Aaron's Essays

Asking questions

Small children are great at asking questions. Their questions are simple, direct, and utterly without ego. If you spend time with 2-5 year olds, you’ll find yourself answering a near infinite series of progressively more challenging questions. This is one of the main paths by which children learn about the world.

ex. Child question: “Why is the sky blue?”


Adults are generally bad at asking questions. Adult questions come with long preambles, caveats, and agendas. Adult questions don’t seem to be designed to learn about the world, they seem to be designed to avoid looking dumb.

ex. Adult question: “I know the sky isn’t actually blue, and that if I look up, I know there’s nothing between me and outer space except air and maybe some clouds. Beyond that, there’s outer space. Outer space is effectively black because there isn’t light. Yet the sky often, though not always, appears to be various shades of blue. What triggers that blue? Is it something about how light interacts with air? Or gravity? Magnetic fields?”

The contrast between child questions and adult questions is striking, but it isn’t a function of age, it’s a function of conditioning. At some point, children get social signals that asking questions makes them appear stupid. This can come from a frustrated parent, annoyed teacher, or mocking friend.

Kids internalize this feedback and start to ask fewer questions as they grow up. The scope of things about which adults ask questions get narrower and narrower, shying away from things that are new or hard to understand. At the same time, adult questions get longer and longer. Most of this length isn’t important for the question, it is meant to prevent the audience from thinking the questioner is stupid.

These questions aren’t terribly useful to either the questioner or the questionee. If you’re not sure what this type of question looks like, watch a congressional hearing. The questions are meant to prove points, not produce knowledge.

Adult questions reinforce their own badness. When an adult spends most of a question proving his own intelligence, the question doesn’t produce new knowledge. This reinforces the perception that questions aren’t worth practicing. Without that practice, the questions continue to get worse.

I’ve been thinking about questions because I noticed that I’ve become increasingly self conscious about not knowing things, and cover for that by asking adult questions. Some of this is a function of the fact that I’m older and some of it is a function of the fact that my job involves giving a significant amount of advice. I noticed, as well, that conversations in which I ask simpler questions are more enjoyable and more interesting. These conversations lead to new ideas and better plans.

I’m particularly conscious of this dynamic in conversations with founders. Many founders believe that adult questions are the best questions because of their interactions with investors. On the one hand, investors tell founders that asking questions is a good thing. On the other hand, founders are judged for asking questions that are “too basic.” Founders who do things without asking questions first are generally rewarded for being “fast” whereas those that ask some questions first are derided as being “slow.”

However, successful startups are not defined by the questions that a founder asks. They are defined by what a founder does with the information at hand, and how quickly. Founders who acquire new information quickly and act are more successful. Founders who dither or get caught in infinite loops of ignorant execution generally fail.

Founders and investors would have better conversations if each side dropped the pretense of needing to “look” smart and asked child questions. This isn’t a complex task, but it does require focus.

The way I’m working on this is to stop myself each time I’m about to ask a question and figure out what I actually want to know, and then see if I can just ask that specific question in no more than a single simple sentence. If the question does not produce a satisfying answer, I’ll try to understand if it is because I worded the question poorly, or did not provide enough context. If I need more context, I’ll add one or two sentences of context, and ask the question again.

This strategy does not always work: I don’t always catch myself in time, I sometimes ask poorly worded questions, sometimes the question requires more context than is practical. However, I’ve found that the act of thinking about the question I want to ask and examining it in this child/adult dichotomy has helped me ask better questions, have better conversations, and learn more things. That’s enough of a reason to continue doing it.

Fooled by experts

Experts are generally right until they're wrong. Unfortunately, it's very easy to get fooled into thinking that experts are always right. This is because they are...experts. They are authoritative and knowledgeable. This is especially true when it comes to trying new things in existing fields. We are biased into believing that knowing a lot about something confers an ability to predict the future.

The problem with expertise is that it doesn't necessarily come paired with an openness to new ideas. Expertise can be used to shut down new ideas and avenues of exploration just as easily as it can aid in invention. I've been guilty of this when hearing about new ideas. In fact, it often feels easier to shut things down than to use what I know to figure out how to make something new actually work.[1] Perversely, being negative may make someone seem like more of an expert than they are, creating a negative feedback cycle.

When you don't know much about a subject, you're free of the constraints of what has been tried. That means people who are inexpert will often have wilder ideas. Sometimes, that's called naive or stupid. However, when those ideas happen to work, then it's called genius or groundbreaking.

That doesn't mean that every problem can be solved by creative ignoramuses. There seems to be a level of expertise which, when paired with the right environment, is conducive to productive creativity. I wish I had a way to know those levels for different areas.

Without a clear set of rules to use in evaluating expertise, I instead try to evaluate founders based on what I can learn from talking to them. This is especially true in areas that I understand fairly well. While there are certain things I might expect a founder to understand about what they're doing, I'm much more interested in how they think and test assumptions. This is part of how I try to figure out if I'm investing in potential (good) or track record (not so good).

One of the tricks in doing this is to see whether or not the tests that founders run cause them to ask increasingly interesting questions. Coming up with questions is a good sign of creativity, while answering them well builds the expertise necessary to actually get something done.

Technology increases the likelihood that a seemingly naive approach to a problem will work because it reduces the iterative cycle of trying that approach. Given enough time and resources, you could try every single solution[2]. In the real world, though, we have to pick solutions to work on. This is where the good founders are separated from the bad ones. The good ones get better and more open as time goes on, while the bad ones get more closed down and start rejecting ideas out of hand.

The same is true for investors. The best ones use new knowledge to open up new ideas, while the worst use it to close out entire categories of ideas.[3] That's short sighted because the world constantly changes, which makes new things possible. Founding and an investing in those new possibilities is what creates the companies that change the world. That creates new areas in which to be expert, starting the cycle all over again.

__

[1] https://archive.org/stream/ERIC_ED211573#page/n0/mode/2up

[2] And enough monkeys at enough typewriters will eventually produce Hamlet.

[3] This is different than investors who understand the limits of their expertise and focus on specific sectors. An investor can be sector focused and open to lots of new ideas within that sector, just as an investor can say they'll invest in anything only to shoot down every idea.

Thanks to Craig Cannon for feedback.

Why VCs sometimes push companies to burn

Originally posted on the YC blog, here: https://blog.ycombinator.com/why-vcs-sometimes-push-companies-to-burn-too-fast/

In Investors and their incentives, I tried to give a broad breakdown of the incentives that drive the major types of startup investors. I want to dig deeper into a behavior that VCs sometimes exhibit which seems strange from the point of view of founders. Despite praising frugality, VCs sometimes push companies to spend more money, faster. Sometimes this leads to faster growth. More often it leads to empty bank accounts. I've seen a number of companies get killed by this dynamic, and it took me a long time to understand why it happens.

Misaligned Incentives

Though VC firms generally act in ways that they believe are most likely to help a company succeed, their incentives can become misaligned around questions of burn because of uncertainty and time. Since VCs know that most of their returns come from a very small number of bets, they need individual partners to spend large amounts of time with the companies that they've funded. That time is required to help (in the best scenarios) and to figure out which bet is actually worth further investment of time and capital. In an ideal world, VC firms would be able to increase partners and investing capital whenever they want to add investments. That would, however, require infinite accessible capital and ever larger partnerships. The first doesn't exist and the second would be hard to manage and support.

Funds are also under pressure to produce returns within a given timeframe in order to demonstrate quality for raising further funds. These capital raises are important because many VCs make more off of management fees than they do from investing well.[1] The most important thing for the near and medium term success of a firm is its ability to generate more fees. With an average fund life for early stage investors of 10 years - and a pattern of raising a new fund every 2-4 years - VCs must show some kind of progress. There is rarely enough time for a firm to return capital to LPs before raising a new fund, so they use private valuations to mark their portfolios up or down. If a fund can show a material markup, they have an edge in marketing for their capital raise. While funds that don't produce meaningful returns to investors generally fail after a while, it takes a really long time for that to happen.

Combining these two incentives creates a situation in which it is better for a VC firm to push a company to demonstrate success or failure quickly rather than move more slowly. Companies that succeed quickly lower the uncertainty involved in investing further in that company, and justify the amount of time spent by an individual partner on that investment. If a company is doing well it is also likely to attract new investments at higher valuations, which allows the firm to mark up its investment.

Conversely, companies that fail remove themselves as a time commitment for a partner. That's not an ideal outcome, but it means that the VC firm can refocus energy on companies that are doing well and on finding new companies to make up the valuation lost through the failure of one piece of the portfolio. The faster this happens, the better for the investor.

The worst situation for a VC is one in which the firm has made a significant investment in a company that just muddles along, constantly threatening success and failure. These companies require a lot of time and effort to figure out whether or not they can be saved. They generally generate significant team drama which investors sometimes mediate. They often present difficult bridge financing questions, and they rarely function as good marketing fodder.

How Founders Should Respond

Figuring out how to deal with these pressures is important, and varies depending on your relationship with your investor. The most important thing to remember is that the CEO controls the bank account. Investors can pressure founders to spend faster, but they cannot force them to do so. Founders need to have their own understanding of where and when to spend money, and what rate of spending makes sense.

Next, remember that promises of more funding made when things are going well aren't worth much. All that matters is the money currently in your bank account. It can be easy, during heady days of growth, to justify any and all expenditures. I've talked to founders who have done just that, saying that they need to hire more engineers to develop new features for new users or that they need to spend on advertising to ramp up acquisition. Invariably, they talk about how investor x or y told them that they're special and had promised to keep funding them no matter what. Investors will often point to the burn at successful companies like Uber to prove that spending is good. Founders will sometimes accept this logic without thinking deeply enough about whether or not the lesson applies to their own situations.[2] A company can get away with this so long as growth is working, but as soon as that company goes sideways, spending becomes problematic.[3]

The closer your company edges towards “not likely to return a meaningful part of the fund,” the faster investor capital will dry up, no matter the promises made early on. This is where investor incentives diverge from those of the company. As the investor gets surer that the company will fail, the investor usually pulls back. As the founder gets closer to failure, the founder's need for active engagement and help grows. This mismatch can kill companies and relationships.

It may not be surprising, then, that the more money a founder has in the bank, the stronger their position when fundraising. In fact, we've seen a few YC companies raise financing rounds without having spent any of the money from a prior round. They do this because they get offered very friendly terms. They can do this because they've figured out how to grow without spending lots of money.
The Second-Worst Case

When incentives start to diverge, one of two things can happen. The first case is a bit better, though not great for the founders. There are times where the VC will see a failing company and decide that success is still possible. When this happens, the VC may invest more into the company but will only do so on terms materially more favorable to the investor than would otherwise be the case. This results in significant loss of control and equity for the founder, and may involve the founder being replaced. Whether or not this happens depends on how much leverage the founder has left, and as a bank account approaches 0, leverage decreases asymptotically.

Still, the company may survive.

The Worst Case

The company runs out of money.
__
[1] https://hbr.org/2014/08/venture-capitalists-get-paid-well-to-lose-money
[2] I addressed a bit of this here: http://www.aaronkharris.com/exceptionalism
[3] My partner, Dalton Caldwell, wrote a great post about what you should do if you find yourself in this situation: http://www.themacro.com/articles/2016/01/advice-startups-running-out-of-money/.

Thanks to Craig Cannon, Paul Buchheit and Dalton Caldwell for your edits. Thanks to Paul Graham for initially explaining this dynamic.

Fundraising isn't predictable

I recently had a conversation with a founder who'd just finished a number of unsatisfying conversations with investors. The founder's company was growing well, had crossed $1mm ARR, but still couldn't raise money. According to conventional wisdom, that should have been enough to open some checkbooks. Unfortunately, that's not how fundraising works.

This is strange because of the way we generally talk about startups. In attempting to set up the kinds of experiments on which startups are built, we tend to put precise milestones in place to gauge results. This is great when measuring progress, but it implies a false sense of precision with regard to raising money.

I've started borrowing a concept from insurance to frame this problem: Parametric triggers. Parametric triggers are a tool used by insurance companies to govern payouts from policies. These triggers remove subjectivity from the claims process by using independently verifiable, empirical evidence as the decision tree for claims.[1] Applying this concept to fundraising would mean that there is a precise set of things a startup could do to get funding. That's how we'd all like it to work, but there are no specific events that will automatically get money from an investor.

In reality, fundraising happens for companies under a ranged set of conditions. Those ranges and the accompanying outcomes are generally different for different companies, or even for the same company at different times. To make matters even trickier, those ranges are only apparent in hindsight when looking at aggregate information about a large number of deals. What those ranges don't tell you is where the actually successful companies sat in those ranges.

This dynamic is apparent in the aftermath of Demo Day at YC. If you were to isolate on a single characteristic of companies within a batch (say, revenue), you'd find companies with impressive revenue growth failing to raise capital, and ones without much revenue raising significant capital. In a parametric world, this wouldn't happen.

The reason this does happen is that parametric triggers are a great tool for deciding what to do about something that happened in the past. In insurance, parametric triggers determine a specific payout for a specific event - an earthquake of 4.7 magnitude will pay a policy, while one of 4.6 will not. Parametric insurance doesn't account for costs after the event or the payout. The system has all the information it needs to make a decision at the decision point.

Investing in startups isn't about what has happened. Investing in startups is about what will happen. Investors are trying to find small companies that will become multibillion dollar ones. They use past performance as an indication of the quality of the founders and the idea, but no metric is a perfect predictor of the future. The ranged conditions that investors often use as filters - i.e. 15-20% growth per month for SAAS companies [2] - are useful tools to use when sorting for worthwhile conversations given limited time. Those filters are never sufficient for actually making an investment, and every investor has a different framework for making that final decision.

When investors turn down a deal, the good ones will give the founders at least one reason. This may take the form of “we don't understand how you can scale,” “we think there are too many competitors,” or “we don't think you're growing fast enough.” While it is important to think about the reasons given, never assume that the only thing between you and the investment is proving that you can solve for that condition.

What you actually need to do is figure out what pieces of evidence you can use to create a future for your company that investors would be foolish to ignore. This will likely require some mix of your traction, the market you are attacking, and how impressive you and your team are. You also need to recognize that there are contra-indicators of success which investors have been trained to recognize. If they see a small ultimate market with no adjacencies, then incredible early growth has no chance of producing a big company, so they will pass.

Much has been written on how to pitch and fundraise, which is part of the problem. Founders are drowning in examples that seek to turn fundraising into a deterministic process. Accepting that it isn't is a counterintuitive and critical lesson in fundraising successfully.

Here are examples of meaningful achievements that won't automatically get you funded but will get an investor's attention and can be used as part of the story you tell:
  • $1mm ARR
  • 10% weekly growth
  • Repeat founders who were previously funded
  • 2 Fortune 500 pilots, 5 in pipeline
  • Engagement rates that are higher than Facebook

Fundraising is a frustrating process that resists well-meaning attempts to make it into a predictable science. That's actually quite fitting, though, since while much about startups can be measured, that measurement never guarantees success. Success comes from using the unique advantages you have and combining them into a coherent, self-reinforcing story and product.
__

[1] https://www.wikiwand.com/en/Parametric_insurance
[2] http://tomtunguz.com/mrr-growth/

Thanks to Craig Cannon, Dalton Caldwell, Sam Altman, and Andy Weissman for your edits.

Investors and their incentives

It is incredibly important to understand the incentives of investors when you are raising money. This used to be fairly easy - you raised money from Venture Capitalists who wanted to see a big return on their investment. The best of these investors were incredibly focused on doing the one thing they did well: investing in technology companies.

The world looks different now. There are a lot of different types of startup investors, each with a different approach to investing and different incentives for doing so. While they’d all prefer not to lose money, the differences in how their incentives and expectations are structured are critical in understanding how to decide whether or not to work with them, how to negotiate with them, and how to interact with them over the lifetime of your relationship.

Below, I've tried to highlight the key motivations and structures of each of the investors you are likely to encounter. While I haven't attempted to catalogue all the ways in which the incentives of investors influence their behavior - sometimes to the detriment of the startup founders - this should work as a starting point to think through those issues.

One big incentive difference that I've chosen not to address here is preferred vs. common shareholders. That creates a whole other set of questions that is out of scope. Assume the investors below will hold preferred shares.

VC Firm
The classic startup investor that is primarily working to achieve returns for the funds they’ve raised from a set of Limited Partners. The partners at these funds are usually paid in two ways: they earn a % of the funds that they’ve raised and they earn a % of the return they make on the total fund from which they invest in you. These firms generally raise further funds based on the performance of earlier investments.

While investors make money off their management fees (the % of what they've raised), the big pay outs only happen when they return a multiple of the total capital raised. Because of how concentrated venture returns are, VCs generally invest only in deals which they believe will return all or most of their returns.

Angel [1]
Usually a wealthy individual who invests their own money in startups. Sometimes these are people that made money through their own startups. Sometimes they are coming from different industries, and sometimes they inherited their wealth. It can be useful to know which category of angel you are talking to.

Angels have become hugely important to early stage startups as the number of companies have grown beyond the financing ability of traditional funds. More importantly, the wide distribution of Angels has materially decreased the access barrier which founders face when raising capital.

There are a lot of different incentives at play for Angels. While many of them are hoping for huge financial returns, many also want to be able to brag about having invested in a hugely successful startup. What's strange is that Angels often just want to brag about investing in startups, regardless of the success of those companies. Investing in startups has become a badge of pride in a lot of circles.

Though most angels invest as a part time activity, there is a wide range in the amount of time and effort they devote to investing. Knowing that they have other things on their minds should change how you approach and interact with them. You also need to understand how experienced they are, the size at which they invest, and how much time and energy they'll actually spend on their investing.

At one end of this spectrum, you'll find angels who are essentially small seed funds. They invest significant amounts of money in startups, and do so frequently. In the best case, these investors have a deep understanding of the companies in which they invest and can be incredibly helpful. These investors understand that they are likely to lose all their committed capital, understand the standard terms and instruments through which startups raise money, and are generally low maintenance.

At the other end of this spectrum, you'll find small, inexperienced investors who just want to invest in startups wherever they find them. These investors will put a small number of small checks into whatever company they can get into. Because they are unfamiliar with how startups work, they are often hard to deal with and incredibly focused on the risk of losing their principal. These are investors you should avoid unless you absolutely need them.

Accelerator
Accelerators are, for the most part, a subset of VC firms where there are professional investors who raise funds from outside parties to invest in startups. They usually fund more companies than classic VC firms, and do so in batches. Similarly to a VC, most accelerators need to demonstrate return to investors to continue to raise funds.

Accelerators also have a non-financial incentive - they need breakout successes to prove that they are providing operational help to companies and actually “accelerating them.”

Syndicate
Syndicates come in a number of different flavors, but are generally groups of angels who come together to invest in a single deal. Usually, these syndicates have a lead who will be your main point of contact. This lead usually puts money at risk for a return and gets some sort of performance fee for the other funds that they bring into the deal.

Watch out for syndicates where the syndicate runner is taking a management fee on funds raised and is encouraging you to raise more money than you think is wise. Also, be careful with syndicates since you don't know ahead of time who will invest in your company, and might find that one of the investors is a direct competitor who just wants some privileged information.

Crowdfunder
Any and all comers on the internet who pre-order an unbuilt item in the hopes that it will some day get built and be cool/solve a problem. They’re not expecting financial return, but do want regular updates on what’s going on with the project and when they can expect delivery. Hiding problems is a lot worse than transparently failing to deliver.

Family and Friends
While your friends and family are probably hoping to get a huge return from their investment, they’re more likely motivated by wanting you to get a chance to succeed and be happy. They are unlikely to try to negotiate terms, though in contrast you will feel the worst for losing their money. Take money from them if you need it and only after making sure they fully and completely understand that they are likely to lose anything they give you. You also need to be okay with losing all of the money someone you care about gives you. If you don’t think this is true, don’t take money from them because it will likely hurt your relationship.

Family office
These are the private investment vehicles for super high net worth individuals and families. Whereas some individuals invest their own money as angels, those that get to a certain scale often employ staffs of portfolio managers and investment professionals. There are many different structures here. Some family offices are structured like single limited parter hedge funds with a high tolerance for risk, and others are structured more conservatively. Whatever the risk tolerance, the staffs of family offices generally get carry on their investments as well as salary, which introduces some of the incentive dynamics present at VCs.

Generally, these entities are very concerned with not losing their principal. Whereas a VC fund that loses an entire fund will have a hard time raising other funds, a family office that loses all its principal has no recourse for more funds. Generally, if you get an investment from a family office, it will come from a small portion of a portion of an overall investment portfolio.

Corporate investor (direct)
There are a lot of corporations that like to talk about investing in startups. Some of them actually do this, and some do not. When an investment comes directly from the company's balance sheet at the direction of a particular business line, the corporation is usually looking for strategic value from the investment. Most of these companies know that investing in startups is unlikely to change the valuation of the investor.[2]

This means they either want an inside edge to acquire you at some point, or believe that investing in you will help improve their bottom line. They may want to prevent you from selling to competitors in the same space, and may have other confusing and onerous ideas. This is because their incentives are different that those of most startups - they are more concerned with how you can help them than with how they can help you get gigantic.

Corporate investor (venture arm)
While corporate venture arms have some of the misaligned incentives of direct corporates, they generally have a mandate to generate financial returns for the company's balance sheet. This means they act more like VCs, and are typically more conversant with how startups work.

Government
Governments have many reasons for investing in startups. There are many government grants available in various countries that are designed to promote startups as way to increase job growth. There are also government agencies with their own venture funds, generally designed to fund technology that will help the government in the long run.

University endowment
University endowments are similar to family offices, but they represent an endowment. They are designed to produce returns to fund the university over time. These entities do not generally invest in early stage companies except in conjunction with a fund with whom they have a strong relationship.

Seed fund
Seed funds are VC funds that write smaller checks at early stages of companies. They usually have the same incentives and structures.

Hedge fund
Hedge funds are largely unrestricted pools of capital. Traditionally, these were focused on public market investments, though in the last few years have started investing in startups. Generally, they invest in later stages and are looking for returns on capital as they have LPs and similar incentive structures to VC funds.

Mutual fund
These are large pools of capital run by portfolio managers. They don't have LPs, rather they have large groups of retail investors who buy shares in the funds, which capital they can then deploy. These funds only invest in late stage startups, because they need to deploy a lot of capital to have any kind of impact at the portfolio level. These managers are paid based on performance, and often have reporting requirements that cause them to publish their internal marks for private companies. This has caused a lot of consternation lately as Fidelity has been publishing widely oscillating valuations for a number of companies like Dropbox.

Sovereign wealth fund
These are the largest pools of capital in the world, and are essentially very large family offices for entire countries. These funds are large enough to invest in any and all asset classes that the managers believe will produce a return on investment. Like mutual funds, these funds rarely invest directly in startups - they are more likely to invest in funds that do.

In the last few years, however, a number of them have begun to invest directly in startups. While the managers of these funds are generally paid on a performance basis, there are a lot of other complicated incentives in place at certain funds deriving from political requirements. These are pretty hard to parse.


__
[1] I don't actually love the term Angel as it seems insufficiently precise given the wide range of investors to whom it applies. While it would probably make sense to find other terms for subgroups, the key thing to know is that. Maybe we'll try to rename them down the road.
[2] Yahoo's investment in Alibaba is a rare exception to this rule.

Thanks to Paul Buchheit, Geoff Ralston, Daniel Gackle, and Dalton Caldwell for your help writing this.