When to Shut Down Your Company

I originally wrote this essay four years ago and published it over at YC's blog at the time. A number of recent conversations I've had with founders made me realize that the framework below remains useful. As the market continues through its downswing, more founders will face difficult decisions about how long and how far to push their startups. This is as emotionally fraught as ever, so I hope that resurfacing this essay will help a few people through the decision making process


It makes sense that founders and investors spend so much time talking about things that go well. If we spent all of our time dwelling on the companies that failed, we wouldn’t have time for much else.

When people do talk about company failure, they often do so in a way meant to make them seem wise by breaking down all the lessons they’d learned through failing. I did something like this when we shut Tutorspree down. I think that was a valuable exercise, and maybe it even helped some people. Mostly, though, it was cathartic.

Founders lack a coherent way to think about when to shut down.[1] Founders do not always get to choose to shut down.[2] However, most of the time, it is the founder’s choice. It’s a personal decision. It’s a hard and painful decision. It’s an emotional, fraught decision. However, shutting down doesn’t have to be a blind decision.

The unintuitive thing about figuring out if you should shut down your company is that it isn’t the path of least resistance. The “easiest” thing to do for a struggling company is to fall into zombie mode – neither growing nor truly dead. This is easy because it doesn’t require an active decision, it just involves continuing to do the bare minimum to keep the company alive. This involves a series of seemingly small compromises that lead to stasis or failure.

Shutting down is hard because it means publicly admitting that you were wrong, unlucky, or incompetent.[3] Because of this difficulty, we’ve evolved a set of terms that often mean “shut down” without saying “shut down.” In no particular order these are: pivot, hard pivot, rebrand, strategic shift, change customer focus, and platform switch.[4]

Stakeholders

Shutting down is hard because it generally means disappointing people who believed in you. Some of this disappointment is real, some of it is imagined. Founders who have raised money are usually most concerned about disappointing or upsetting their investors. Investors are often upset when a bet fails, generally in proportion to the amount of capital they’ve given you, and in the speed with which things went wrong since they invested. However, investors know that most bets are going to fail, and will usually get over their emotions. The thing that bothers investors more than shutting down is when founders either misrepresent the status of a business and suddenly shut down without warning, or slowly bleed out a business over years while taking up a lot of the investor’s time.

Founders often worry that shutting down will disappoint their customers. This is true, but companies that aren’t doing well generally see their products degrade, which also upsets customers. Unless your product provides a critical, lifesaving purpose, shutting down cleanly and with transparency is much preferred to a slow fade into obsolescence.[5]

Founders should worry about the impact that shutting down has on their employees. The biggest burden a founder has is to meet payroll. The biggest emotional investment that founders make – especially early on – is convincing great people to take a leap of faith and accept an offer to work their butts off on a long shot. This dynamic is why transparency around the decision to shutdown and the timeline of it is so important.

The worst thing a founder can do to an employee is to tell them that everything is amazing, and then to one day tell them that she a) no longer has a job and b) that the company cannot pay her what was promised and c) that you’ve known all this for a long time but didn’t want to tell her because you were worried about her feelings. It is much better to make a clean decision to shut down in full view of of your employees with enough time and money left to help employees find new jobs and move on.

A Framework for Shutting Down

The decision to actually shut down an idea can be made through answering the following questions:[6]

  1. Do you have any ideas left to grow your startup?
  2. Can you drive that growth profitably?
  3. Do you want to work on the startup that results from that growth?
  4. Do you want to work with your co-founders on the startup that results from that growth?

The first two questions here are quantitative and driven by the experiments that founders are constantly running on their products and users. Unfortunately, while they are quantitative, there’s no deterministic way to know if you’ve actually tried everything in the right way. At some point, you have to make a call based on accumulated evidence.[7]

The second two are qualitative and introduce the largest challenges around making a decision comes when the answers to these questions are mixed between yes and no. For instance – if you enjoy working with your co-founders on a business that produces positive cash flows but is not growing, you probably shouldn’t shut down. However, if you hate the people you’re working with on a business that is growing rapidly, you should either shut it down, or find a way to keep going without the team. These are personal decisions that no one else can make for you, but it is important to separate the different threads of the decision so that you can actually evaluate them as objectively as possible.

The Other Side

I was terrified to shut my company down, even though all the evidence said it was the right decision.[8] The thing that scared me most was the that I had no idea what would happen afterwards. However, I realized that I had exactly one life to live, and that staying in a bad situation of my own making out of fear was a dumb thing to do.

Having spoken with many founders who have gone through similar situations, I’ve found that most of them are incredibly depressed after shutting down their startups for a few months. Afterwards, things start to get better. This may be hard to believe but shutting down releases so much tension and clears away so many burdensome expectations that it allows for brand new creative ideas and approaches. It’s a necessary part of the startup cycle, and one that founders need to approach more openly and rigorously.

__

Notes
[1] Throughout this essay, I use “shut down” interchangeably for ideas and companies. The two differ only slightly.

[2] Running out of money unexpectedly, often as a result of a failed fundraise, can force a founder to shut down.

[3] Incompetent seems harsh, but it would be strange to believe that you could be competent at all the new things required in building a company on the first or even tenth try.

[4] I am sure this is a non-exhaustive list.

[5] In this case, shutting down is still critical, but the decision should involve a long timeline, over communication, and a best effort to find alternatives.

[6] I could also state this as the inverse. Deciding to work on a startup should come down to these questions all being answered with “yes.”

[7] It’s important to make sure that you’ve actually tried a lot of things. Founders often quit far too early on an idea that needs more time.

[8] The answers to questions 2, 3, and 4, were all “no.” The good news is that Ryan, Josh, and I all like each other as humans.

Thanks to Sam Altman and Craig Cannon for reading drafts of this way back.

VCs are Scared When They Should be Greedy

I originally wrote this essay for The Information, where it was published on July 6, 2022.

I've read a lot of the advice that venture capitalists have given founders on the current state of the markets. Most of that advice focuses on how founders need to adjust to survive the deteriorating conditions—cutting cash burn by firing underperforming employees, slowing hiring, shrinking marketing budgets and so on. But this advice is only half the story.

Founders cannot afford to simply survive. They need to build companies that attract investment. That may seem impossible right now while the  markets are collapsing, but venture capitalists still have billions of dollars sitting in just-raised funds that they need to invest. With the exception of crossover funds that have lost massive amounts of money and are nervous about losing limited partners, smart investors are not winding down new funds or reducing fees—they’re taking meetings with founders. For VCs, now is a great time to find deals.

There’s never been more cash available for startups, but the people holding that cash (investors) are telling founders there’s no money available for investing. Founders need to understand the logic behind this contradiction and how to operate within it, because rapidly growing companies still (usually) need to raise capital to build their businesses.

The fact that the markets are bad doesn’t mean founders get to hide. The opposite is likely true—when markets are bad, the best founders will find ways to raise the capital they need to exit the turmoil in a dominant position. And if history is any guide, it is precisely during this downturn that we’ll see the emergence of a new class of tech powerhouses.

Let’s be clear: The current situation is entirely different from the market routs of 2000, 2008 or 2020.

In 2000, the Nasdaq superheated due to the large number of companies that skyrocketed into the public markets fueled by fanciful metrics disengaged from revenue. When that bubble burst, it did so because it became clear, almost all at once, that a lot of the emperors had no clothes…or business models. Critically, the venture market at the time was tiny relative to today’s ecosystem.

Then came 2008, when I got to watch the market collapse from the relatively safe vantage point of hedge fund Bridgewater Associates. When the mortgage-backed security market went bust, investors realized they had no idea what they were actually holding or what those holdings were worth, and didn’t know who they could trust to tell them. As the damage rippled outward, it triggered the implosion of multiple major financial institutions. The full impact of those collapses was difficult to understand because of the tangled web of leverage and counterparty agreements that made the financial system run…which is also what caused it to collapse.

Then came the coronavirus-related market shock of 2020. That was a strange one. It felt like the start of a war. For about a month, everyone assumed the absolute worst. And then the markets—public, private and everything in between—went nuts. New initial public offerings flooded the market and share prices jumped 100% in their first days of trading, while private companies raised ever larger private rounds from investors afraid to miss the next $100 billion company. This flywheel kept running until sometime around the beginning of this year, when it all ground down to a halt.

The basic logic of this slowdown is uncertainty. Growth is slowing, and we may be headed for a recession. At the same time, the Federal Reserve Board is hiking interest rates to try to control inflation. When interest rates go up, venture capitalists have to adjust the models they use to value portfolio companies. These models are designed to produce a current value for future cash flows. The interest rate is used to discount the value of those cash flows, so for tech companies—for whom most of the cash flow is far into the future—small interest rate changes produce huge changes in value. The faster this happens, or the more uncertainty there is, the more chaos there is around pricing.

Here’s what isn’t uncertain: whether investors are suddenly going to run out of money. In fact, the opposite is true. Venture capitalists are sitting on more cash than they’ve ever had. Sure, some of them are scared because they might not see much of a return on their most recent deals, and general partners are also feeling queasy about just how much their public stock portfolios have pulled back. But those are temporary conditions. In contrast with the scenario in 2000, most of today’s tech companies are real businesses. They’re simply carrying extremely rich valuations based on a worldview that suddenly looks outdated. It’s only a matter of time before investors shake off their fear and start doing deals again.

True, a few things need to happen before venture capitalists loosen their grip on their wallets again. First, some of their portfolio companies now either have new capital needs or are suddenly willing to take dilution to create a larger margin of safety going forward. That will take more of investors’ capital reserves than they may have initially planned, and it will for sure eat into the capital they have available for new deals, which will make them hesitate. This is, however, a finite exercise. Once venture capitalists have done the math on how much they need to hold back, they’ll know how much they can deploy. While there may be less money available than six months ago, the overall pie is so much bigger than it used to be that this shouldn’t actually matter.

The next thing investors have to worry about is how to price deals. This strikes me as a bit funny because there has never been a science to pricing a venture deal. That said, fear is contagious, and no matter how distant the connection is between public market pricing and private market investing, it works its way back, like so:

  1. The public market investor gets hammered.
  2. The crossover investor sees that, gets scared and becomes unsure how to price a deal.
  3. The late-stage investor sees fear in the crossover investor, the early-stage investor sees fear in the late-stage investor, and suddenly everyone is scared…
  4. …except seed investors, who have always been buying lottery tickets and are likely to keep on gambling.

For this backlog to clear, individual investors need to see someone go first. A venture capitalist won’t lose their job by sitting on the sidelines and tweeting about fiscal responsibility, but they certainly could by sending a term sheet for a company that quickly does a down round.

This is where founders need to get smart. Investors know that the general market paralysis gives them—for the time being, at least—more leverage and more time than they’ve enjoyed in quite a while. A founder willing to play into an investor’s need for certainty is a founder that’s going to come away with a deal.

That means, above all, planning for the worst—not simply figuring out a way for their business to limp along indefinitely, but rather understanding how much they can grow and how long they can last with the cash they have on hand.

Once founders have a sense of that, they need to invest more heavily in their relationships with investors. This has always been important, and now it is even more so. Early-stage investing is as much an emotional decision as it is a rational economic one. Investors are more likely to back founders they know and trust, people they’ve seen execute over time. Founders who work with this dynamic will be better positioned to raise capital when they need it.

That said, founders also need to lower their expectations—about the amount of money they can raise, the price at which they can do it and the amount of time they’ll need to complete a round. Rounds are going to occur further apart, and fundraising processes are going to take longer to complete until the market changes again. And make no mistake, the market will change again, because that’s how cycles work. They go down, they go up.

Despite the doom and gloom, there’s money available for good startups who tell good stories to good investors. That’s harder to do now than it was, but it may also lead to bigger and better things. Fundraising has never been a sure thing. It just felt that way based on the press coverage and the stories you probably heard from other founders. The current market is just reminding us of that.

In defense of exploding term sheets

There’s a lot of content out there arguing that exploding term sheets—ones that expire after a set amount of time, whether or not they’re accepted—are bad. Some of this comes from founders who do not appreciate the pressure; some if it comes from VCs who want to emphasize their own founder-friendliness. The reality, though, is that all term sheets explode. The only variable is when.
 
I’ll even go one step further: If term sheets didn’t explode, fundraising would be much harder for founders. During a well-run fundraising process, founders are largely in control. They can dictate when to pitch and when to release information. They’re the ones who choose if and when, ultimately, to sign an offer. At the extreme, this can produce comically narrow demands: Pitch Monday, offers Tuesday, decision on Friday. (To be fair, sometimes this works.) Venture capitalists are left in control of one thing: whether and when to say yes or no.
 
Now, in fully liquid or public markets, investors can take as long as they want to decide whether to buy or sell—and if the fundamentals change in the interim, well, they can go on their merry way. But in VC rounds, the default assumption is that, once a VC says yes to an investment, they can’t take it back.
 
This is kind of weird.
 
Term sheets aren’t actually contracts, and yet they’re considered to be (mostly) binding. That means once an investor makes an offer, they’re exposed to two types of risk: first, of course, that their investment turns out to be a bad one, but also the indefinite opportunity cost of spending a lot of time working on, thinking about, and allocating funds to a deal they might not win.
 
If we multiply that second issue by the number of deals an investor might want to do, you start to run into an even bigger problem. Investors can only evaluate or commit to a certain number of deals at a time. An investor who cannot control the lifespan of her offers would quickly run out of bandwidth to evaluate new deals. This would be bad for everyone. If an investor is willing to make a bet, she should be able to know when to expect an answer.
 
Process uncertainty also creates issues for founders, who have a right to know when to expect a yes or no. But once an offer is made, founders would prefer if it stayed open until they can make a decision, however long that takes. VCs prefer an offer that expires relatively quickly.
 
To me, it certainly seems as if both sides want something reasonable. What’s missing is a common set of rules to determine what’s fair. So, here’s what I propose: Founders should have at least as long to decide whether to accept the terms as the VC took to decide to make an offer. I'll explain.
 
If an investor makes a decision to invest within 24 hours of the start of a fundraising process, then a term sheet with a 24-hour clock seems reasonable. Think of it as a bonus for VCs who make decisions quickly. Their time in uncertainty is small, so they should be able to process more deals in less time than slower moving funds.
 
At the other end of the spectrum, investors who need more time to deliberate should extend founders the same courtesy. The VC could even cast this as an advantage—“Yes, we take time to get to know you and think deeply and stuff. But when we make a decision, we make it with conviction, and we will romance you and woo you and convince you, dear founder, to take our offer.”
 
VCs could also publicly commit to a certain timeline. “Once we meet you for a pitch meeting, we’ll make a decision within a week. Once we give you a term sheet, you’ll have a week to decide.” A day on each side could work. A month on each side could work. As long as both sides understand the terms ahead of time, the trade is even.
 
From a founder’s point of view, too, all of this should seem fair. And by that same token, if a founder needs an offer by a specific date or time, they should return a decision on that same time scale.
 
That said, negotiations are messy, and the reciprocity principle won’t work perfectly every time. Sometimes a founder will push the deadline too hard and scare everyone off. Sometimes a VC will do the same. But fun fact: exploding term sheets don’t literally explode. A VC can always decide to extend another offer if they’re still interested after their deadline has passed, or even ignore the deadline altogether. This is simply a negotiating tactic.

The hot round fallacy

I originally wrote this essay for The Information, where it was published on Feb 16.

Here's a fun little paradox in fundraising: Founders and the media love to talk about how quickly funding rounds happen these days—think about the big, rapid-fire raises from OpenSeaDataBricks and others—but when you watch a fundraising process from start to finish, it’s never as fast as it seems. Even the tidiest-looking processes, ones that seem to kick off on day X and sign term sheets on day Y, aren’t what they appear to be.

The best-run rounds are always running, just at different levels of intensity. This breaks some brains because founders want to believe you’re either fundraising or you’re not. In reality, great founders are always fundraising. It’s just that sometimes they call it “fundraising” and sometimes they call it “building relationships.”

This is an important point for founders to internalize because the illusion of fast, aka hot rounds hurts companies. Specifically, founders come to expect that they can emerge from a hole, say “We’re raising” and be done in a few days.[1] The early-stage market hasn’t taken quite the same beating as later-stage or public markets, but still, now is no time to be stupid. The truth is that nearly every fast round I've ever seen (and I’ve seen many) actually started months earlier with careful relationship- and narrative-building or extensive work by an investor—or in many cases both.

Let’s look at this from the investor’s perspective. Every venture capitalist has some quality bar a company needs to clear to trigger an investment. That usually consists of two key components: the overall market or opportunity, and the founder and company in the context of that opportunity. While different investors may consider these in different orders or with different weights, they all form views on each.

Developing a feel for a particular market requires research. That might involve commissioning a consulting firm[2] to perform due diligence on a particular region and its regulatory structure, or it may mean simply talking to potential users and building a view on what solutions may or may not work. This can take weeks or months, most of it before an investor ever talks to the founder of a prospective investment. Some investors run this process quickly, some run it slowly, and to be sure, some don’t run it at all.[3]

But just because a venture capitalist has all the information necessary to make a quick investment decision doesn’t mean they will—there’s still the company to consider. Here again, we run into different speeds and philosophies. When I worked at Y Combinator, we made most of our decisions in 10 minutes. Other investors need hours with founders, often distributed over months or years. Once a venture capitalist has these pieces, the question becomes whether the price of the deal makes sense.

When you understand this framework, the pace at which many follow-on rounds seem to occur doesn’t look quite so fast. These (mostly) aren’t discrete fundraising decisions. They’re continuations of situations where an investor did all or most of the work to get to a “yes” at a given round but lost, then reevaluated the price, made another decision to say “yes” and convinced the founder to take more money.

Founders can use this understanding to their advantage by realizing that they do not have to be passive bystanders in the investor knowledge-gathering process. Think about it this way: An investor evaluating a market needs information. That investor needs to talk to people. Founders are people who happen to know quite a lot about their markets and also generally like talking about them, and investors are usually happy to listen![4] As they listen, they can’t help but be influenced by a convincing story and growing metrics.

That said, founders have to be smart about what information they share and when they do it.

Those who spend too much time with investors and who share too much information too often inevitably hurt their companies. It can be hard for founders to see the balance here, so as someone who’s seen many rounds happen,[5] I’d offer these rules of thumb:

  1. Meet with investors who can actually make the decision to lead a round. This is not necessarily someone with the title of partner.
  2. Meet with those investors once per quarter—maybe once a month in extreme cases.
  3. Share high-level metrics, but not enough to let someone form a full picture.
  4. Prepare concrete asks that can’t be executed during the conversation, such as: “Here are three portfolio companies that would be good customers. Can you introduce me to the heads of analytics at each?”
  5. Track follow-ups.

Maybe someone wants to make the semantic argument that this isn’t fundraising. There’s no deck, there’s no ask, there’s no official invite to a fundraising kickoff party.[6] But it is a meeting between someone who wants to provide funding in exchange for equity and someone who is willing to sell equity for money. If that isn’t fundraising, I don’t know what is. When founders run this part of the process effectively, they’re able to target their later, formal fundraising efforts solely at investors they know are interested.

From there, things should move along at a nice clip, in no small part because the founder is now effectively running an auction. I’ve written extensively about this in the past, so I won’t rehash all of it here, but the main point is that competition forces investors to move quickly. This is in part because good investors generally know when other good investors are interested, but it’s also about the way founders act when they have options versus when they don’t. This is hard to fake, and it works.

I’ve seen founders attempt to force a quick process by setting arbitrarily short timelines, by refusing to produce reasonable data or answer questions, by generally being jerks—all things they believe are effective based on what they’ve read, not based on the work they’ve put in. Startup orthodoxy confirms what these founders want to believe: that they can keep their heads down pretty much forever and get chased by money whenever they need it. Maybe that was true at one time. But markets change, and founders who adapt reap the benefits.

__

Thank you to Jillian Goodman for your help writing this.

__

[1] We’re focused here on priced rounds: Series A, Series B and up. Seed rounds are different.

[2] This is a strategy for rapidly building knowledge on many opportunities that I first saw happening at Tiger Global Management and that other funds have adopted.

[3] There are plenty of investors who operate on gut, but this happens less and less as you move through Series A to B and up.

[4] Venture capitalists like to have meetings. This is a major part of their jobs. Go have a meeting! They’ll listen for longer than your parents and you will get free coffee out of it.

[5] When asked this question, I generally come back to Potter Stewart’s famous quote from a distinctly different context: “I know it when I see it."

[6] Not a thing. At least not yet.

A framework for choosing what to do

A year ago, I left YC and felt a bit lost about what to do next.[1] A lot of people were pretty sure they knew what I should do. But...I didn't and most of their suggestions didn’t excite me.

As I thought through my possible paths, I realized that I didn't have a useful framework for making a decision. My gut wasn't helpful since too many things seemed interesting. I started asking friends for advice, and Henrik Werdelin gave me the structure I needed to figure out what I should try next.

As I worked through his advice and its implications, something odd happened - a whole lot of my friends started hitting the same confusing decision point through which I’d just gone. Maybe we're all going through natural career evolutions, maybe these are early onset midlife crises, and maybe a few years of inescapable existential dread about a never ending pandemic is just screwing up all our preconceived notions about how our lives should/need to work.

Whatever the case, I keep having the same conversation, so writing it out seems useful.[2]

The 1st, and most important thing to keep in mind is that no career decision is perfect, no decision without risk. Oh, and no one actually cares what you do. This is especially hard for high achievers to internalize, harder still for people who have succeeded publicly. Maybe this is hardest of all for people who feel as if nothing they've done has matched their self perception and internal expectations.

The next thing to do is to start making lists. In particular, three lists:

The 1st list is an amalgam of all the qualities you want in your next role and what you get from having that role. This is expansive by design. For instance:

  • Work with people who are actually funny

  • Easy commute

  • Do not work weekends

  • Earn at least x

  • Coworkers are the same in person as on social media

  • Build on personal expertise

  • Creates interesting future options

  • Other stuff

It is ok to have items on here that are intensely personal and maybe unexpected. Perhaps you want less risk or ambition than you've shown in the past. Maybe you know that you’ll never be happy with a boss. Write it all out.

List two is all the jobs that you think sound interesting. Try to go way beyond your existing set of experiences to things you always thought were interesting but never seriously explored. You're going to throw most of this out, but that's ok. You could write:

  • Get a PhD is American History

  • Run for office

  • Join a government agency

  • Start a software company

  • Be a baker

  • Hedge fund - work at or start

  • Start a nonprofit

Obviously this list will be informed by what you know and what you want. Some of it will be impractical. That's ok.

The final list is made up of people you know and with whom you'd actually work. Sometimes this list is long and sometimes it is short. There's no right answer here.

Once you have these lists, you start cross referencing them to see what mix of qualities, roles and people seem like the best trade.

There's no science here. You could decide that salary is the most important criteria of all and use that as the single lens. Usually, the balance is more nuanced. What helped me most was forcing myself to list all the options, interests, and criteria. I'd never done that and it required me to think more deeply about what I did and did not want.

This process took me some time, and I ended up choosing a set of what I enjoyed in a way that capitalized some unfair advantages I'd built in my career. I had come to an answer, but I was still too scared to try it. At that point, another friend, Omri Dahan, gave me the final piece I needed. He pointed out that I was still trying to optimize for long term perfection, and suggested that I should, instead, take a risk for six months. I hadn’t internalized the idea that very few career decisions are permanent. I thought I was choosing the job I’d have for the rest of my life, but I wasn’t. I was choosing the next thing to learn about and try. I was fortunate to be able to run that experiment. Eight months later, it's become clear to me that I made the right choice from the perspective of how much I've been able to learn and build. Perhaps more important to me, the problem set keeps getting more complex and interesting. Still, though, I know this is just another experiment that I’ll run it until it no longer makes sense.

__

[1] I'll note that "what to do" and "job" here aren't necessarily the right terms. I'm using them as stand ins for the the thing you do that isn't family/personal life. Some people will go through this exercise and decide that there's nothing for them outside of personal/family life. I'm happy for them, but my wiring needs to have a balance of family and work.

[2] By writing from personal experience, I know that this process won't work for everyone and every situation. But if you find it useful, great!