Founding a Startup, Just One More Time

I’m a 2X founder and a startup investor. What would I do differently today?

My entire career has been in startups.

I co-founded two tech startups, and was the first employee at another. Those products were used by Fortune 500 companies, governments and universities: even now, they form a core part of Canada’s national government intranet, and are how NBC News sends footage back to its newsroom. I was also an engineer at Medium, the platform you’re reading this on. I love to create new products, and I love to do interdisciplinary work. Building a product business with a human mission is fulfilling to me: it feels like all my muscles are being used. Like most founders, I prefer to use a broad range of skills than go deep on one.

Matter’s seventh cohort

Since January 2017, I’ve been the Director of Investments for San Francisco at Matter Ventures, which supports entrepreneurs building a more informed, inclusive, and empathetic society through a five month accelerator program. I’m responsible for sourcing the startups we invest in west of the Mississippi, as well as helping them raise money and develop their businesses. It’s been, in parts, a challenging, inspiring, heartbreaking, life-affirming journey.

Best of all, I’m surrounded by incredible people building important products and services that I’m glad exist in the world. My colleagues, who are design, product and business finance experts, are no less inspiring. I feel like I’m constantly leveling up from the people I get to spend my time around, and the focused accelerator program I help to run.

Given this context, and my background, it’s inevitable that I sometimes find myself thinking about what it would be like to start something new. I’m not about to found a new venture, but it’s fun to think about.

So, why not? I’ve decided to give myself some space for that thought experiment.

Knowing what I know now, from the founders I work with, my background in startups, and what I’ve learned from working at a values-based accelerator: if I was to do it all again, what choices would I make?

Two ground rules going in:

  1. This piece comes from me personally, so doesn’t necessarily represent the views of my employer. It does represent mine. I’m pretty opinionated.
  2. This isn’t meant to be advice. I’m thinking about what I would do. If you’re working on a venture, you need to think for yourself and make your own decisions. This isn’t a recipe book.

Let’s go.

0. Getting my feet wet

Before anything else — before incorporation, before any kind of code or formal team — there’s one key first step.

No, not a logo. Not a name. (Those can come later.)

There are good reasons to found a startup, and bad reasons:

“I’m going to found a startup and make millions of dollars” is a terrible reason. Not only is this line of thinking the domain of the Wall Street über-bro, there are far easier ways to make a lot of money. (Go work on the actual Wall Street, for one.)

“I’m going to work for myself and have a great working culture that I get to define” is, in my opinion, also a bad reason to found a startup. A great working culture is part of the means, not the end. It’s part of what makes the venture sustainable, but it isn’t its reason for existing.

A good reason is: “I want to solve a big problem that real people have”.

Actually solving a problem requires a pragmatic approach. It’s not enough to see that the problem exists: seeing it is one thing, but I’m going to need to deeply understand the people who have the problem, and through them, holistically understand the problem itself. I’ll need to talk to those people in order to get to know their needs and motivations; I’ll need to do research around the wider context and talk to subject-matter experts; I’ll need to feel the problem, even if it isn’t mine. And I have to be open to serendipity: I might discover an amazing insight on my journey that leads me in a different direction.

None of this has to take months. Anyone can assess this information quickly, by cold calling and asking the right questions to the right people. It’s work I would need to do.

Yes, I could scratch my own itch and solve one of my own problems. But as a thirty-something white man living in the San Francisco Bay Area, my sorts of problems are over-solved. If I’m going to make a dent in the universe, I want it to be helping other people. That isn’t just a social need: communities that are often overlooked by tech startups can be great target markets.

You also need to consider founder/problem fit. For example, should the problem of representation in media by solved by a straight, white dude, or might it be better if the masthead looks more like this?

There are techniques for discovering all this, of course. It’s one of the things Matter is very good at teaching, based on the design thinking methodology pioneered at Stanford’s d.School and elsewhere. The Program team are expert practitioners, and the workshops are incredibly effective. You learn these skills during the design thinking bootcamp at the very beginning of the accelerator program, and they come up again and again across the five months. If you learn everything you can about the people you’re trying to help, holistically, solutions you might never have thought of begin to present themselves.

Even then, you still need to validate your assumptions: it’s not enough to assume you’ve created a great solution for someone. You need to test, and test constantly.

Concreteness and detail here are important. I’m a very up-in-the-clouds thinker; I often fall into the trap of having too much focus on the long-term mission and not enough on the short-term strategy. I’m far from alone. I see a lot of teams who pick “millennials” as their target market. I once picked “content creators”. Neither of those are meaningful: you need to have a deep understanding of who these people are, in a way that can’t be derived from statistical studies or broad market analyses. Constructing personas, as many startups do, also won’t help you here. Neither abstracted demographic groups nor fictional people that you’ve invented yourself will give you surprising insights in the way that specific, real people can through qualitative conversation.

I’d create the roughest of rough prototypes and put it in front of a few well-chosen people to gauge their reactions. The lower fidelity the prototype, the more interesting the feedback will be: show someone a realistic wireframe mockup and they’ll give you feedback on intricacies of the UI, but show them a rough sketch, and they’ll give you feedback on the idea. It’s not about asking leading questions; just shut up and let them react to what you’ve shown them.

It’s a cycle: I’d incorporate their feedback, iterate my prototype, and test again. Once I felt like I was on what at least could be the right track, I’d start experiment with writing some code and making a functional prototype.

Here’s the important bit: all of this counts as the basics. I wouldn’t quit my job; I wouldn’t incorporate a company; I wouldn’t do anything formal. I might create a working name and draw an amateur logo, because it’s fun. But the bulk of my work would be teasing out a solution, validating my initial assumptions about whether that solution is desirable, whether it could be viable as a business, and beginning to build something that works. (As long, of course, as it was feasible — i.e., that it was physically possible to build it, with the time and resources at my disposal.)

Questions we ask of every Matter startup

Key question: once I’ve gone through the process of grokking the problem I want to solve and determining the direction I might want to go in, how do I feel creatively? Does it make me want to jump up on a chair out of excitement, like my first startup did? Does validating my hypotheses make me want to sing, or do I otherwise feel highly driven to work on it?

If not, I’m not going to take it further. This has to be something I’d feel elated to work on for the next decade of my life, through good times and bad. If it’s an unemotional undertaking, I’m probably not going to see it through. Building a startup is not something sane people do, and the struggle isn’t worth it if I just feel “meh” about it.

On the other hand, if I’ve validated my assumptions, and I can’t stop thinking about it no matter how hard I try …

1. The beginning

Key question: how far can I get without going full-time?

This isn’t something you hear a lot about in traditional startup propaganda, but keeping your job and having a side project may be a smart thing to do. If you’re able, drawing a salary and doing great work for an employer during the day can help you develop your first prototype by night. When I built Elgg, I worked for universities during the day, and then switched gears and was building my own software in the evening.

This isn’t an ability everyone has. It’s undeniably easier for 23 year olds, as I was at the time, and doesn’t necessarily work as well for parents, for example. People from low income backgrounds may struggle to find this time, which disproportionately affects founders from underrepresented groups. Many people (like me) also find themselves less able to work long hours as they get older.

It’s also true that many employers make you sign IP agreements that give them ownership over anything that you create that could be relevant to their line of business. For that reason, you should certainly never use company equipment to work on your own projects, and you shouldn’t do it during working hours. If you’re particularly unlucky, it may be that you can’t work on something on the side at all.

What this looks like is different for everyone, but I would want to go as long as I could without having to cut off my salary. This is particularly true now that I live in the United States, which has a brutal healthcare and welfare system. There’s no doubt in my mind that universal healthcare and the social safety net in the UK made me feel able to found my first startup, and there’s no way I want to be without health insurance or a stable roof over my head today. It’s not worth the risk.

My first rule of being a startup founder is: don’t materially harm yourself (or anyone else). This also applies financially; I once built up so much credit card debt that I had effectively red-lined a sports car. These are bad situations to be in, and the odds in startups are such that you shouldn’t bet on being able to immediately dig yourself out of a hole. (I was lucky.)

Ideally, I wouldn’t want to leave my job until I had a pretty good idea that what I was working on would be a viable business. I certainly wouldn’t leave my job until I had some kind of functional prototype: very few people will fund or buy into just an idea. An InVision mockup also doesn’t count: having to actually execute forces you to answer a different set of questions. I would also want to have enough personal runway — the length of time before the money runs out — to last for at least six months. A year would be better.

That doesn’t mean I would need the $55,000 it takes to live in San Francisco for that time; moving to shared housing, or even in with family members, could help. I’m not independently wealthy. It would be sad if only wealthier people could play this game, and I’m not interested in perpetuating a world where that holds true.

I would want to be within driving distance of San Francisco, or be able to take semi-regular trips in. My first startup was founded in Edinburgh, Scotland; my second in Austin, Texas; my third in San Francisco. I’ve seen first-hand that very few ecosystems can compete with the concentration of institutional knowledge, supportive culture, and infrastructure. I spent the majority of my first startup fighting against the idea — I shouldn’t need to live in Silicon Valley! — but I regret it. I actively harmed my business on ideological grounds.

2. Fellow travelers

Key question: who else do I need on the journey?

A lot of investors frown on single founders, and for good reason. Founding a startup is an emotionally rigorous journey. That’s a euphemism: it’s absolutely fucking shit a great deal of the time.

When you have someone to talk to, who’s embarking on the journey with you, you feel less alone. You get to spread the load: if you come down with something simple like a cold, it’s not all resting on your shoulders, so you can take care of yourself. A small, high-performing team can run faster than a single person — and early-stage startups need to run very, very fast.

There’s also another, more existential reason why I might need a larger team: if my skills, experience, and perspective aren’t enough to solve the problem. (They’re probably not.)

Some of the Elgg team in 2008

For example: I’m an engineer, founder, and investor. I know what those fields are like. I’m a straight, white man. I know what that lived experience is like. But if I need to serve people with skills that don’t fall into those categories, or who have different lived experiences that I might not fully understand, I’m going to need to find other people who can take ownership and help shape the product.

Everyone needs to reach an audience of people with diverse backgrounds, which means that everybody’s team needs to be diverse, too. A gaggle of white dudes can only do so much. Widening your gene pool of ideas will make you more resilient. And, really, who wants to work on a homogeneous team?

Even though I’m technical, my preference would be to have a technical co-founder. I’ve been a founding CTO and a technical CEO. Balancing the needs of product development and the needs of building a business is very hard, and it’s easy for both to suffer as a result.

As I wrote in another post, this needs to be a full collaborator will ownership over the business:

Not a technical collaborator, who only weighs in on technical issues, but a real, honest-to-god partner in crime. Someone who can help you navigate your startup challenges, and happens to have a complementary set of skills. Your technical lead is your product lead; put in that light, you need to be looking for much more than just an ability to code.

Of course, circumstances vary. There’s a chance I won’t find the right person straight away. In a founding team, it’s important that everybody brings something concrete to the table. That might be strong business or development skills; it might be deep domain knowledge; it might be extensive, trusted contacts. It can’t be nebulous qualities like “being very organized” or “getting shit done.” And they have to be in it for the right reasons: motivated by mission rather than solely by personal gain.

It’s worth saying that while going on the journey with other people is better, it’s not a hard and fast requirement to have co-founders. Rather than force an awkward relationship with somebody who isn’t really right, in the event I couldn’t find someone who is, I’d grit my teeth and go it alone until I could.

3. Incorporation and structure

Key question: which ownership structure will help me achieve my goals faster?

Incorporation isn’t something that needs to happen immediately, but I’ll need to do it before I deal with money in any way, and before I can formalize equity for any co-founders. I’ve seen this done really badly — in my first startup, we had to rewrite the entire set of formation documents as soon as we tried to raise investment. The best practices here are pretty cut and dry: if an investor or other decision-maker is looking at my business, the last thing I want to become a roadblock is my legal documentation. So making it as unremarkable as possible is a really good idea.

In the United States, that means forming a Delaware C-Corp (or, potentially, a Delaware Public Benefit Corporation, if I want to enshrine a specific public benefit into the legal documents of the company). All intellectual property is signed over to the company, and everyone has legal agreements that say any future work is also owned by the company.

It’s not a fun thing to think about, but sometimes, teams don’t work out, and somebody leaves. It’s a terrible signal for investors if you have an ex-founder who still owns a huge amount of the company: they could potentially obstruct deals, for example, diluting the influence of the remaining founders. There are lots of reasons why every founding member needs to have their shares on a vesting schedule, ideally over four years, with a one-year cliff. If someone leaves early, they don’t keep their full complement of shares — and if they leave before the first year is out, they don’t keep a stake at all. Everyone would have to stick around for four years to get their full allocation — unless the startup was acquired, in which case vesting would accelerate.

When investors in my first startup asked for this, I was skeptical. Do they want to shut me out and take my company from me? It’s a reasonable thing to ask, but vesting establishes trust, both in a founder’s intentions to stick around, and in the ability for the company to continue to operate even if the worst happens.

How much should each founder own? Founder equity is hard to calculate. There are tools out there that purport to help you figure it out, but honestly, I think they’re often too harsh. On one hand, if you have a 50/50 ownership stake, how can you tie-break decisions? (50/50 splits are another red flag that some investors watch out for.) On the other, if a founder only owns 5% of the company, what’s their incentive to stick around? Finding a balance is important, and also a signal that the founders can have difficult conversations. If I had one other co-founder, I’d probably aim for a 60/40 split, but it’s the kind of thing there isn’t a set formula for.

Once this has been determined, it’s important to know the difference between stake and influence. Even in a 50/50 startup relationship, it’s not true that all decisions are shared. If they were, everything would come to a grinding halt. Clearly understanding which decisions need a vote (perhaps even at a formal board level) and where each of you has an ultimate area of responsibility is important. Making decisions quickly and without argument is vital, and that’s what you need to optimize for.

A startup is a formal venture, not a project or a whim. It’s important to have these conversations early, including about what happens if everything goes wrong, so everyone is on the same page and has the same set of expectations. And if you can’t have those conversations, maybe that’s a sign that you don’t have the right team — or even that you’re not ready to be a founder. (Employees who can’t stomach the inherent risk are not ready to work for early-stage startups, either.)

Me? I’ve been in a startup with a 50/50 split with unclear responsibilities, and one in which we were very bad at discussing what would happen in a worst case scenario. We did okay, but I’m determined that in any future startup, clarity and transparency will be key, with an overall structure that encourages investment (if that’s what we want).

4. Technology

Key question: with the time, skills and resources we have, can we build what we need to — and if so, how?

I love programming.

I really do. Being able to sit down and create something from nothing is rewarding. Building software is surprisingly similar to building a written narrative, shitty first drafts and all. While we call it “engineering,” it’s an expressive, creative pursuit. It’s as much about art as it is about logic.

But writing software in a startup is a means to an end.

Everything is in service to the business; the code is a supporting player. You’re almost certainly not inventing transformative new technology in an early-stage business — you’re commercializing it. Disappearing for a year to do R&D isn’t a luxury you have. You need to build the simplest thing possible in order to get feedback and iterate your product. In lean startup methodology, you call that the minimum viable product: the smallest thing that will satisfy early adopters and turn them into customers. At Matter, we talk about the minimum desirable experience: the smallest thing that solves a problem for the real people you’ve identified.

The actual technology you pick to build your product needs to be driven by a business decision, too. Some developers laugh at PHP, but if that’s what you need to get the job done, great; look at WordPress, or Facebook. Some consider Node to be a fad; Netflix, PayPal and Medium are just a few of its heavy users, so I’m pretty sure it’s well-established at this point. Ruby on Rails, Go, Python, Swift, Objective-C, Java, etc etc etc; if you can use it to build the product you need, it’s the right thing. At the same time, remember that you’ll also want to easily hire developers, so Erlang or Haskell might not be your best choices. Annual licenses are a drag on the business, so maybe don’t use C# and IIS — unless, on the other hand, those technologies really are the best for the job. Hey, they might be. You do you.

I would probably use Node and React for a modern web application; your mileage might vary. As an open source veteran, I used to believe in fighting against lock-in at all costs, and there are still very good business reasons for this. Being able to easily change infrastructure providers means you reduce supplier influence in your business. If you build your platform on DynamoDB, for example, it’s going to be time-consuming and expensive to move away from Amazon Web Services. But the truth is, there’s always some amount of lock-in, and the ease and price of just picking a proprietary service off the shelf rather than building your own often outweighs — and should outweigh — other considerations.

As far as the act of building goes, I’m not a huge believer in heavy organizational frameworks like Scrum. I do like building in two-week sprints punctuated by tactical goal-setting meetings (and, depending on the team, daily standups). I’ve found that this mix works well for my in a startup context, and finds the right ground between needing to give developers the time and space to get their teeth into a problem and the need for agility and optionality. But again: your mileage might vary. There’s no set path.

5. Making money

Key question: $$$?

As a founder, you’re probably not Ev Williams. (Unless you actually are. Hi, Ev.) I know I’m not. The strategy of building something enormous and waiting to see how the community pans out before you figure out how to make revenue is something you can do if you’ve sold a few companies before and are able to raise money based on that traction. Particularly if you’re a first-time founder, but also if you’re yet to make your millions, you’ll probably need to have a concrete and detailed revenue plan — or at least a revenue hypothesis — far earlier in your development.

Businesses make money. It’s part of the definition of what separates a venture from a project. Startups aren’t just apps or services that you build for an audience: they’re entities that create and capture value at scale, but are at an early enough stage in their development that how, exactly, they’ll do that has yet to be fine-tuned.

Some startups exist solely to be acquired. They’re ventures established to capture a market or otherwise disrupt an existing, larger business, such that it’s easier for that business to buy them than not. I think that’s both risky and ideologically unwise: a model born almost entirely from greed. It’s far more interesting to create something that solves real human problems, in a sustainable way; these ventures are also far more likely to succeed.

Sustainability, for me, means obtaining these things:

  • Profitability, so that you can continue to evolve and grow your venture
  • Lifestyle sustainability, so you don’t burn out, and can have a family (if you want) and a life like a more-or-less normal human being
  • Scalability, so the ability of your venture to grow is relatively untethered from the effort and resources you have to put into it
  • Ethical strategy, so there won’t come a point where your customers (or the market) will turn against you for harmful practices, and you can sleep at night
  • Environmental and cultural balance, so you’re not making the world a worse place than you found it

All of these have an impact on how you structure your startup, build its culture, and figure out how you’ll make revenue.

Scalability is non-negotiable. Take a consultancy business, where you are paid for work on a per-project basis. Every project is different, so you have to put in a lot of work each time (even if that work diminishes a little over time, as you build a toolkit of reusable resources). Those projects are potentially lucrative, could be ethical, and might be environmentally and culturally neutral. But if you don’t somehow land the next contract, or if you’re unable to do a lot more work, you suddenly aren’t making any more money. These businesses are linearly tethered to the work you put into them. They’re not scalable.

There’s nothing inherently wrong with that, but it limits the growth you can achieve, and while your team is small it’s inherently unsafe. If you suddenly see more demand (a nice problem to have!), you can’t necessarily meet it without letting the quality of your work suffer. You can’t get sick, and someone has to be hustling for new big-ticket customers all the time. As the joke about freelancing goes, you’re free to work whichever 24 hours of the day you choose.

I’ve been on this treadmill before, and it’s not a lot of fun. It’s often suggested as a business model for open source or bootstrapped product development — the idea is that the product itself is free or low cost, but development is supported through consultancy support gigs. In truth, this barely works, because you no longer have the freedom to take the product where it needs to go in order to fulfill the larger need. Because the needs of each customer are so different, the product can easily lose focus, and because foundational development is rarely something clients see a need to pay for, the underlying platform can be neglected.

Subscriptions and advertising for a core product are two more scalable revenue models. While both have their merits, I would strongly prefer the former for my own startup: if your product provides real value, you need comparatively few customers to break even with a small team. Subscriptions are inherently recurring, so the revenue you receive from your customers is cumulative. Comparatively speaking, advertising requires real scale: millions upon millions of visitors to a mass-market site, who must keep coming back in order for you to see sustained revenue.

Targeted advertising also does real harm: it encourages platforms to spy on their users, discourages the privacy and anonymity that are required to protect vulnerable communities, encourages clickbait and ad-heavy pages, and I even think it has to take a lot of the blame for our current geopolitical climate. The European Union is beginning to fight back with the General Data Protection Regulation, and I’m convinced more legislation will follow.

But of course, I wouldn’t be able to just pick a revenue model on an arbitrary or ideological basis. Everything comes back to those real people I’m trying to help. What kind of model fits best into their world, and who would actually be making the purchasing decision?

I used to make software for higher education. It was important that my products were desirable by both the faculty and students, but neither group was actually the customer. I needed to know not just which department at the institution would buy, but exactly who would be making those decisions. If I knew them as deeply as I knew my users, I could figure out how to make my product enticing to them, and how the payment structure needed to work. (They’re unlikely to pay with a credit card on a monthly basis, for example; more usually, there’s a paper contract and an annual ACH payment.) Frankly, it would have been far better if the faculty had paid directly — but this just wasn’t going to happen, no matter how much I wanted it to.

The same applies to selling data insights. I talk to a lot of startups who want to sell preference, recommendation, or profiling data. While data may be the new oil, this level of detail isn’t enough: why will someone buy your data? Who are they, which insights do they actually need, and which form does it need to take in order to be useful to them? How much do you need to charge for it to be worth it to them? You can’t just ask someone how much they’re willing to pay; people are notoriously bad at reporting this. You need to know the customer deeply, and you have to be able to test your assumptions.

It’s tempting to try a smorgasbord of revenue models, but honestly, optimizing for just one is hard enough. I wouldn’t want to set myself up in a situation where multiple revenue models could potentially cannibalize each other (or muddy each other’s funnels), or where I could lose focus while designing the business. Adding multiple revenue models is something that can happen when you’re a much larger business. I want to focus on a single, simple, scalable, recurring revenue model.

I wouldn’t necessarily execute on it on day one. But I would sure as hell know what it would be.

6. Going to market and growing from there

Key question: how will we acquire customers?

If you build it, they will come.

Just kidding: they really, really won’t.

I need to do a lot more than build a product to get people to use it, so let’s bring it back to those original people I’m trying to help. How can I launch my product with them, and grow from there?

One of the most eye-opening experiences in my career was when I first worked with a large, well-funded startup. Until I joined the team, I thought they were winning by having the most beautiful interface in the market and a thoughtful product. Seems reasonable, right? They had hired some of the best product people in the industry and absolutely had those things, but it turned out they also had some very smart, very well-connected people who were making behind-the-scenes deals all the time. It wasn’t luck or merit; they were making partnerships through brute force sales and direct incentives, leveraging every relationship they had. Those partnerships would then lead, further down the line, to more organic user growth, once they had established a stronger reputation.

A lot of people assume that the tech industry is a meritocracy — not least because that’s the message that’s often cynically conveyed — but it’s not even remotely true. In the same way that the people who make the most money aren’t necessarily the most talented — they just happened to have the right connections and opportunities — the companies that succeed don’t do so because they’ve built the best product. Having a great product is table stakes; you’ve still got to have a smart strategy to bring it to market. You’ve got to have the right connections and opportunities.

You need to do things that don’t scale to get yourself off the ground. Those initial partnerships and customer inroads are vital. Frankly, they’re a lot harder to get if you’re not already part of the “in” crowd, which is another reason why those initial user conversations are so important.

Once I have that first group of loyal, committed users that I’m serving well, how can I grow outwards?

I’ll want to build in a mechanism for organic growth. If I’m building a subscription business, it’s harder to hit viral growth, because every onboarding user has to pass through some kind of a paywall. Still, it’s not impossible, and — depending on the needs of those original users! — naturally allowing each user to bring more users into the system may make the product more useful.

The k-factor of your product is calculated as the average number of invitations each user sends ✖️ the average conversion rate of each invitation. For example, if your average user sends 3 invitations and the conversion rate is a third, you’ve achieved linear growth: a k-factor of 1. Anything above that is considered to be viral growth. (Remember that an “invitation” could be shared content with a link to sign up, or an invitation to collaborate, or hundreds of other things.)

Subscription businesses will often have a k-factor of 0.1–0.3, because of the effect of the paywall. That’s okay, but it’s another reason why partnerships and sales are so vital. The simplest, dumbest version of this is paid customer acquisition through advertising, which can work if your conversion rate is high enough, but having a great reputation and being spread through word of mouth are so much more powerful. That, in turn, means finding ways to increase your credibility. And it implies, probably forever, having strong in-person sales and making deals that will either bring in big customers directly, or create a value incentive to bring more people in organically.

Something else to consider is the sales cycle: the length of time to complete all activities involved in selling to a specific customer. Remember I said I used to sell to higher education? The sales cycle to sell to a university at a campus-wide level can be as much as eighteen months. That’s because of how their budgets work in relation to the academic year, and the time you have to put in convincing them that they can trust you. Institutions will recommend your product to others, if they like it, but that organic process can take years. The result is very high prices, because that’s the only way to make the products viable. That’s actually not okay for a startup, and many enterprise markets function in a similar way. Meanwhile, selling to individuals or smaller companies is likely to be a much shorter, cheaper process — but you still have to sell to them.

My main point is: I’ve got to know how I’m going to reach people, how I’m going to grow, and I will need to fight in each case. Who will I start selling to? How will I grow from there? Which other people can I spread to next? How does that first group act as a springboard to let me better reach them? And most importantly of all, how can I give myself a sustainable, unfair advantage that prevents someone else from eating my lunch?

I’ll want to have hypotheses for all of these questions — and like everything else in my startup, I’ll want to be constantly testing my assumptions. It’s not about getting it right first time, or having a perfect model; it’s about having a direction and constantly doing experiments to get as much feedback as possible.

7. To the moon

This is just a thought experiment. I’m not planning to leave my position at Matter, and I’m not planning on starting anything new on the side right now.

I’ve deliberately been a bit coy about the actual topic of my startup in this piece. Because I meet teams every day, I don’t want to color anyone’s thinking or imply that I’m only interested in one kind of thing. As a media investor with an open source background who cares about privacy and inclusion, you can probably imagine the things I care about most, but I’m also open to be surprised.

Every business is different; the needs of every startup are unique to it. It would be wrong to take my ideas about how I would want to think about a new venture and use it as a recipe. My hope, instead, is that it can spark your own ideas, and help you decide which questions you might want to ask.

Startups are exciting. There’s a rush that comes from building something from scratch, and a feeling of power that comes from building everything on your terms. They’re also not the only kind of business. Just because something isn’t right as a high-growth business, or because it isn’t right for venture investment, that doesn’t mean it’s invalid as an idea, or in any way nonviable. A lot of the ventures I see would be better served as a non-profit organization, and that’s completely fine.

Finally, I hope I’ve conveyed the social responsibility involved in creating these businesses. It’s not just about creating something high-growth that makes money. The technology industry has transformed the way the world works. That means we all need to consider the social effects of what we build, and ensure that we’re world-positive: we’re building a sustainable, viable business while simultaneously leaving the world in a better state than we found it. With great power comes great responsibility. Anything less is strip-mining.

If you’re building a new startup venture, this will be one of the hardest journeys of your life. Believe me, I know.

Good luck; let me know how you get on.