AMA with Marc Andreessen

Read this candid, wide-ranging discussion with one of the most influential Silicon Valley investors.

Avatar Photo of Patrick McKenzie
Patrick McKenzie

Patrick has built four software companies that did business internationally. He now works on Atlas at Stripe.

  1. Introduction
  2. Raising money from VCs
  3. Advice for founders
  4. Life of the mind
  5. Software eating the world

Startups are a curious alchemy of people, knowledge, money, and technology. Access to these building blocks has historically been grossly uneven but is improving over time. Open-source software and cloud services have made the core infrastructure of technology companies easier and cheaper to build than ever before. The internet has collected and distributed a growing body of work and practical know-how on how to build and scale companies.

There still exists a dependence on certain people, though, particularly where the allocation of money is concerned.

Stripe Atlas helps founders worldwide start ambitious companies. One of our aims is to decrease the social distance between our entrepreneurs and investors, experts, and others who can help them. To do this, we created a private Stripe Atlas Forum, and we have been inviting guests to speak directly with Atlas members and answer their questions about building companies.

Marc Andreessen is a co-founder of Andreessen Horowitz and famously coined the phrase “software is eating the world.” Prior to that, he cowrote Mosaic, the first mainstream web browser. While he’s unfortunately no longer an active Twitter user (his @pmarca account made for compulsory reading while he was), he did agree to answer over 30 questions from Stripe Atlas founders. These include questions about how to convince a venture capitalist to give money; what has changed about the tech industry since he cowrote the first web browser; and where opportunities still exist to build huge, meaningful businesses around technology the world needs. Marc graciously gave us permission to post a portion of his answers publicly, for the benefit of the wider community.

If we can help you with access to the people, knowledge, and technology that will help you start and grow your new business, join Stripe Atlas. (And if you need the money, read Marc’s advice below on pitching venture capitalists.)

Raising money from VCs

What is most important for you when you decide to invest in a startup?

There are three different answers, for the three different stages.

At the seed stage, when a startup is brand new, the decision is driven almost entirely by the people. Who are they, and what is their pedigree and track record to cause one to expect that they can build something special?

At the venture stage, when a startup has a prototype or an initial product but not yet a fully functional business, the decision is some combination of the people, as with seed rounds, but also product or market fit: Is there reason to believe that this product in this market at this time is going to take off?

At the growth stage, when a startup is fully in market and building out sales and marketing efforts to expand, the decision becomes far more about the financial characteristics of the business, particularly unit economics: Can the startup profitably sell its product to each customer?

See also these two blog posts for a lot more information on these topics:

How many investing pitches do you receive?

These numbers are approximately but directionally correct:

We receive about 2,000 inbound qualified pitches per year. By “qualified” I mean we already know people directly involved in the company or people associated with the company in some way (angel investors, other VCs, advisors, coaches and mentors, lawyers, or customers).

We think this is 2,000 out of a pool of approximately 4,000 startups that hit the bar where they might plausibly be able to raise venture capital in the year. (These numbers are US-only to simplify the topic.) We don’t see the other 2,000, either because we don’t know anyone associated with the company (probably a fault on our part) or because there is some reason we wouldn’t be a good investor for them (for example, if we have already invested in a competitor in the same space).

Out of the 2,000 we see, we will make somewhere between 20–40 investments per year. So around a 1–2% hit rate.

The very best VCs in the US collectively make perhaps 200 investments per year. Out of those 200, about 15 of them will generate 90%+ of the investment returns for the entire year.

So our job as investors is to try to nail as many of those 15 per year out of the 20–40 investments per year we make.

The very best VCs in the industry seem to be able to invest in maybe 2 or 3 of the 15 each. So, by definition, even the best VC in the business will miss most of the big winners. It’s a humbling business.

Did you ever invest in an unknown startup that contacted you via email the first time?

I don’t think so.

At first blush this might seem insane. Why would a VC only invest in people they already know? Doesn’t that cut off original thinking from people outside of the existing network? Isn’t it a fact that many of the most successful startups come from founders brand new to the industry?

The reason is subtle but important. Getting a warm introduction to a VC is a basic test of networking skills.

VCs are dying for interesting, qualified referrals from people in their network—angel investors, other VCs, advisors, coaches and mentors, lawyers, and customers. All of those people love giving qualified referrals to their favorite VCs. VCs are some of the easiest people in the world to reach via their networks.

It turns out that the skill required to network into a VC is the same as the skill required to network into a customer, into a supplier, into a distribution partner, into the press, into an executive search firm.

And so if a founder can’t navigate a network into a VC firm, it is unlikely that founder has the skills to navigate the other networks required to succeed in building a company.

Although this may sound harsh, it isn’t intended to be. The best startup advice of all time comes from Steve Martin: “Be so good they can’t ignore you.” In this case, that means, be so good at networking that they can’t ignore you. The skills you develop learning how to navigate to VCs will pay off 1,000x in building your startup more generally.

What numbers are more important when investing at seed stage?

For us, at the seed stage, 90%+ of the decision is based on the pedigree and track record of the core team. So we don’t tend to look at numbers very much at that stage. We’re almost always betting on a particularly special team doing something new and interesting—it’s a qualitative evaluation, not quantitative.

How do investors evaluate startups that have done well in their home country and are looking to expand in the US market? Do they look at the past performance and decide, or is traction in the US market a must before the startups start reaching out to investors?

VCs vary on this question. Some actively seek out companies that are successful outside the US, some wait to see if those startups can succeed in the US, and some don’t invest in startups based outside the US at all.

In our case, although we have occasionally invested in particularly special startups based outside the US, such as TransferWise and Improbable, we generally invest either in startups based entirely in the US, or startups that use what might be called the “Israeli model” of building R&D in their home country but building SG&A (sales, marketing, finance, legal, etc.) in the US. As suggested by the name, some of the best Israeli startups have been executing this model for the last 20–30 years; more recently, we are seeing founders from many other countries (Canada, China, Brazil, Argentina, Pakistan, and more) pursue the same model.

The common narrative now is that the startup valuation peak was 2015, and since then fundraising has gotten tougher and the most profligate startups are going out of business. Where do you see things going from this year through the next couple years in terms of valuations? New highs or retrenchment?

JP Morgan was once famously asked if he thought the stock market would rise or fall, and he answered, “It will fluctuate.” Which is also my answer here :-) I think it’s impossible to forecast these things—think of the many people who began predicting a new tech crash starting in the mid-2000s and are still wrong more than a decade later.

That said, I would characterize the current fundraising climate in the US as vigorous but somewhat discriminating. I think you are right that in 2015 the market became somewhat overenthusiastic in that virtually any startup could raise money and many at prices that in retrospect were far too high. Whereas today, high-quality startups that should be able to raise money generally are, but lower quality startups may run into trouble.

What tips would you give first-time entrepreneurs meeting investors to avoid taking money from particular investors who would not be good for the business?

That is a great question. Founders should reference check VCs just like VCs reference check founders. As a founder, I would talk to as many people who have worked with a particular VC in the past as possible—other founders, angel investors, executives, lawyers, etc.

As with anyone in business, any VC should be happy to give you a long list of people they have worked with in the past who you can call. If a VC won’t do that, beware, beware, beware.

There are many questions you can ask references, but I would really focus on how the VC operates under pressure. Everyone can be supportive and behave well when things are going well, but like anyone else, VC behavior varies wildly when the going gets tough, and that’s when founders tend to really regret their choice of investors.

We plan on launching our product in July, and if things go well, we plan to raise our Series A by December. When is the best time to start this process?

I believe in an orchestrated process that doesn’t meander but nevertheless gives VCs time to make a proper decision—many VCs will simply back out of a process if they don’t have enough time to do their work. A good wag is to allow three to four months for the whole process, from introductions and first meetings to closed contracts and money in the bank. So if I were you, I’d probably start conversations around September, which is also when many VCs get back from their August vacations (sad but true).

Advice for founders

What do you recommend for pricing in SaaS before reaching product market fit?

Pricing is highly specific to the product and the market, so it is hard to give general advice.

But if I were to give general advice, I’d say that we see far more SaaS startups underpricing their product than overpricing.

The problem with overpricing seems obvious: We in our daily lives as consumers are more likely to buy products if they are cheaper, and so pricing higher is presumed to reduce sales.

But that’s not how business markets tend to work. In business markets, where customers make what’s called a considered purchase—the result of a reasonably objective and rigorous analysis of options—startups that underprice tend to have the problem I call “too hungry to eat.” By pricing too low, they can’t generate enough revenue per deal to justify the sales and marketing investment required to get the deal at all. In contrast, by pricing higher, the startup can afford to invest in a serious sales and marketing effort that will tend to win a lot more details than a competitor selling a cut-rate product on a shoestring go-to-market budget.

TLDR: When in doubt, double prices :-)

When we talk to a potential investor, should we tell them about previous projects we had that failed?

There’s an old sales slogan: “If you can’t avoid it, feature it.” Tell a story about how you worked your way through different ideas before you got to the one that works—the one you have now. By doing so, you can convey your level of determination and ability to react to changing circumstances—both valuable traits in a founder.

The thing to definitely not do is “hide the ball.” It is a very bad idea to not tell potential investors about negative things that you know they would want to know. In addition to the ethical considerations, there is a very real practical consideration—investors almost always learn the truth, through their diligence work and reference checks. When an investor realizes in diligence that a founder has hid the ball on bad news in the past, it creates the concern that the founder will hide the ball on bad news in the future—that the founder can’t be trusted. It is much better to be the founder who is up front, crisp, and articulate on bad things that have happened in the past and what you’ve learned as a result.

You have stated that a good market can compensate for a lackluster team or product. What markets do you feel are currently underserved?

I don’t tend to have a view on any particular market. I find that successful startups tend to be highly idiosyncratic—they combine multiple elements including product, market, team, business model, timing, culture, strategy, and tactics in unique ways. So I don’t believe in examining any single element, such as markets, in isolation.

That said, some of the markets that we think are particularly underserved by advanced technology and tech startups right now include healthcare, education, real estate, transportation, law, government, defense, and financial services.

Life of the mind

What are the books on your bookshelf with views and statements you disagree with, and what’s the source of disagreement? How often do you revisit these books versus those where you agree with the written content?

That’s a great question.

I’ve been influenced by way too many books to list, but here are a few that I keep coming back to:

  • The Sovereign Individual—Written 20 years ago, this is the most thought-provoking book on the unfolding nature of the 21st century that I’ve yet read. It’s packed with ideas on every page, many that are now fast becoming conventional wisdom, and many that are still heretical. Two related books to read are The Twilight of Sovereignty and Cryptonomicon.

  • The Baroque Cycle—A work of rigorously researched historical fiction with only the slightest overlay of science fiction, this book tells the story of the emergence of the modern world and its systems (democracy, the scientific method, financial markets, etc.) in a way that is wholly fresh. These novels make me think about what a Neal Stephenson of 2300 may write about our times and us.

  • The Innovator’s DilemmaThe Lean Startup, and Zero to One are the defining trilogy of intellectual thought on the art and science of modern technology startups. Virtually every page of each is open to debate, and yet, as a whole, they provide intellectual scaffolding for our endeavors that I wish had existed when I started in 1994.

  • The book I am most looking forward to is The Square and the Tower, on the rise and fall and rise of networks and the eternal battle between networks and hierarchies at all levels of human life.

In the 10 years since you originally wrote that article on product or market fit, have you developed any additional strong or weak signals that you use to determine if a company had product market fit?

We recently wrote two blog posts on exactly this topic:

As a startup entrepreneur, especially one with a potentially huge impact on people’s lives, do you have to constantly try to assess the potential ethical ramifications of business choices, or do you just let things roll, focus on growth, and let the market sort it?

I think that it is always important, in life and work, to have a sense of the ethical impact on one’s actions and output. Virtually all of the actually capable startup founders I know think deeply about the ethical aspects of their businesses, contrary to whatever outside commentators might say.

That said, history shows that it is particularly difficult to forecast either the benefits or downsides of new technology. The most classic example is nuclear weapons—many of the inventors of nuclear technology had justifiably serious concerns over how their work would be used. And yet, nuclear weapons not only helped end World War II and almost certainly saved lives on net for both the US and Japan, but a case can be made that the existence of nuclear weapons and nuclear deterrence prevented a catastrophic third world war between the US and the USSR that could have killed hundreds of millions of people in the decades following WWII.

I think based on this and many other examples, we should be very cautious at forecasting negative implications of new technologies—sometimes the forecasts are correct, but more often they simply show a lack of imagination and foresight as to the positive benefits to come.

Software eating the world

Have you seen any changes in either the backgrounds or personalities of founders in new companies versus in the previous 10–20 years? If so, what changes or trends have you seen?

I think there are two major changes that are somewhat opposite of one another:

First, there’s no question a far larger number of very young founders are starting companies. The rise of accelerators, incubators, angel investors, seed funds, online funding platforms, and the like have made it much easier to start a company than 20 years ago, and so far, more companies are being started, and by younger and less experienced people. This is wonderful because it is expanding both the number of experiments that the startup ecosystem can run each year and the talent base of the people building startups.

Second, 20 years ago, most startups were what I call “tool builders”—they built tools like chips, operating systems, routers, or databases and sold those tools to business or consumer customers to use however the customers saw fit. But today, more startups are what we call “full stack”—instead of building tools, they are building technology, and then using it to directly enter end markets in competition with incumbents. These full-stack startups are more operationally intense and tend to require more experienced founders and executives. And so, one sees older, more experienced, more operationally competent founders and core team members for many of those startups.

Real estate is a concern for many businesses (especially startups). Do you see any future disruption in that industry due to software? If so, what kind and to what degree?

I think it’s an absolutely huge issue, particularly for startups based in popular urban areas like the SF Bay Area, and it’s an issue across both commercial and residential real estate (offices and housing). We are wide open for disruptive startups that can make an impact on either commercial or residential real estate, and we have a number of investments already, of which the most successful so far is Airbnb.

I think there are two broad categories of obvious potential disruption.

The first is disruption that makes commercial and residential real estate as it is currently understood better, cheaper, easier to access, or all of the above. I would include in this category disruptions in transportation that make real estate more generally accessible—for example, self-driving cars should make commuting time much more productive and hence open up outlying areas to more housing for professionals.

The second is disruption that helps remove the need for geographic colocation entirely. A running joke in Silicon Valley is the help wanted ad that goes something like this: “Engineers wanted for startup building software to enable collaboration on complex projects online. Must be willing to relocate to San Francisco.” As funny and accurate as that is, I am a believer that telepresence and collaboration software will eventually enable far more remote work and virtual teams, reducing the importance of colocated real estate. It can’t happen soon enough.

Historically great technologies are adopted by great minds as hobbies. What hobby trends are you currently seeing?

Great question. We call this “what nerds do on nights and weekends.” This is the single most reliable source of new ideas in our industry that will ultimately be adopted much more broadly.

Right now we see huge nerd night-and-weekend energy in areas like cryptocurrency, biohacking, quantified self, synthetic biology, virtual reality, drones, and self-driving cars.

Do you see any opportunities, or downsides, in highly regulated fields like health tech and legal tech?

Yes—both opportunities and downsides.

The opportunities are massive—highly regulated industries like healthcare and law tend to be very large and very poorly served by advanced technology. In economic terms, they are both a high percentage of GDP and have low productivity growth, so the opportunity for startups is very large.

But the downsides, or risks, are also significant. These are markets that tend to have three difficult characteristics. First, they are highly regulated, meaning they are just hard to enter in general. Second, they tend to suffer from “regulatory capture”—the existing incumbent companies in these markets tend to have wired the regulatory system for their own benefit and to exclude competition. And third, they tend to have high levels of government subsidy for consumers, meaning that the government is a major payer and in some cases the only payer—and it can be very hard to get the government to pay for something new, even when it’s much better.

In short, these markets are very appealing for the very best founders who are prepared to deal with the additional complexity and difficulty.

Disclaimer: This guide is not intended to and does not constitute legal or tax advice, recommendations, mediation, or counseling under any circumstance. This guide and your use thereof does not create an attorney-client relationship with Stripe, Orrick, or PwC. The guide solely represents the thoughts of the author and is neither endorsed by nor does it necessarily reflect Orrick’s belief. Orrick does not warrant or guarantee the accurateness, completeness, adequacy, or currency of the information in the guide. You should seek the advice of a competent attorney or accountant licensed to practice in your jurisdiction for advice on your particular problem.

Ready to get started? Get in touch or create an account.

Create an account and start accepting payments—no contracts or banking details required. Or, contact us to design a custom package for your business.