since such a check would comprise less than 0.2% of their portfolio. But smaller funds like us are perfectly happy to write $250K-$500K checks. There are also active angel investors and family offices who are happy to invest at even earlier stages.

It’s always worthwhile to diligence your investors by chatting with entrepreneurs who’ve worked with them. In crypto especially, it’s rarely a good idea to take money from just anyone willing to throw it at you. The delta between “smart money” and “dumb money” is particularly large in an opaque industry like crypto. There have been many horror stories about investors kicking out founders, suing the company, or blocking subsequent rounds of funding. You want investors who actually understand what you’re building and are aligned with you for the long run.

That said, entrepreneurs often assume that raising money is simply a grading problem. You, as an entrepreneur, receive a grade based on the ranking of the VC firm that you raise from. But this is the wrong way to think about it. Raising money is not just a grading problem, but also a matching problem. VCs are matching you to what they need in their portfolio. If they are already heavy on exchanges, or they have a strong DeFi thesis, or they don’t have enough smart contract platform exposure, that will often determine whether you are a better fit for one VC or another. Alternatively, if they already know you, they will be better off investing in you than in another entrepreneur with whom they have a weaker relationship (or whom they just don’t want to work with!).

So talk to lots of investors. And a word to the wise: don’t over-optimize for valuation, especially in the early stages. I know this sounds self-serving coming from an investor, but you’ll hear this again and again. Valuation, early on, just doesn’t matter as much as picking the right partners, setting a good growth trajectory for your company, and not over-raising or overpromising to your investors. If you’re successful, you’ll make most of the money later on, not on your early fundraises.

It’s counterintuitive enough to be worth spelling out: raising too much money usually spells doom for a company. We all know of huge ICO projects that over-raised capital and are now sitting on their hands, unsure how to iterate on their tens of millions of dollars, locked into a business plan that no longer makes sense. And of course, all that money you raise is not an exit — it doesn’t end up in your bank account. It sits in your company’s coffers to try to somehow build your business. When there’s no clear path forward, companies tend to stagnate and devolve into politics and infighting. It’s not a place you want to end up.

If you’re doing an equity raise, your deal mechanics should look like standard startup financing (I recommend Venture Deals if you want to go deep here). But for token raises, there’s a whole separate playbook.

Token deal structure

Much ado has been made of corporate structure in crypto when it comes to tokens. Generally speaking, SAFTs are now out of favor compared to equity that comes with token rights (i.e., shareholders of your company will automatically be granted any tokens the company creates). But most good crypto investors can do either equity or tokens. Discuss it with your investors and your legal counsel.

You want to think carefully about the token distribution. Some rough rules of thumb are that you shouldn’t give the team more than 15–20% of the token supply, and investors no more than 30% of the supply. If VCs own more than that, your coin risks being panned as a “VC coin.” You want it to be more widely distributed than that.

On the other end of the spectrum, some token founders will sell far too little of their token. By selling a small portion of the network at as high a valuation as possible, they hope to maximize their ownership stake and make a bunch of money on paper. This is a huge, huge mistake. Say it with me — tokens 👏 are 👏 not 👏 equity.

The point of your token distribution is to distribute your token as widely as possible. Why? Because tokens are supposed to be global, decentralized, and money-like. In other words, a token becomes valuable because it is distributed. Owning 80% of a company would make you a savvy owner, but owning 80% of a token would make that token worthless. You should be excited distributing tokens because it makes your stake in that token more valuable.

Should you do a foundation? Where should you do it? What should the arrangement be between a foundation and the company?

As much as I’d love to definitely answer these questions, the reality is—nobody is really sure, and there are no best practices yet. In short, we’ve seen just about every model go horribly south, and it’s too early to say that any model has gone well. We’ve seen just about everything, and my guess is that this will continue evolving in the near future. Your best bet is probably just to talk to your investors, the entrepreneurs you trust, and sketch out a good approach you are are all willing to bet on.

Go to market

Go to market, distribution, there are many terms for the same thing. It’s the most neglected thing in crypto.

How will you attract your initial users? What distribution channels can you use? What kind of viral loop or referral program will you use to attract new users? What do you expect your customer acquisition cost (CAC) to be? What about your CAC-payback? You should have a plan more concrete than “promoting through influencers” or “market making.” To be successful in 2019, you’ll need some creativity.

If you’re doing a token, you also need to consider how to distribute your token supply beyond your pre-launch investors. Unfortunately, there’s no easy answer here either.

ICOs used to be the standard, but today they’re less and less common. CoinList lets you reach accredited investors, but many of those folks are just looking to flip your token, and there’s a lot of cross-over with VCs. Earn lets you distribute the token to Coinbase users while educating them on your product, but it’s not as turn-key. IEOs can be kind of shady depending on the venue, but they let you distribute the coin to a more global investor base (just be sure you’re legally covered, especially when it comes to US customers). Airdrops are a thing, but I don’t know of any non-fork coin that’s successfully been adopted on the basis of airdrops. There are creative approaches such as Handshake airdropping to Github accounts and Stellar airdropping to Keybase users, but the unfortunate truth is that most airdrops are never claimed.

If you’re building pure crypto, open source your code. Once you’re post-launch, if you want any hope of being eventually decentralized, this is a prerequisite. You have to bring the community into the fold of what you’re building, and they won’t trust it if they can’t participate in its development. You will also gradually want to extract your company from a central operational role in the business if possible. Good models to learn from here are Maker, Cosmos, and Ethereum. Cosmos in particular established a lot of the best practices around new network launches: see Game of Stakes and their phased releases. All software will have inevitably have bugs, but you want to catch as many of these as possible in realistic adversarial conditions.

Unlike the Internet, crypto is global from day one. That implies that no matter where your company is founded, you must eventually build a global team with boots on the ground around the world. Whether you’re from the US, Europe, or Asia, you want to have a presence in each of those geographies to build awareness and relay the needs of your different communities. It’s not enough to simply focus on your local geography.

Credit: Chappius Halder (note: this data doesn’t track China, which likely has more crypto traders than North America)

Finally, you must keep iterating on UI/UX. This is probably the most important frontier for crypto, and I suspect more and more businesses will differentiate on user experience rather than core technology. Taylor Monahan has written eloquently on the importance of communicating clearly to your user, and Austin Griffith has continually pushed the frontier of crypto onboarding.

Build stuff! Get users! Hire great people! Iterate!

From here on out, I don’t think there’s any more general advice I can give you. Your company will have its own unique needs and challenges, and it’s on you and your team to figure them out.

Wherever you are in this process, I hope this guide has been helpful in nudging you in the right direction. And I look forward to seeing what you build!

Thanks to Ivan Bogatyy, Tarun Chitra, Ali Yahya, Alexander Pack, Casey Caruso, Leland Lee, and Lucas Ryan for reviewing drafts of this post.