Paul Graham's Essays

Apple's Mistake

November 2009

I don't think Apple realizes how badly the App Store approval process is broken. Or rather, I don't think they realize how much it matters that it's broken.

The way Apple runs the App Store has harmed their reputation with programmers more than anything else they've ever done. Their reputation with programmers used to be great. It used to be the most common complaint you heard about Apple was that their fans admired them too uncritically. The App Store has changed that. Now a lot of programmers have started to see Apple as evil.

How much of the goodwill Apple once had with programmers have they lost over the App Store? A third? Half? And that's just so far. The App Store is an ongoing karma leak.

* * *

How did Apple get into this mess? Their fundamental problem is that they don't understand software.

They treat iPhone apps the way they treat the music they sell through iTunes. Apple is the channel; they own the user; if you want to reach users, you do it on their terms. The record labels agreed, reluctantly. But this model doesn't work for software. It doesn't work for an intermediary to own the user. The software business learned that in the early 1980s, when companies like VisiCorp showed that although the words "software" and "publisher" fit together, the underlying concepts don't. Software isn't like music or books. It's too complicated for a third party to act as an intermediary between developer and user. And yet that's what Apple is trying to be with the App Store: a software publisher. And a particularly overreaching one at that, with fussy tastes and a rigidly enforced house style.

If software publishing didn't work in 1980, it works even less now that software development has evolved from a small number of big releases to a constant stream of small ones. But Apple doesn't understand that either. Their model of product development derives from hardware. They work on something till they think it's finished, then they release it. You have to do that with hardware, but because software is so easy to change, its design can benefit from evolution. The standard way to develop applications now is to launch fast and iterate. Which means it's a disaster to have long, random delays each time you release a new version.

Apparently Apple's attitude is that developers should be more careful when they submit a new version to the App Store. They would say that. But powerful as they are, they're not powerful enough to turn back the evolution of technology. Programmers don't use launch-fast-and-iterate out of laziness. They use it because it yields the best results. By obstructing that process, Apple is making them do bad work, and programmers hate that as much as Apple would.

How would Apple like it if when they discovered a serious bug in OS X, instead of releasing a software update immediately, they had to submit their code to an intermediary who sat on it for a month and then rejected it because it contained an icon they didn't like?

By breaking software development, Apple gets the opposite of what they intended: the version of an app currently available in the App Store tends to be an old and buggy one. One developer told me:

As a result of their process, the App Store is full of half-baked applications. I make a new version almost every day that I release to beta users. The version on the App Store feels old and crappy. I'm sure that a lot of developers feel this way: One emotion is "I'm not really proud about what's in the App Store", and it's combined with the emotion "Really, it's Apple's fault."

Another wrote:

I believe that they think their approval process helps users by ensuring quality. In reality, bugs like ours get through all the time and then it can take 4-8 weeks to get that bug fix approved, leaving users to think that iPhone apps sometimes just don't work. Worse for Apple, these apps work just fine on other platforms that have immediate approval processes.

Actually I suppose Apple has a third misconception: that all the complaints about App Store approvals are not a serious problem. They must hear developers complaining. But partners and suppliers are always complaining. It would be a bad sign if they weren't; it would mean you were being too easy on them. Meanwhile the iPhone is selling better than ever. So why do they need to fix anything?

They get away with maltreating developers, in the short term, because they make such great hardware. I just bought a new 27" iMac a couple days ago. It's fabulous. The screen's too shiny, and the disk is surprisingly loud, but it's so beautiful that you can't make yourself care.

So I bought it, but I bought it, for the first time, with misgivings. I felt the way I'd feel buying something made in a country with a bad human rights record. That was new. In the past when I bought things from Apple it was an unalloyed pleasure. Oh boy! They make such great stuff. This time it felt like a Faustian bargain. They make such great stuff, but they're such assholes. Do I really want to support this company?

* * *

Should Apple care what people like me think? What difference does it make if they alienate a small minority of their users?

There are a couple reasons they should care. One is that these users are the people they want as employees. If your company seems evil, the best programmers won't work for you. That hurt Microsoft a lot starting in the 90s. Programmers started to feel sheepish about working there. It seemed like selling out. When people from Microsoft were talking to other programmers and they mentioned where they worked, there were a lot of self-deprecating jokes about having gone over to the dark side. But the real problem for Microsoft wasn't the embarrassment of the people they hired. It was the people they never got. And you know who got them? Google and Apple. If Microsoft was the Empire, they were the Rebel Alliance. And it's largely because they got more of the best people that Google and Apple are doing so much better than Microsoft today.

Why are programmers so fussy about their employers' morals? Partly because they can afford to be. The best programmers can work wherever they want. They don't have to work for a company they have qualms about.

But the other reason programmers are fussy, I think, is that evil begets stupidity. An organization that wins by exercising power starts to lose the ability to win by doing better work. And it's not fun for a smart person to work in a place where the best ideas aren't the ones that win. I think the reason Google embraced "Don't be evil" so eagerly was not so much to impress the outside world as to inoculate themselves against arrogance. [1]

That has worked for Google so far. They've become more bureaucratic, but otherwise they seem to have held true to their original principles. With Apple that seems less the case. When you look at the famous 1984 ad now, it's easier to imagine Apple as the dictator on the screen than the woman with the hammer. [2] In fact, if you read the dictator's speech it sounds uncannily like a prophecy of the App Store.

We have triumphed over the unprincipled dissemination of facts.

We have created, for the first time in all history, a garden of pure ideology, where each worker may bloom secure from the pests of contradictory and confusing truths.

The other reason Apple should care what programmers think of them is that when you sell a platform, developers make or break you. If anyone should know this, Apple should. VisiCalc made the Apple II.

And programmers build applications for the platforms they use. Most applications—most startups, probably—grow out of personal projects. Apple itself did. Apple made microcomputers because that's what Steve Wozniak wanted for himself. He couldn't have afforded a minicomputer. [3] Microsoft likewise started out making interpreters for little microcomputers because Bill Gates and Paul Allen were interested in using them. It's a rare startup that doesn't build something the founders use.

The main reason there are so many iPhone apps is that so many programmers have iPhones. They may know, because they read it in an article, that Blackberry has such and such market share. But in practice it's as if RIM didn't exist. If they're going to build something, they want to be able to use it themselves, and that means building an iPhone app.

So programmers continue to develop iPhone apps, even though Apple continues to maltreat them. They're like someone stuck in an abusive relationship. They're so attracted to the iPhone that they can't leave. But they're looking for a way out. One wrote:

While I did enjoy developing for the iPhone, the control they place on the App Store does not give me the drive to develop applications as I would like. In fact I don't intend to make any more iPhone applications unless absolutely necessary. [4]

Can anything break this cycle? No device I've seen so far could. Palm and RIM haven't a hope. The only credible contender is Android. But Android is an orphan; Google doesn't really care about it, not the way Apple cares about the iPhone. Apple cares about the iPhone the way Google cares about search.

* * *

Is the future of handheld devices one locked down by Apple? It's a worrying prospect. It would be a bummer to have another grim monoculture like we had in the 1990s. In 1995, writing software for end users was effectively identical with writing Windows applications. Our horror at that prospect was the single biggest thing that drove us to start building web apps.

At least we know now what it would take to break Apple's lock. You'd have to get iPhones out of programmers' hands. If programmers used some other device for mobile web access, they'd start to develop apps for that instead.

How could you make a device programmers liked better than the iPhone? It's unlikely you could make something better designed. Apple leaves no room there. So this alternative device probably couldn't win on general appeal. It would have to win by virtue of some appeal it had to programmers specifically.

One way to appeal to programmers is with software. If you could think of an application programmers had to have, but that would be impossible in the circumscribed world of the iPhone, you could presumably get them to switch.

That would definitely happen if programmers started to use handhelds as development machines—if handhelds displaced laptops the way laptops displaced desktops. You need more control of a development machine than Apple will let you have over an iPhone.

Could anyone make a device that you'd carry around in your pocket like a phone, and yet would also work as a development machine? It's hard to imagine what it would look like. But I've learned never to say never about technology. A phone-sized device that would work as a development machine is no more miraculous by present standards than the iPhone itself would have seemed by the standards of 1995.

My current development machine is a MacBook Air, which I use with an external monitor and keyboard in my office, and by itself when traveling. If there was a version half the size I'd prefer it. That still wouldn't be small enough to carry around everywhere like a phone, but we're within a factor of 4 or so. Surely that gap is bridgeable. In fact, let's make it an RFS. Wanted: Woman with hammer.

Notes

[1] When Google adopted "Don't be evil," they were still so small that no one would have expected them to be, yet.

[2] The dictator in the 1984 ad isn't Microsoft, incidentally; it's IBM. IBM seemed a lot more frightening in those days, but they were friendlier to developers than Apple is now.

[3] He couldn't even afford a monitor. That's why the Apple I used a TV as a monitor.

[4] Several people I talked to mentioned how much they liked the iPhone SDK. The problem is not Apple's products but their policies. Fortunately policies are software; Apple can change them instantly if they want to. Handy that, isn't it?

Thanks to Sam Altman, Trevor Blackwell, Ross Boucher, James Bracy, Gabor Cselle, Patrick Collison, Jason Freedman, John Gruber, Joe Hewitt, Jessica Livingston, and Robert Morris for reading drafts of this.

What Startups Are Really Like

October 2009

(This essay is derived from a talk at the 2009 Startup School.)

I wasn't sure what to talk about at Startup School, so I decided to ask the founders of the startups we'd funded. What hadn't I written about yet?

I'm in the unusual position of being able to test the essays I write about startups. I hope the ones on other topics are right, but I have no way to test them. The ones on startups get tested by about 70 people every 6 months.

So I sent all the founders an email asking what surprised them about starting a startup. This amounts to asking what I got wrong, because if I'd explained things well enough, nothing should have surprised them.

I'm proud to report I got one response saying:

What surprised me the most is that everything was actually fairly predictable!

The bad news is that I got over 100 other responses listing the surprises they encountered.

There were very clear patterns in the responses; it was remarkable how often several people had been surprised by exactly the same thing. These were the biggest:

1. Be Careful with Cofounders

This was the surprise mentioned by the most founders. There were two types of responses: that you have to be careful who you pick as a cofounder, and that you have to work hard to maintain your relationship.

What people wished they'd paid more attention to when choosing cofounders was character and commitment, not ability. This was particularly true with startups that failed. The lesson: don't pick cofounders who will flake.

Here's a typical reponse:

You haven't seen someone's true colors unless you've worked with them on a startup.

The reason character is so important is that it's tested more severely than in most other situations. One founder said explicitly that the relationship between founders was more important than ability:

I would rather cofound a startup with a friend than a stranger with higher output. Startups are so hard and emotional that the bonds and emotional and social support that come with friendship outweigh the extra output lost.

We learned this lesson a long time ago. If you look at the YC application, there are more questions about the commitment and relationship of the founders than their ability.

Founders of successful startups talked less about choosing cofounders and more about how hard they worked to maintain their relationship.

One thing that surprised me is how the relationship of startup founders goes from a friendship to a marriage. My relationship with my cofounder went from just being friends to seeing each other all the time, fretting over the finances and cleaning up shit. And the startup was our baby. I summed it up once like this: "It's like we're married, but we're not fucking."

Several people used that word "married." It's a far more intense relationship than you usually see between coworkers—partly because the stresses are so much greater, and partly because at first the founders are the whole company. So this relationship has to be built of top quality materials and carefully maintained. It's the basis of everything.

2. Startups Take Over Your Life

Just as the relationship between cofounders is more intense than it usually is between coworkers, so is the relationship between the founders and the company. Running a startup is not like having a job or being a student, because it never stops. This is so foreign to most people's experience that they don't get it till it happens. [1]

I didn't realize I would spend almost every waking moment either working or thinking about our startup. You enter a whole different way of life when it's your company vs. working for someone else's company.

It's exacerbated by the fast pace of startups, which makes it seem like time slows down:

I think the thing that's been most surprising to me is how one's perspective on time shifts. Working on our startup, I remember time seeming to stretch out, so that a month was a huge interval.

In the best case, total immersion can be exciting:

It's surprising how much you become consumed by your startup, in that you think about it day and night, but never once does it feel like "work."

Though I have to say, that quote is from someone we funded this summer. In a couple years he may not sound so chipper.

3. It's an Emotional Roller-coaster

This was another one lots of people were surprised about. The ups and downs were more extreme than they were prepared for.

In a startup, things seem great one moment and hopeless the next. And by next, I mean a couple hours later.

The emotional ups and downs were the biggest surprise for me. One day, we'd think of ourselves as the next Google and dream of buying islands; the next, we'd be pondering how to let our loved ones know of our utter failure; and on and on.

The hard part, obviously, is the lows. For a lot of founders that was the big surprise:

How hard it is to keep everyone motivated during rough days or weeks, i.e. how low the lows can be.

After a while, if you don't have significant success to cheer you up, it wears you out:

Your most basic advice to founders is "just don't die," but the energy to keep a company going in lieu of unburdening success isn't free; it is siphoned from the founders themselves.

There's a limit to how much you can take. If you get to the point where you can't keep working anymore, it's not the end of the world. Plenty of famous founders have had some failures along the way.

4. It Can Be Fun

The good news is, the highs are also very high. Several founders said what surprised them most about doing a startup was how fun it was:

I think you've left out just how fun it is to do a startup. I am more fulfilled in my work than pretty much any of my friends who did not start companies.

What they like most is the freedom:

I'm surprised by how much better it feels to be working on something that is challenging and creative, something I believe in, as opposed to the hired-gun stuff I was doing before. I knew it would feel better; what's surprising is how much better.

Frankly, though, if I've misled people here, I'm not eager to fix that. I'd rather have everyone think starting a startup is grim and hard than have founders go into it expecting it to be fun, and a few months later saying "This is supposed to be fun? Are you kidding?"

The truth is, it wouldn't be fun for most people. A lot of what we try to do in the application process is to weed out the people who wouldn't like it, both for our sake and theirs.

The best way to put it might be that starting a startup is fun the way a survivalist training course would be fun, if you're into that sort of thing. Which is to say, not at all, if you're not.

5. Persistence Is the Key

A lot of founders were surprised how important persistence was in startups. It was both a negative and a positive surprise: they were surprised both by the degree of persistence required

Everyone said how determined and resilient you must be, but going through it made me realize that the determination required was still understated.

and also by the degree to which persistence alone was able to dissolve obstacles:

If you are persistent, even problems that seem out of your control (i.e. immigration) seem to work themselves out.

Several founders mentioned specifically how much more important persistence was than intelligence.

I've been surprised again and again by just how much more important persistence is than raw intelligence.

This applies not just to intelligence but to ability in general, and that's why so many people said character was more important in choosing cofounders.

6. Think Long-Term

You need persistence because everything takes longer than you expect. A lot of people were surprised by that.

I'm continually surprised by how long everything can take. Assuming your product doesn't experience the explosive growth that very few products do, everything from development to dealmaking (especially dealmaking) seems to take 2-3x longer than I always imagine.

One reason founders are surprised is that because they work fast, they expect everyone else to. There's a shocking amount of shear stress at every point where a startup touches a more bureaucratic organization, like a big company or a VC fund. That's why fundraising and the enterprise market kill and maim so many startups. [2]

But I think the reason most founders are surprised by how long it takes is that they're overconfident. They think they're going to be an instant success, like YouTube or Facebook. You tell them only 1 out of 100 successful startups has a trajectory like that, and they all think "we're going to be that 1."

Maybe they'll listen to one of the more successful founders:

The top thing I didn't understand before going into it is that persistence is the name of the game. For the vast majority of startups that become successful, it's going to be a really long journey, at least 3 years and probably 5+.

There is a positive side to thinking longer-term. It's not just that you have to resign yourself to everything taking longer than it should. If you work patiently it's less stressful, and you can do better work:

Because we're relaxed, it's so much easier to have fun doing what we do. Gone is the awkward nervous energy fueled by the desperate need to not fail guiding our actions. We can concentrate on doing what's best for our company, product, employees and customers.

That's why things get so much better when you hit ramen profitability. You can shift into a different mode of working.

7. Lots of Little Things

We often emphasize how rarely startups win simply because they hit on some magic idea. I think founders have now gotten that into their heads. But a lot were surprised to find this also applies within startups. You have to do lots of different things:

It's much more of a grind than glamorous. A timeslice selected at random would more likely find me tracking down a weird DLL loading bug on Swedish Windows, or tracking down a bug in the financial model Excel spreadsheet the night before a board meeting, rather than having brilliant flashes of strategic insight.

Most hacker-founders would like to spend all their time programming. You won't get to, unless you fail. Which can be transformed into: If you spend all your time programming, you will fail.

The principle extends even into programming. There is rarely a single brilliant hack that ensures success:

I learnt never to bet on any one feature or deal or anything to bring you success. It is never a single thing. Everything is just incremental and you just have to keep doing lots of those things until you strike something.

Even in the rare cases where a clever hack makes your fortune, you probably won't know till later:

There is no such thing as a killer feature. Or at least you won't know what it is.

So the best strategy is to try lots of different things. The reason not to put all your eggs in one basket is not the usual one, which applies even when you know which basket is best. In a startup you don't even know that.

8. Start with Something Minimal

Lots of founders mentioned how important it was to launch with the simplest possible thing. By this point everyone knows you should release fast and iterate. It's practically a mantra at YC. But even so a lot of people seem to have been burned by not doing it:

Build the absolute smallest thing that can be considered a complete application and ship it.

Why do people take too long on the first version? Pride, mostly. They hate to release something that could be better. They worry what people will say about them. But you have to overcome this:

Doing something "simple" at first glance does not mean you aren't doing something meaningful, defensible, or valuable.

Don't worry what people will say. If your first version is so impressive that trolls don't make fun of it, you waited too long to launch. [3]

One founder said this should be your approach to all programming, not just startups, and I tend to agree.

Now, when coding, I try to think "How can I write this such that if people saw my code, they'd be amazed at how little there is and how little it does?"

Over-engineering is poison. It's not like doing extra work for extra credit. It's more like telling a lie that you then have to remember so you don't contradict it.

9. Engage Users

Product development is a conversation with the user that doesn't really start till you launch. Before you launch, you're like a police artist before he's shown the first version of his sketch to the witness.

It's so important to launch fast that it may be better to think of your initial version not as a product, but as a trick for getting users to start talking to you.

I learned to think about the initial stages of a startup as a giant experiment. All products should be considered experiments, and those that have a market show promising results extremely quickly.

Once you start talking to users, I guarantee you'll be surprised by what they tell you.

When you let customers tell you what they're after, they will often reveal amazing details about what they find valuable as well what they're willing to pay for.

The surprise is generally positive as well as negative. They won't like what you've built, but there will be other things they would like that would be trivially easy to implement. It's not till you start the conversation by launching the wrong thing that they can express (or perhaps even realize) what they're looking for.

10. Change Your Idea

To benefit from engaging with users you have to be willing to change your idea. We've always encouraged founders to see a startup idea as a hypothesis rather than a blueprint. And yet they're still surprised how well it works to change the idea.

Normally if you complain about something being hard, the general advice is to work harder. With a startup, I think you should find a problem that's easy for you to solve. Optimizing in solution-space is familiar and straightforward, but you can make enormous gains playing around in problem-space.

Whereas mere determination, without flexibility, is a greedy algorithm that may get you nothing more than a mediocre local maximum:

When someone is determined, there's still a danger that they'll follow a long, hard path that ultimately leads nowhere.

You want to push forward, but at the same time twist and turn to find the most promising path. One founder put it very succinctly:

Fast iteration is the key to success.

One reason this advice is so hard to follow is that people don't realize how hard it is to judge startup ideas, particularly their own. Experienced founders learn to keep an open mind:

Now I don't laugh at ideas anymore, because I realized how terrible I was at knowing if they were good or not.

You can never tell what will work. You just have to do whatever seems best at each point. We do this with YC itself. We still don't know if it will work, but it seems like a decent hypothesis.

11. Don't Worry about Competitors

When you think you've got a great idea, it's sort of like having a guilty conscience about something. All someone has to do is look at you funny, and you think "Oh my God, they know."

These alarms are almost always false:

Companies that seemed like competitors and threats at first glance usually never were when you really looked at it. Even if they were operating in the same area, they had a different goal.

One reason people overreact to competitors is that they overvalue ideas. If ideas really were the key, a competitor with the same idea would be a real threat. But it's usually execution that matters:

All the scares induced by seeing a new competitor pop up are forgotten weeks later. It always comes down to your own product and approach to the market.

This is generally true even if competitors get lots of attention.

Competitors riding on lots of good blogger perception aren't really the winners and can disappear from the map quickly. You need consumers after all.

Hype doesn't make satisfied users, at least not for something as complicated as technology.

12. It's Hard to Get Users

A lot of founders complained about how hard it was to get users, though.

I had no idea how much time and effort needed to go into attaining users.

This is a complicated topic. When you can't get users, it's hard to say whether the problem is lack of exposure, or whether the product's simply bad. Even good products can be blocked by switching or integration costs:

Getting people to use a new service is incredibly difficult. This is especially true for a service that other companies can use, because it requires their developers to do work. If you're small, they don't think it is urgent. [4]

The sharpest criticism of YC came from a founder who said we didn't focus enough on customer acquisition:

YC preaches "make something people want" as an engineering task, a never ending stream of feature after feature until enough people are happy and the application takes off. There's very little focus on the cost of customer acquisition.

This may be true; this may be something we need to fix, especially for applications like games. If you make something where the challenges are mostly technical, you can rely on word of mouth, like Google did. One founder was surprised by how well that worked for him:

There is an irrational fear that no one will buy your product. But if you work hard and incrementally make it better, there is no need to worry.

But with other types of startups you may win less by features and more by deals and marketing.

13. Expect the Worst with Deals

Deals fall through. That's a constant of the startup world. Startups are powerless, and good startup ideas generally seem wrong. So everyone is nervous about closing deals with you, and you have no way to make them.

This is particularly true with investors:

In retrospect, it would have been much better if we had operated under the assumption that we would never get any additional outside investment. That would have focused us on finding revenue streams early.

My advice is generally pessimistic. Assume you won't get money, and if someone does offer you any, assume you'll never get any more.

If someone offers you money, take it. You say it a lot, but I think it needs even more emphasizing. We had the opportunity to raise a lot more money than we did last year and I wish we had.

Why do founders ignore me? Mostly because they're optimistic by nature. The mistake is to be optimistic about things you can't control. By all means be optimistic about your ability to make something great. But you're asking for trouble if you're optimistic about big companies or investors.

14. Investors Are Clueless

A lot of founders mentioned how surprised they were by the cluelessness of investors:

They don't even know about the stuff they've invested in. I met some investors that had invested in a hardware device and when I asked them to demo the device they had difficulty switching it on.

Angels are a bit better than VCs, because they usually have startup experience themselves:

VC investors don't know half the time what they are talking about and are years behind in their thinking. A few were great, but 95% of the investors we dealt with were unprofessional, didn't seem to be very good at business or have any kind of creative vision. Angels were generally much better to talk to.

Why are founders surprised that VCs are clueless? I think it's because they seem so formidable.

The reason VCs seem formidable is that it's their profession to. You get to be a VC by convincing asset managers to trust you with hundreds of millions of dollars. How do you do that? You have to seem confident, and you have to seem like you understand technology. [5]

15. You May Have to Play Games

Because investors are so bad at judging you, you have to work harder than you should at selling yourself. One founder said the thing that surprised him most was

The degree to which feigning certitude impressed investors.

This is the thing that has surprised me most about YC founders' experiences. This summer we invited some of the alumni to talk to the new startups about fundraising, and pretty much 100% of their advice was about investor psychology. I thought I was cynical about VCs, but the founders were much more cynical.

A lot of what startup founders do is just posturing. It works.

VCs themselves have no idea of the extent to which the startups they like are the ones that are best at selling themselves to VCs. [6] It's exactly the same phenomenon we saw a step earlier. VCs get money by seeming confident to LPs, and founders get money by seeming confident to VCs.

16. Luck Is a Big Factor

With two such random linkages in the path between startups and money, it shouldn't be surprising that luck is a big factor in deals. And yet a lot of founders are surprised by it.

I didn't realize how much of a role luck plays and how much is outside of our control.

If you think about famous startups, it's pretty clear how big a role luck plays. Where would Microsoft be if IBM insisted on an exclusive license for DOS?

Why are founders fooled by this? Business guys probably aren't, but hackers are used to a world where skill is paramount, and you get what you deserve.

When we started our startup, I had bought the hype of the startup founder dream: that this is a game of skill. It is, in some ways. Having skill is valuable. So is being determined as all hell. But being lucky is the critical ingredient.

Actually the best model would be to say that the outcome is the product of skill, determination, and luck. No matter how much skill and determination you have, if you roll a zero for luck, the outcome is zero.

These quotes about luck are not from founders whose startups failed. Founders who fail quickly tend to blame themselves. Founders who succeed quickly don't usually realize how lucky they were. It's the ones in the middle who see how important luck is.

17. The Value of Community

A surprising number of founders said what surprised them most about starting a startup was the value of community. Some meant the micro-community of YC founders:

The immense value of the peer group of YC companies, and facing similar obstacles at similar times.

which shouldn't be that surprising, because that's why it's structured that way. Others were surprised at the value of the startup community in the larger sense:

How advantageous it is to live in Silicon Valley, where you can't help but hear all the cutting-edge tech and startup news, and run into useful people constantly.

The specific thing that surprised them most was the general spirit of benevolence:

One of the most surprising things I saw was the willingness of people to help us. Even people who had nothing to gain went out of their way to help our startup succeed.

and particularly how it extended all the way to the top:

The surprise for me was how accessible important and interesting people are. It's amazing how easily you can reach out to people and get immediate feedback.

This is one of the reasons I like being part of this world. Creating wealth is not a zero-sum game, so you don't have to stab people in the back to win.

18. You Get No Respect

There was one surprise founders mentioned that I'd forgotten about: that outside the startup world, startup founders get no respect.

In social settings, I found that I got a lot more respect when I said, "I worked on Microsoft Office" instead of "I work at a small startup you've never heard of called x."

Partly this is because the rest of the world just doesn't get startups, and partly it's yet another consequence of the fact that most good startup ideas seem bad:

If you pitch your idea to a random person, 95% of the time you'll find the person instinctively thinks the idea will be a flop and you're wasting your time (although they probably won't say this directly).

Unfortunately this extends even to dating:

It surprised me that being a startup founder does not get you more admiration from women.

I did know about that, but I'd forgotten.

19. Things Change as You Grow

The last big surprise founders mentioned is how much things changed as they grew. The biggest change was that you got to program even less:

Your job description as technical founder/CEO is completely rewritten every 6-12 months. Less coding, more managing/planning/company building, hiring, cleaning up messes, and generally getting things in place for what needs to happen a few months from now.

In particular, you now have to deal with employees, who often have different motivations:

I knew the founder equation and had been focused on it since I knew I wanted to start a startup as a 19 year old. The employee equation is quite different so it took me a while to get it down.

Fortunately, it can become a lot less stressful once you reach cruising altitude:

I'd say 75% of the stress is gone now from when we first started. Running a business is so much more enjoyable now. We're more confident. We're more patient. We fight less. We sleep more.

I wish I could say it was this way for every startup that succeeded, but 75% is probably on the high side.

The Super-Pattern

There were a few other patterns, but these were the biggest. One's first thought when looking at them all is to ask if there's a super-pattern, a pattern to the patterns.

I saw it immediately, and so did a YC founder I read the list to. These are supposed to be the surprises, the things I didn't tell people. What do they all have in common? They're all things I tell people. If I wrote a new essay with the same outline as this that wasn't summarizing the founders' responses, everyone would say I'd run out of ideas and was just repeating myself.

What is going on here?

When I look at the responses, the common theme is that starting a startup was like I said, but way more so. People just don't seem to get how different it is till they do it. Why? The key to that mystery is to ask, how different from what? Once you phrase it that way, the answer is obvious: from a job. Everyone's model of work is a job. It's completely pervasive. Even if you've never had a job, your parents probably did, along with practically every other adult you've met.

Unconsciously, everyone expects a startup to be like a job, and that explains most of the surprises. It explains why people are surprised how carefully you have to choose cofounders and how hard you have to work to maintain your relationship. You don't have to do that with coworkers. It explains why the ups and downs are surprisingly extreme. In a job there is much more damping. But it also explains why the good times are surprisingly good: most people can't imagine such freedom. As you go down the list, almost all the surprises are surprising in how much a startup differs from a job.

You probably can't overcome anything so pervasive as the model of work you grew up with. So the best solution is to be consciously aware of that. As you go into a startup, you'll be thinking "everyone says it's really extreme." Your next thought will probably be "but I can't believe it will be that bad." If you want to avoid being surprised, the next thought after that should be: "and the reason I can't believe it will be that bad is that my model of work is a job."

Notes

[1] Graduate students might understand it. In grad school you always feel you should be working on your thesis. It doesn't end every semester like classes do.

[2] The best way for a startup to engage with slow-moving organizations is to fork off separate processes to deal with them. It's when they're on the critical path that they kill you—when you depend on closing a deal to move forward. It's worth taking extreme measures to avoid that.

[3] This is a variant of Reid Hoffman's principle that if you aren't embarrassed by what you launch with, you waited too long to launch.

[4] The question to ask about what you've built is not whether it's good, but whether it's good enough to supply the activation energy required.

[5] Some VCs seem to understand technology because they actually do, but that's overkill; the defining test is whether you can talk about it well enough to convince limited partners.

[6] This is the same phenomenon you see with defense contractors or fashion brands. The dumber the customers, the more effort you expend on the process of selling things to them rather than making the things you sell.

Thanks: to Jessica Livingston for reading drafts of this, and to all the founders who responded to my email.

Related:

Startups in 13 Sentences

The Hardest Lessons for Startups to Learn

How Not to Die

The 18 Mistakes That Kill Startups

A Fundraising Survival Guide

Persuade xor Discover

September 2009

When meeting people you don't know very well, the convention is to seem extra friendly. You smile and say "pleased to meet you," whether you are or not. There's nothing dishonest about this. Everyone knows that these little social lies aren't meant to be taken literally, just as everyone knows that "Can you pass the salt?" is only grammatically a question.

I'm perfectly willing to smile and say "pleased to meet you" when meeting new people. But there is another set of customs for being ingratiating in print that are not so harmless.

The reason there's a convention of being ingratiating in print is that most essays are written to persuade. And as any politician could tell you, the way to persuade people is not just to baldly state the facts. You have to add a spoonful of sugar to make the medicine go down.

For example, a politician announcing the cancellation of a government program will not merely say "The program is canceled." That would seem offensively curt. Instead he'll spend most of his time talking about the noble effort made by the people who worked on it.

The reason these conventions are more dangerous is that they interact with the ideas. Saying "pleased to meet you" is just something you prepend to a conversation, but the sort of spin added by politicians is woven through it. We're starting to move from social lies to real lies.

Here's an example of a paragraph from an essay I wrote about labor unions. As written, it tends to offend people who like unions.

People who think the labor movement was the creation of heroic union organizers have a problem to explain: why are unions shrinking now? The best they can do is fall back on the default explanation of people living in fallen civilizations. Our ancestors were giants. The workers of the early twentieth century must have had a moral courage that's lacking today.

Now here's the same paragraph rewritten to please instead of offending them:

Early union organizers made heroic sacrifices to improve conditions for workers. But though labor unions are shrinking now, it's not because present union leaders are any less courageous. An employer couldn't get away with hiring thugs to beat up union leaders today, but if they did, I see no reason to believe today's union leaders would shrink from the challenge. So I think it would be a mistake to attribute the decline of unions to some kind of decline in the people who run them. Early union leaders were heroic, certainly, but we should not suppose that if unions have declined, it's because present union leaders are somehow inferior. The cause must be external. [1]

It makes the same point: that it can't have been the personal qualities of early union organizers that made unions successful, but must have been some external factor, or otherwise present-day union leaders would have to be inferior people. But written this way it seems like a defense of present-day union organizers rather than an attack on early ones. That makes it more persuasive to people who like unions, because it seems sympathetic to their cause.

I believe everything I wrote in the second version. Early union leaders did make heroic sacrifices. And present union leaders probably would rise to the occasion if necessary. People tend to; I'm skeptical about the idea of "the greatest generation." [2]

If I believe everything I said in the second version, why didn't I write it that way? Why offend people needlessly?

Because I'd rather offend people than pander to them, and if you write about controversial topics you have to choose one or the other. The degree of courage of past or present union leaders is beside the point; all that matters for the argument is that they're the same. But if you want to please people who are mistaken, you can't simply tell the truth. You're always going to have to add some sort of padding to protect their misconceptions from bumping against reality.

Most writers do. Most writers write to persuade, if only out of habit or politeness. But I don't write to persuade; I write to figure out. I write to persuade a hypothetical perfectly unbiased reader.

Since the custom is to write to persuade the actual reader, someone who doesn't will seem arrogant. In fact, worse than arrogant: since readers are used to essays that try to please someone, an essay that displeases one side in a dispute reads as an attempt to pander to the other. To a lot of pro-union readers, the first paragraph sounds like the sort of thing a right-wing radio talk show host would say to stir up his followers. But it's not. Something that curtly contradicts one's beliefs can be hard to distinguish from a partisan attack on them, but though they can end up in the same place they come from different sources.

Would it be so bad to add a few extra words, to make people feel better? Maybe not. Maybe I'm excessively attached to conciseness. I write code the same way I write essays, making pass after pass looking for anything I can cut. But I have a legitimate reason for doing this. You don't know what the ideas are until you get them down to the fewest words. [3]

The danger of the second paragraph is not merely that it's longer. It's that you start to lie to yourself. The ideas start to get mixed together with the spin you've added to get them past the readers' misconceptions.

I think the goal of an essay should be to discover surprising things. That's my goal, at least. And most surprising means most different from what people currently believe. So writing to persuade and writing to discover are diametrically opposed. The more your conclusions disagree with readers' present beliefs, the more effort you'll have to expend on selling your ideas rather than having them. As you accelerate, this drag increases, till eventually you reach a point where 100% of your energy is devoted to overcoming it and you can't go any faster.

It's hard enough to overcome one's own misconceptions without having to think about how to get the resulting ideas past other people's. I worry that if I wrote to persuade, I'd start to shy away unconsciously from ideas I knew would be hard to sell. When I notice something surprising, it's usually very faint at first. There's nothing more than a slight stirring of discomfort. I don't want anything to get in the way of noticing it consciously.

Notes

[1] I had a strange feeling of being back in high school writing this. To get a good grade you had to both write the sort of pious crap you were expected to, but also seem to be writing with conviction. The solution was a kind of method acting. It was revoltingly familiar to slip back into it.

[2] Exercise for the reader: rephrase that thought to please the same people the first version would offend.

[3] Come to think of it, there is one way in which I deliberately pander to readers, because it doesn't change the number of words: I switch person. This flattering distinction seems so natural to the average reader that they probably don't notice even when I switch in mid-sentence, though you tend to notice when it's done as conspicuously as this.

Thanks to Jessica Livingston and Robert Morris for reading drafts of this.

Note: An earlier version of this essay began by talking about why people dislike Michael Arrington. I now believe that was mistaken, and that most people don't dislike him for the same reason I did when I first met him, but simply because he writes about controversial things.

Post-Medium Publishing

September 2009

Publishers of all types, from news to music, are unhappy that consumers won't pay for content anymore. At least, that's how they see it.

In fact consumers never really were paying for content, and publishers weren't really selling it either. If the content was what they were selling, why has the price of books or music or movies always depended mostly on the format? Why didn't better content cost more? [1]

A copy of Time costs $5 for 58 pages, or 8.6 cents a page. The Economist costs $7 for 86 pages, or 8.1 cents a page. Better journalism is actually slightly cheaper.

Almost every form of publishing has been organized as if the medium was what they were selling, and the content was irrelevant. Book publishers, for example, set prices based on the cost of producing and distributing books. They treat the words printed in the book the same way a textile manufacturer treats the patterns printed on its fabrics.

Economically, the print media are in the business of marking up paper. We can all imagine an old-style editor getting a scoop and saying "this will sell a lot of papers!" Cross out that final S and you're describing their business model. The reason they make less money now is that people don't need as much paper.

A few months ago I ran into a friend in a cafe. I had a copy of the New York Times, which I still occasionally buy on weekends. As I was leaving I offered it to him, as I've done countless times before in the same situation. But this time something new happened. I felt that sheepish feeling you get when you offer someone something worthless. "Do you, er, want a printout of yesterday's news?" I asked. (He didn't.)

Now that the medium is evaporating, publishers have nothing left to sell. Some seem to think they're going to sell content—that they were always in the content business, really. But they weren't, and it's unclear whether anyone could be.

Selling

There have always been people in the business of selling information, but that has historically been a distinct business from publishing. And the business of selling information to consumers has always been a marginal one. When I was a kid there were people who used to sell newsletters containing stock tips, printed on colored paper that made them hard for the copiers of the day to reproduce. That is a different world, both culturally and economically, from the one publishers currently inhabit.

People will pay for information they think they can make money from. That's why they paid for those stock tip newsletters, and why companies pay now for Bloomberg terminals and Economist Intelligence Unit reports. But will people pay for information otherwise? History offers little encouragement.

If audiences were willing to pay more for better content, why wasn't anyone already selling it to them? There was no reason you couldn't have done that in the era of physical media. So were the print media and the music labels simply overlooking this opportunity? Or is it, rather, nonexistent?

What about iTunes? Doesn't that show people will pay for content? Well, not really. iTunes is more of a tollbooth than a store. Apple controls the default path onto the iPod. They offer a convenient list of songs, and whenever you choose one they ding your credit card for a small amount, just below the threshold of attention. Basically, iTunes makes money by taxing people, not selling them stuff. You can only do that if you own the channel, and even then you don't make much from it, because a toll has to be ignorable to work. Once a toll becomes painful, people start to find ways around it, and that's pretty easy with digital content.

The situation is much the same with digital books. Whoever controls the device sets the terms. It's in their interest for content to be as cheap as possible, and since they own the channel, there's a lot they can do to drive prices down. Prices will fall even further once writers realize they don't need publishers. Getting a book printed and distributed is a daunting prospect for a writer, but most can upload a file.

Is software a counterexample? People pay a lot for desktop software, and that's just information. True, but I don't think publishers can learn much from software. Software companies can charge a lot because (a) many of the customers are businesses, who get in trouble if they use pirated versions, and (b) though in form merely information, software is treated by both maker and purchaser as a different type of thing from a song or an article. A Photoshop user needs Photoshop in a way that no one needs a particular song or article.

That's why there's a separate word, "content," for information that's not software. Software is a different business. Software and content blur together in some of the most lightweight software, like casual games. But those are usually free. To make money the way software companies do, publishers would have to become software companies, and being publishers gives them no particular head start in that domain. [2]

The most promising countertrend is the premium cable channel. People still pay for those. But broadcasting isn't publishing: you're not selling a copy of something. That's one reason the movie business hasn't seen their revenues decline the way the news and music businesses have. They only have one foot in publishing.

To the extent the movie business can avoid becoming publishers, they may avoid publishing's problems. But there are limits to how well they'll be able to do that. Once publishing—giving people copies—becomes the most natural way of distributing your content, it probably doesn't work to stick to old forms of distribution just because you make more that way. If free copies of your content are available online, then you're competing with publishing's form of distribution, and that's just as bad as being a publisher.

Apparently some people in the music business hope to retroactively convert it away from publishing, by getting listeners to pay for subscriptions. It seems unlikely that will work if they're just streaming the same files you can get as mp3s.

Next

What happens to publishing if you can't sell content? You have two choices: give it away and make money from it indirectly, or find ways to embody it in things people will pay for.

The first is probably the future of most current media. Give music away and make money from concerts and t-shirts. Publish articles for free and make money from one of a dozen permutations of advertising. Both publishers and investors are down on advertising at the moment, but it has more potential than they realize.

I'm not claiming that potential will be realized by the existing players. The optimal ways to make money from the written word probably require different words written by different people.

It's harder to say what will happen to movies. They could evolve into ads. Or they could return to their roots and make going to the theater a treat. If they made the experience good enough, audiences might start to prefer it to watching pirated movies at home. [3] Or maybe the movie business will dry up, and the people working in it will go to work for game developers.

I don't know how big embodying information in physical form will be. It may be surprisingly large; people overvalue physical stuff. There should remain some market for printed books, at least.

I can see the evolution of book publishing in the books on my shelves. Clearly at some point in the 1960s the big publishing houses started to ask: how cheaply can we make books before people refuse to buy them? The answer turned out to be one step short of phonebooks. As long as it isn't floppy, consumers still perceive it as a book.

That worked as long as buying printed books was the only way to read them. If printed books are optional, publishers will have to work harder to entice people to buy them. There should be some market, but it's hard to foresee how big, because its size will depend not on macro trends like the amount people read, but on the ingenuity of individual publishers. [4]

Some magazines may thrive by focusing on the magazine as a physical object. Fashion magazines could be made lush in a way that would be hard to match digitally, at least for a while. But this is probably not an option for most magazines.

I don't know exactly what the future will look like, but I'm not too worried about it. This sort of change tends to create as many good things as it kills. Indeed, the really interesting question is not what will happen to existing forms, but what new forms will appear.

The reason I've been writing about existing forms is that I don't know what new forms will appear. But though I can't predict specific winners, I can offer a recipe for recognizing them. When you see something that's taking advantage of new technology to give people something they want that they couldn't have before, you're probably looking at a winner. And when you see something that's merely reacting to new technology in an attempt to preserve some existing source of revenue, you're probably looking at a loser.

Notes

[1] I don't like the word "content" and tried for a while to avoid using it, but I have to admit there's no other word that means the right thing. "Information" is too general.

Ironically, the main reason I don't like "content" is the thesis of this essay. The word suggests an undifferentiated slurry, but economically that's how both publishers and audiences treat it. Content is information you don't need.

[2] Some types of publishers would be at a disadvantage trying to enter the software business. Record labels, for example, would probably find it more natural to expand into casinos than software, because the kind of people who run them would be more at home at the mafia end of the business spectrum than the don't-be-evil end.

[3] I never watch movies in theaters anymore. The tipping point for me was the ads they show first.

[4] Unfortunately, making physically nice books will only be a niche within a niche. Publishers are more likely to resort to expedients like selling autographed copies, or editions with the buyer's picture on the cover.

Thanks to Michael Arrington, Trevor Blackwell, Steven Levy, Robert Morris, and Geoff Ralston for reading drafts of this.

The List of N Things

September 2009

I bet you the current issue of Cosmopolitan has an article whose title begins with a number. "7 Things He Won't Tell You about Sex," or something like that. Some popular magazines feature articles of this type on the cover of every issue. That can't be happening by accident. Editors must know they attract readers.

Why do readers like the list of n things so much? Mainly because it's easier to read than a regular article. [1] Structurally, the list of n things is a degenerate case of essay. An essay can go anywhere the writer wants. In a list of n things the writer agrees to constrain himself to a collection of points of roughly equal importance, and he tells the reader explicitly what they are.

Some of the work of reading an article is understanding its structure—figuring out what in high school we'd have called its "outline." Not explicitly, of course, but someone who really understands an article probably has something in his brain afterward that corresponds to such an outline. In a list of n things, this work is done for you. Its structure is an exoskeleton.

As well as being explicit, the structure is guaranteed to be of the simplest possible type: a few main points with few to no subordinate ones, and no particular connection between them.

Because the main points are unconnected, the list of n things is random access. There's no thread of reasoning you have to follow. You could read the list in any order. And because the points are independent of one another, they work like watertight compartments in an unsinkable ship. If you get bored with, or can't understand, or don't agree with one point, you don't have to give up on the article. You can just abandon that one and skip to the next. A list of n things is parallel and therefore fault tolerant.

There are times when this format is what a writer wants. One, obviously, is when what you have to say actually is a list of n things. I once wrote an essay about the mistakes that kill startups, and a few people made fun of me for writing something whose title began with a number. But in that case I really was trying to make a complete catalog of a number of independent things. In fact, one of the questions I was trying to answer was how many there were.

There are other less legitimate reasons for using this format. For example, I use it when I get close to a deadline. If I have to give a talk and I haven't started it a few days beforehand, I'll sometimes play it safe and make the talk a list of n things.

The list of n things is easier for writers as well as readers. When you're writing a real essay, there's always a chance you'll hit a dead end. A real essay is a train of thought, and some trains of thought just peter out. That's an alarming possibility when you have to give a talk in a few days. What if you run out of ideas? The compartmentalized structure of the list of n things protects the writer from his own stupidity in much the same way it protects the reader. If you run out of ideas on one point, no problem: it won't kill the essay. You can take out the whole point if you need to, and the essay will still survive.

Writing a list of n things is so relaxing. You think of n/2 of them in the first 5 minutes. So bang, there's the structure, and you just have to fill it in. As you think of more points, you just add them to the end. Maybe you take out or rearrange or combine a few, but at every stage you have a valid (though initially low-res) list of n things. It's like the sort of programming where you write a version 1 very quickly and then gradually modify it, but at every point have working code—or the style of painting where you begin with a complete but very blurry sketch done in an hour, then spend a week cranking up the resolution.

Because the list of n things is easier for writers too, it's not always a damning sign when readers prefer it. It's not necessarily evidence readers are lazy; it could also mean they don't have much confidence in the writer. The list of n things is in that respect the cheeseburger of essay forms. If you're eating at a restaurant you suspect is bad, your best bet is to order the cheeseburger. Even a bad cook can make a decent cheeseburger. And there are pretty strict conventions about what a cheeseburger should look like. You can assume the cook isn't going to try something weird and artistic. The list of n things similarly limits the damage that can be done by a bad writer. You know it's going to be about whatever the title says, and the format prevents the writer from indulging in any flights of fancy.

Because the list of n things is the easiest essay form, it should be a good one for beginning writers. And in fact it is what most beginning writers are taught. The classic 5 paragraph essay is really a list of n things for n = 3. But the students writing them don't realize they're using the same structure as the articles they read in Cosmopolitan. They're not allowed to include the numbers, and they're expected to spackle over the gaps with gratuitous transitions ("Furthermore...") and cap the thing at either end with introductory and concluding paragraphs so it will look superficially like a real essay. [2]

It seems a fine plan to start students off with the list of n things. It's the easiest form. But if we're going to do that, why not do it openly? Let them write lists of n things like the pros, with numbers and no transitions or "conclusion."

There is one case where the list of n things is a dishonest format: when you use it to attract attention by falsely claiming the list is an exhaustive one. I.e. if you write an article that purports to be about the 7 secrets of success. That kind of title is the same sort of reflexive challenge as a whodunit. You have to at least look at the article to check whether they're the same 7 you'd list. Are you overlooking one of the secrets of success? Better check.

It's fine to put "The" before the number if you really believe you've made an exhaustive list. But evidence suggests most things with titles like this are linkbait.

The greatest weakness of the list of n things is that there's so little room for new thought. The main point of essay writing, when done right, is the new ideas you have while doing it. A real essay, as the name implies, is dynamic: you don't know what you're going to write when you start. It will be about whatever you discover in the course of writing it.

This can only happen in a very limited way in a list of n things. You make the title first, and that's what it's going to be about. You can't have more new ideas in the writing than will fit in the watertight compartments you set up initially. And your brain seems to know this: because you don't have room for new ideas, you don't have them.

Another advantage of admitting to beginning writers that the 5 paragraph essay is really a list of n things is that we can warn them about this. It only lets you experience the defining characteristic of essay writing on a small scale: in thoughts of a sentence or two. And it's particularly dangerous that the 5 paragraph essay buries the list of n things within something that looks like a more sophisticated type of essay. If you don't know you're using this form, you don't know you need to escape it.

Notes

[1] Articles of this type are also startlingly popular on Delicious, but I think that's because delicious/popular is driven by bookmarking, not because Delicious users are stupid. Delicious users are collectors, and a list of n things seems particularly collectible because it's a collection itself.

[2] Most "word problems" in school math textbooks are similarly misleading. They look superficially like the application of math to real problems, but they're not. So if anything they reinforce the impression that math is merely a complicated but pointless collection of stuff to be memorized.

Russian Translation

The Anatomy of Determination

September 2009

Like all investors, we spend a lot of time trying to learn how to predict which startups will succeed. We probably spend more time thinking about it than most, because we invest the earliest. Prediction is usually all we have to rely on.

We learned quickly that the most important predictor of success is determination. At first we thought it might be intelligence. Everyone likes to believe that's what makes startups succeed. It makes a better story that a company won because its founders were so smart. The PR people and reporters who spread such stories probably believe them themselves. But while it certainly helps to be smart, it's not the deciding factor. There are plenty of people as smart as Bill Gates who achieve nothing.

In most domains, talent is overrated compared to determination—partly because it makes a better story, partly because it gives onlookers an excuse for being lazy, and partly because after a while determination starts to look like talent.

I can't think of any field in which determination is overrated, but the relative importance of determination and talent probably do vary somewhat. Talent probably matters more in types of work that are purer, in the sense that one is solving mostly a single type of problem instead of many different types. I suspect determination would not take you as far in math as it would in, say, organized crime.

I don't mean to suggest by this comparison that types of work that depend more on talent are always more admirable. Most people would agree it's more admirable to be good at math than memorizing long strings of digits, even though the latter depends more on natural ability.

Perhaps one reason people believe startup founders win by being smarter is that intelligence does matter more in technology startups than it used to in earlier types of companies. You probably do need to be a bit smarter to dominate Internet search than you had to be to dominate railroads or hotels or newspapers. And that's probably an ongoing trend. But even in the highest of high tech industries, success still depends more on determination than brains.

If determination is so important, can we isolate its components? Are some more important than others? Are there some you can cultivate?

The simplest form of determination is sheer willfulness. When you want something, you must have it, no matter what.

A good deal of willfulness must be inborn, because it's common to see families where one sibling has much more of it than another. Circumstances can alter it, but at the high end of the scale, nature seems to be more important than nurture. Bad circumstances can break the spirit of a strong-willed person, but I don't think there's much you can do to make a weak-willed person stronger-willed.

Being strong-willed is not enough, however. You also have to be hard on yourself. Someone who was strong-willed but self-indulgent would not be called determined. Determination implies your willfulness is balanced by discipline.

That word balance is a significant one. The more willful you are, the more disciplined you have to be. The stronger your will, the less anyone will be able to argue with you except yourself. And someone has to argue with you, because everyone has base impulses, and if you have more will than discipline you'll just give into them and end up on a local maximum like drug addiction.

We can imagine will and discipline as two fingers squeezing a slippery melon seed. The harder they squeeze, the further the seed flies, but they must both squeeze equally or the seed spins off sideways.

If this is true it has interesting implications, because discipline can be cultivated, and in fact does tend to vary quite a lot in the course of an individual's life. If determination is effectively the product of will and discipline, then you can become more determined by being more disciplined. [1]

Another consequence of the melon seed model is that the more willful you are, the more dangerous it is to be undisciplined. There seem to be plenty of examples to confirm that. In some very energetic people's lives you see something like wing flutter, where they alternate between doing great work and doing absolutely nothing. Externally this would look a lot like bipolar disorder.

The melon seed model is inaccurate in at least one respect, however: it's static. In fact the dangers of indiscipline increase with temptation. Which means, interestingly, that determination tends to erode itself. If you're sufficiently determined to achieve great things, this will probably increase the number of temptations around you. Unless you become proportionally more disciplined, willfulness will then get the upper hand, and your achievement will revert to the mean.

That's why Julius Caesar thought thin men so dangerous. They weren't tempted by the minor perquisites of power.

The melon seed model implies it's possible to be too disciplined. Is it? I think there probably are people whose willfulness is crushed down by excessive discipline, and who would achieve more if they weren't so hard on themselves. One reason the young sometimes succeed where the old fail is that they don't realize how incompetent they are. This lets them do a kind of deficit spending. When they first start working on something, they overrate their achievements. But that gives them confidence to keep working, and their performance improves. Whereas someone clearer-eyed would see their initial incompetence for what it was, and perhaps be discouraged from continuing.

There's one other major component of determination: ambition. If willfulness and discipline are what get you to your destination, ambition is how you choose it.

I don't know if it's exactly right to say that ambition is a component of determination, but they're not entirely orthogonal. It would seem a misnomer if someone said they were very determined to do something trivially easy.

And fortunately ambition seems to be quite malleable; there's a lot you can do to increase it. Most people don't know how ambitious to be, especially when they're young. They don't know what's hard, or what they're capable of. And this problem is exacerbated by having few peers. Ambitious people are rare, so if everyone is mixed together randomly, as they tend to be early in people's lives, then the ambitious ones won't have many ambitious peers. When you take people like this and put them together with other ambitious people, they bloom like dying plants given water. Probably most ambitious people are starved for the sort of encouragement they'd get from ambitious peers, whatever their age. [2]

Achievements also tend to increase your ambition. With each step you gain confidence to stretch further next time.

So here in sum is how determination seems to work: it consists of willfulness balanced with discipline, aimed by ambition. And fortunately at least two of these three qualities can be cultivated. You may be able to increase your strength of will somewhat; you can definitely learn self-discipline; and almost everyone is practically malnourished when it comes to ambition.

I feel like I understand determination a bit better now. But only a bit: willfulness, discipline, and ambition are all concepts almost as complicated as determination. [3]

Note too that determination and talent are not the whole story. There's a third factor in achievement: how much you like the work. If you really love working on something, you don't need determination to drive you; it's what you'd do anyway. But most types of work have aspects one doesn't like, because most types of work consist of doing things for other people, and it's very unlikely that the tasks imposed by their needs will happen to align exactly with what you want to do.

Indeed, if you want to create the most wealth, the way to do it is to focus more on their needs than your interests, and make up the difference with determination.

Notes

[1] Loosely speaking. What I'm claiming with the melon seed model is more like determination is proportionate to wd^m - k|w - d|^n, where w is will and d discipline.

[2] Which means one of the best ways to help a society generally is to create events and institutions that bring ambitious people together. It's like pulling the control rods out of a reactor: the energy they emit encourages other ambitious people, instead of being absorbed by the normal people they're usually surrounded with.

Conversely, it's probably a mistake to do as some European countries have done and try to ensure none of your universities is significantly better than the others.

[3] For example, willfulness clearly has two subcomponents, stubbornness and energy. The first alone yields someone who's stubbornly inert. The second alone yields someone flighty. As willful people get older or otherwise lose their energy, they tend to become merely stubborn.

Thanks to Sam Altman, Jessica Livingston, and Robert Morris for reading drafts of this.

Italian Translation

Portuguese Translation

What Kate Saw in Silicon Valley

August 2009

Kate Courteau is the architect who designed Y Combinator's office. Recently we managed to recruit her to help us run YC when she's not busy with architectural projects. Though she'd heard a lot about YC since the beginning, the last 9 months have been a total immersion.

I've been around the startup world for so long that it seems normal to me, so I was curious to hear what had surprised her most about it. This was her list:

1. How many startups fail. Kate knew in principle that startups were very risky, but she was surprised to see how constant the threat of failure was—not just for the minnows, but even for the famous startups whose founders came to speak at YC dinners.

2. How much startups' ideas change. As usual, by Demo Day about half the startups were doing something significantly different than they started with. We encourage that. Starting a startup is like science in that you have to follow the truth wherever it leads. In the rest of the world, people don't start things till they're sure what they want to do, and once started they tend continue on their initial path even if it's mistaken.

3. How little money it can take to start a startup. In Kate's world, everything is still physical and expensive. You can barely renovate a bathroom for the cost of starting a startup.

4. How scrappy founders are. That was her actual word. I agree with her, but till she mentioned this it never occurred to me how little this quality is appreciated in most of the rest of the world. It wouldn't be a compliment in most organizations to call someone scrappy.

What does it mean, exactly? It's basically the diminutive form of belligerent. Someone who's scrappy manages to be both threatening and undignified at the same time. Which seems to me exactly what one would want to be, in any kind of work. If you're not threatening, you're probably not doing anything new, and dignity is merely a sort of plaque.

5. How tech-saturated Silicon Valley is. "It seems like everybody here is in the industry." That isn't literally true, but there is a qualitative difference between Silicon Valley and other places. You tend to keep your voice down, because there's a good chance the person at the next table would know some of the people you're talking about. I never felt that in Boston. The good news is, there's also a good chance the person at the next table could help you in some way.

6. That the speakers at YC were so consistent in their advice. Actually, I've noticed this too. I always worry the speakers will put us in an embarrassing position by contradicting what we tell the startups, but it happens surprisingly rarely.

When I asked her what specific things she remembered speakers always saying, she mentioned: that the way to succeed was to launch something fast, listen to users, and then iterate; that startups required resilience because they were always an emotional rollercoaster; and that most VCs were sheep.

I've been impressed by how consistently the speakers advocate launching fast and iterating. That was contrarian advice 10 years ago, but it's clearly now the established practice.

7. How casual successful startup founders are. Most of the famous founders in Silicon Valley are people you'd overlook on the street. It's not merely that they don't dress up. They don't project any kind of aura of power either. "They're not trying to impress anyone."

Interestingly, while Kate said that she could never pick out successful founders, she could recognize VCs, both by the way they dressed and the way they carried themselves.

8. How important it is for founders to have people to ask for advice. (I swear I didn't prompt this one.) Without advice "they'd just be sort of lost." Fortunately, there are a lot of people to help them. There's a strong tradition within YC of helping other YC-funded startups. But we didn't invent that idea: it's just a slightly more concentrated form of existing Valley culture.

9. What a solitary task startups are. Architects are constantly interacting face to face with other people, whereas doing a technology startup, at least, tends to require long stretches of uninterrupted time to work. "You could do it in a box."

By inverting this list, we can get a portrait of the "normal" world. It's populated by people who talk a lot with one another as they work slowly but harmoniously on conservative, expensive projects whose destinations are decided in advance, and who carefully adjust their manner to reflect their position in the hierarchy.

That's also a fairly accurate description of the past. So startup culture may not merely be different in the way you'd expect any subculture to be, but a leading indicator.

The Trouble with the Segway

July 2009

The Segway hasn't delivered on its initial promise, to put it mildly. There are several reasons why, but one is that people don't want to be seen riding them. Someone riding a Segway looks like a dork.

My friend Trevor Blackwell built his own Segway, which we called the Segwell. He also built a one-wheeled version, the Eunicycle, which looks exactly like a regular unicycle till you realize the rider isn't pedaling. He has ridden them both to downtown Mountain View to get coffee. When he rides the Eunicycle, people smile at him. But when he rides the Segwell, they shout abuse from their cars: "Too lazy to walk, ya fuckin homo?"

Why do Segways provoke this reaction? The reason you look like a dork riding a Segway is that you look smug. You don't seem to be working hard enough.

Someone riding a motorcycle isn't working any harder. But because he's sitting astride it, he seems to be making an effort. When you're riding a Segway you're just standing there. And someone who's being whisked along while seeming to do no work—someone in a sedan chair, for example—can't help but look smug.

Try this thought experiment and it becomes clear: imagine something that worked like the Segway, but that you rode with one foot in front of the other, like a skateboard. That wouldn't seem nearly as uncool.

So there may be a way to capture more of the market Segway hoped to reach: make a version that doesn't look so easy for the rider. It would also be helpful if the styling was in the tradition of skateboards or bicycles rather than medical devices.

Curiously enough, what got Segway into this problem was that the company was itself a kind of Segway. It was too easy for them; they were too successful raising money. If they'd had to grow the company gradually, by iterating through several versions they sold to real users, they'd have learned pretty quickly that people looked stupid riding them. Instead they had enough to work in secret. They had focus groups aplenty, I'm sure, but they didn't have the people yelling insults out of cars. So they never realized they were zooming confidently down a blind alley.

Ramen Profitable

July 2009

Now that the term "ramen profitable" has become widespread, I ought to explain precisely what the idea entails.

Ramen profitable means a startup makes just enough to pay the founders' living expenses. This is a different form of profitability than startups have traditionally aimed for. Traditional profitability means a big bet is finally paying off, whereas the main importance of ramen profitability is that it buys you time. [1]

In the past, a startup would usually become profitable only after raising and spending quite a lot of money. A company making computer hardware might not become profitable for 5 years, during which they spent $50 million. But when they did they might have revenues of $50 million a year. This kind of profitability means the startup has succeeded.

Ramen profitability is the other extreme: a startup that becomes profitable after 2 months, even though its revenues are only $3000 a month, because the only employees are a couple 25 year old founders who can live on practically nothing. Revenues of $3000 a month do not mean the company has succeeded. But it does share something with the one that's profitable in the traditional way: they don't need to raise money to survive.

Ramen profitability is an unfamiliar idea to most people because it only recently became feasible. It's still not feasible for a lot of startups; it would not be for most biotech startups, for example; but it is for many software startups because they're now so cheap. For many, the only real cost is the founders' living expenses.

The main significance of this type of profitability is that you're no longer at the mercy of investors. If you're still losing money, then eventually you'll either have to raise more or shut down. Once you're ramen profitable this painful choice goes away. You can still raise money, but you don't have to do it now.

* * *

The most obvious advantage of not needing money is that you can get better terms. If investors know you need money, they'll sometimes take advantage of you. Some may even deliberately stall, because they know that as you run out of money you'll become increasingly pliable.

But there are also three less obvious advantages of ramen profitability. One is that it makes you more attractive to investors. If you're already profitable, on however small a scale, it shows that (a) you can get at least someone to pay you, (b) you're serious about building things people want, and (c) you're disciplined enough to keep expenses low.

This is reassuring to investors, because you've addressed three of their biggest worries. It's common for them to fund companies that have smart founders and a big market, and yet still fail. When these companies fail, it's usually because (a) people wouldn't pay for what they made, e.g. because it was too hard to sell to them, or the market wasn't ready yet, (b) the founders solved the wrong problem, instead of paying attention to what users needed, or (c) the company spent too much and burned through their funding before they started to make money. If you're ramen profitable, you're already avoiding these mistakes.

Another advantage of ramen profitability is that it's good for morale. A company tends to feel rather theoretical when you first start it. It's legally a company, but you feel like you're lying when you call it one. When people start to pay you significant amounts, the company starts to feel real. And your own living expenses are the milestone you feel most, because at that point the future flips state. Now survival is the default, instead of dying.

A morale boost on that scale is very valuable in a startup, because the moral weight of running a startup is what makes it hard. Startups are still very rare. Why don't more people do it? The financial risk? Plenty of 25 year olds save nothing anyway. The long hours? Plenty of people work just as long hours in regular jobs. What keeps people from starting startups is the fear of having so much responsibility. And this is not an irrational fear: it really is hard to bear. Anything that takes some of that weight off you will greatly increase your chances of surviving.

A startup that reaches ramen profitability may be more likely to succeed than not. Which is pretty exciting, considering the bimodal distribution of outcomes in startups: you either fail or make a lot of money.

The fourth advantage of ramen profitability is the least obvious but may be the most important. If you don't need to raise money, you don't have to interrupt working on the company to do it.

Raising money is terribly distracting. You're lucky if your productivity is a third of what it was before. And it can last for months.

I didn't understand (or rather, remember) precisely why raising money was so distracting till earlier this year. I'd noticed that startups we funded would usually grind to a halt when they switched to raising money, but I didn't remember exactly why till YC raised money itself. We had a comparatively easy time of it; the first people I asked said yes; but it took months to work out the details, and during that time I got hardly any real work done. Why? Because I thought about it all the time.

At any given time there tends to be one problem that's the most urgent for a startup. This is what you think about as you fall asleep at night and when you take a shower in the morning. And when you start raising money, that becomes the problem you think about. You only take one shower in the morning, and if you're thinking about investors during it, then you're not thinking about the product.

Whereas if you can choose when you raise money, you can pick a time when you're not in the middle of something else, and you can probably also insist that the round close fast. You may even be able to avoid having the round occupy your thoughts, if you don't care whether it closes.

* * *

Ramen profitable means no more than the definition implies. It does not, for example, imply that you're "bootstrapping" the startup—that you're never going to take money from investors. Empirically that doesn't seem to work very well. Few startups succeed without taking investment. Maybe as startups get cheaper it will become more common. On the other hand, the money is there, waiting to be invested. If startups need it less, they'll be able to get it on better terms, which will make them more inclined to take it. That will tend to produce an equilibrium. [2]

Another thing ramen profitability doesn't imply is Joe Kraus's idea that you should put your business model in beta when you put your product in beta. He believes you should get people to pay you from the beginning. I think that's too constraining. Facebook didn't, and they've done better than most startups. Making money right away was not only unnecessary for them, but probably would have been harmful. I do think Joe's rule could be useful for many startups, though. When founders seem unfocused, I sometimes suggest they try to get customers to pay them for something, in the hope that this constraint will prod them into action.

The difference between Joe's idea and ramen profitability is that a ramen profitable company doesn't have to be making money the way it ultimately will. It just has to be making money. The most famous example is Google, which initially made money by licensing search to sites like Yahoo.

Is there a downside to ramen profitability? Probably the biggest danger is that it might turn you into a consulting firm. Startups have to be product companies, in the sense of making a single thing that everyone uses. The defining quality of startups is that they grow fast, and consulting just can't scale the way a product can. [3] But it's pretty easy to make $3000 a month consulting; in fact, that would be a low rate for contract programming. So there could be a temptation to slide into consulting, and telling yourselves you're a ramen profitable startup, when in fact you're not a startup at all.

It's ok to do a little consulting-type work at first. Startups usually have to do something weird at first. But remember that ramen profitability is not the destination. A startup's destination is to grow really big; ramen profitability is a trick for not dying en route.

Notes

[1] The "ramen" in "ramen profitable" refers to instant ramen, which is just about the cheapest food available.

Please do not take the term literally. Living on instant ramen would be very unhealthy. Rice and beans are a better source of food. Start by investing in a rice cooker, if you don't have one.

Rice and Beans for 2n

  olive oil or butter
  n yellow onions
  other fresh vegetables; experiment
  3n cloves garlic
  n 12-oz cans white, kidney, or black beans
  n cubes Knorr beef or vegetable bouillon
  n teaspoons freshly ground black pepper
  3n teaspoons ground cumin
  n cups dry rice, preferably brown

Put rice in rice cooker. Add water as specified on rice package. (Default: 2 cups water per cup of rice.) Turn on rice cooker and forget about it.

Chop onions and other vegetables and fry in oil, over fairly low heat, till onions are glassy. Put in chopped garlic, pepper, cumin, and a little more fat, and stir. Keep heat low. Cook another 2 or 3 minutes, then add beans (don't drain the beans), and stir. Throw in the bouillon cube(s), cover, and cook on lowish heat for at least 10 minutes more. Stir vigilantly to avoid sticking.

If you want to save money, buy beans in giant cans from discount stores. Spices are also much cheaper when bought in bulk. If there's an Indian grocery store near you, they'll have big bags of cumin for the same price as the little jars in supermarkets.

[2] There's a good chance that a shift in power from investors to founders would actually increase the size of the venture business. I think investors currently err too far on the side of being harsh to founders. If they were forced to stop, the whole venture business would work better, and you might see something like the increase in trade you always see when restrictive laws are removed.

Investors are one of the biggest sources of pain for founders; if they stopped causing so much pain, it would be better to be a founder; and if it were better to be a founder, more people would do it.

[3] It's conceivable that a startup could grow big by transforming consulting into a form that would scale. But if they did that they'd really be a product company.

Thanks to Jessica Livingston for reading drafts of this.

Maker's Schedule, Manager's Schedule

July 2009

One reason programmers dislike meetings so much is that they're on a different type of schedule from other people. Meetings cost them more.

There are two types of schedule, which I'll call the manager's schedule and the maker's schedule. The manager's schedule is for bosses. It's embodied in the traditional appointment book, with each day cut into one hour intervals. You can block off several hours for a single task if you need to, but by default you change what you're doing every hour.

When you use time that way, it's merely a practical problem to meet with someone. Find an open slot in your schedule, book them, and you're done.

Most powerful people are on the manager's schedule. It's the schedule of command. But there's another way of using time that's common among people who make things, like programmers and writers. They generally prefer to use time in units of half a day at least. You can't write or program well in units of an hour. That's barely enough time to get started.

When you're operating on the maker's schedule, meetings are a disaster. A single meeting can blow a whole afternoon, by breaking it into two pieces each too small to do anything hard in. Plus you have to remember to go to the meeting. That's no problem for someone on the manager's schedule. There's always something coming on the next hour; the only question is what. But when someone on the maker's schedule has a meeting, they have to think about it.

For someone on the maker's schedule, having a meeting is like throwing an exception. It doesn't merely cause you to switch from one task to another; it changes the mode in which you work.

I find one meeting can sometimes affect a whole day. A meeting commonly blows at least half a day, by breaking up a morning or afternoon. But in addition there's sometimes a cascading effect. If I know the afternoon is going to be broken up, I'm slightly less likely to start something ambitious in the morning. I know this may sound oversensitive, but if you're a maker, think of your own case. Don't your spirits rise at the thought of having an entire day free to work, with no appointments at all? Well, that means your spirits are correspondingly depressed when you don't. And ambitious projects are by definition close to the limits of your capacity. A small decrease in morale is enough to kill them off.

Each type of schedule works fine by itself. Problems arise when they meet. Since most powerful people operate on the manager's schedule, they're in a position to make everyone resonate at their frequency if they want to. But the smarter ones restrain themselves, if they know that some of the people working for them need long chunks of time to work in.

Our case is an unusual one. Nearly all investors, including all VCs I know, operate on the manager's schedule. But Y Combinator runs on the maker's schedule. Rtm and Trevor and I do because we always have, and Jessica does too, mostly, because she's gotten into sync with us.

I wouldn't be surprised if there start to be more companies like us. I suspect founders may increasingly be able to resist, or at least postpone, turning into managers, just as a few decades ago they started to be able to resist switching from jeans to suits.

How do we manage to advise so many startups on the maker's schedule? By using the classic device for simulating the manager's schedule within the maker's: office hours. Several times a week I set aside a chunk of time to meet founders we've funded. These chunks of time are at the end of my working day, and I wrote a signup program that ensures all the appointments within a given set of office hours are clustered at the end. Because they come at the end of my day these meetings are never an interruption. (Unless their working day ends at the same time as mine, the meeting presumably interrupts theirs, but since they made the appointment it must be worth it to them.) During busy periods, office hours sometimes get long enough that they compress the day, but they never interrupt it.

When we were working on our own startup, back in the 90s, I evolved another trick for partitioning the day. I used to program from dinner till about 3 am every day, because at night no one could interrupt me. Then I'd sleep till about 11 am, and come in and work until dinner on what I called "business stuff." I never thought of it in these terms, but in effect I had two workdays each day, one on the manager's schedule and one on the maker's.

When you're operating on the manager's schedule you can do something you'd never want to do on the maker's: you can have speculative meetings. You can meet someone just to get to know one another. If you have an empty slot in your schedule, why not? Maybe it will turn out you can help one another in some way.

Business people in Silicon Valley (and the whole world, for that matter) have speculative meetings all the time. They're effectively free if you're on the manager's schedule. They're so common that there's distinctive language for proposing them: saying that you want to "grab coffee," for example.

Speculative meetings are terribly costly if you're on the maker's schedule, though. Which puts us in something of a bind. Everyone assumes that, like other investors, we run on the manager's schedule. So they introduce us to someone they think we ought to meet, or send us an email proposing we grab coffee. At this point we have two options, neither of them good: we can meet with them, and lose half a day's work; or we can try to avoid meeting them, and probably offend them.

Till recently we weren't clear in our own minds about the source of the problem. We just took it for granted that we had to either blow our schedules or offend people. But now that I've realized what's going on, perhaps there's a third option: to write something explaining the two types of schedule. Maybe eventually, if the conflict between the manager's schedule and the maker's schedule starts to be more widely understood, it will become less of a problem.

Those of us on the maker's schedule are willing to compromise. We know we have to have some number of meetings. All we ask from those on the manager's schedule is that they understand the cost.

Thanks to Sam Altman, Trevor Blackwell, Paul Buchheit, Jessica Livingston, and Robert Morris for reading drafts of this.

Related:

How to Do What You Love

Good and Bad Procrastination

Turkish Translation

French Translation