October 2006In the Q & A period after a recent talk, someone asked what madestartups fail. After standing there gaping for a few seconds Irealized this was kind of a trick question. It's equivalent toasking how to make a startup succeed — if you avoid every cause offailure, you succeed — and that's too big a question to answer onthe fly.Afterwards I realized it could be helpful to look at the problemfrom this direction. If you have a list of all the things youshouldn't do, you can turn that into a recipe for succeeding justby negating. And this form of list may be more useful in practice.It's easier to catch yourself doing something you shouldn't thanalways to remember to do something you should.[1]In a sense there's just one mistake that kills startups: not makingsomething users want. If you make something users want, you'llprobably be fine, whatever else you do or don't do. And if youdon't make something users want, then you're dead, whatever elseyou do or don't do. So really this is a list of 18 things thatcause startups not to make something users want. Nearly all failurefunnels through that.1. Single FounderHave you ever noticed how few successful startups were founded byjust one person? Even companies you think of as having one founder,like Oracle, usually turn out to have more. It seems unlikely thisis a coincidence.What's wrong with having one founder? To start with, it's a voteof no confidence. It probably means the founder couldn't talk anyof his friends into starting the company with him. That's prettyalarming, because his friends are the ones who know him best.But even if the founder's friends were all wrong and the companyis a good bet, he's still at a disadvantage. Starting a startupis too hard for one person. Even if you could do all the workyourself, you need colleagues to brainstorm with, to talk you outof stupid decisions, and to cheer you up when things go wrong.The last one might be the most important. The low points in astartup are so low that few could bear them alone. When you havemultiple founders, esprit de corps binds them together in a waythat seems to violate conservation laws. Each thinks "I can't letmy friends down." This is one of the most powerful forces in humannature, and it's missing when there's just one founder.2. Bad LocationStartups prosper in some places and not others. Silicon Valleydominates, then Boston, then Seattle, Austin, Denver, and New York. Afterthat there's not much. Even in New York the number of startups percapita is probably a 20th of what it is in Silicon Valley. In townslike Houston and Chicago and Detroit it's too small to measure.Why is the falloff so sharp? Probably for the same reason it isin other industries. What's the sixth largest fashion center inthe US? The sixth largest center for oil, or finance, or publishing?Whatever they are they're probably so far from the top that it wouldbe misleading even to call them centers.It's an interesting question why cities become startup hubs, butthe reason startups prosper in them is probably the same as it isfor any industry: that's where the experts are. Standards arehigher; people are more sympathetic to what you're doing; the kindof people you want to hire want to live there; supporting industriesare there; the people you run into in chance meetings are in thesame business. Who knows exactly how these factors combine to booststartups in Silicon Valley and squish them in Detroit, but it'sclear they do from the number of startups per capita in each.3. Marginal NicheMost of the groups that apply to Y Combinator suffer from a commonproblem: choosing a small, obscure niche in the hope of avoidingcompetition.If you watch little kids playing sports, you notice that below acertain age they're afraid of the ball. When the ball comes nearthem their instinct is to avoid it. I didn't make a lot of catchesas an eight year old outfielder, because whenever a fly ball camemy way, I used to close my eyes and hold my glove up more forprotection than in the hope of catching it.Choosing a marginal project is the startup equivalent of my eightyear old strategy for dealing with fly balls. If you make anythinggood, you're going to have competitors, so you may as well facethat. You can only avoid competition by avoiding good ideas.I think this shrinking from big problems is mostly unconscious.It's not that people think of grand ideas but decide to pursuesmaller ones because they seem safer. Your unconscious won't evenlet you think of grand ideas. So the solution may be to think aboutideas without involving yourself. What would be a great idea forsomeone else to do as a startup?4. Derivative IdeaMany of the applications we get are imitations of some existingcompany. That's one source of ideas, but not the best. If youlook at the origins of successful startups, few were started inimitation of some other startup. Where did they get their ideas?Usually from some specific, unsolved problem the founders identified.Our startup made software for making online stores. When we startedit, there wasn't any; the few sites you could order from werehand-made at great expense by web consultants. We knew that ifonline shopping ever took off, these sites would have to be generatedby software, so we wrote some. Pretty straightforward.It seems like the best problems to solve are ones that affect youpersonally. Apple happened because Steve Wozniak wanted a computer,Google because Larry and Sergey couldn't find stuff online, Hotmailbecause Sabeer Bhatia and Jack Smith couldn't exchange email atwork.So instead of copying the Facebook, with some variation that theFacebook rightly ignored, look for ideas from the other direction.Instead of starting from companies and working back to the problemsthey solved, look for problems and imagine the company that mightsolve them. [2]What do people complain about? What do you wish there was?5. ObstinacyIn some fields the way to succeed is to have a vision of what youwant to achieve, and to hold true to it no matter what setbacks youencounter. Starting startups is not one of them. The stick-to-your-visionapproach works for something like winning an Olympic gold medal,where the problem is well-defined. Startups are more like science,where you need to follow the trail wherever it leads.So don't get too attached to your original plan, because it'sprobably wrong. Most successful startups end up doing somethingdifferent than they originally intended — often so different thatit doesn't even seem like the same company. You have to be preparedto see the better idea when it arrives. And the hardest part ofthat is often discarding your old idea.But openness to new ideas has to be tuned just right. Switchingto a new idea every week will be equally fatal. Is there some kindof external test you can use? One is to ask whether the ideasrepresent some kind of progression. If in each new idea you'reable to re-use most of what you built for the previous ones, thenyou're probably in a process that converges. Whereas if you keeprestarting from scratch, that's a bad sign.Fortunately there's someone you can ask for advice: your users. Ifyou're thinking about turning in some new direction and your usersseem excited about it, it's probably a good bet.6. Hiring Bad ProgrammersI forgot to include this in the early versions of the list,because nearly all the founders I know are programmers. This isnot a serious problem for them. They might accidentally hire someonebad, but it's not going to kill the company. In a pinch they cando whatever's required themselves.But when I think about what killed most of the startups in thee-commerce business back in the 90s, it was bad programmers. A lotof those companies were started by business guys who thought theway startups worked was that you had some clever idea and then hiredprogrammers to implement it. That's actually much harder than itsounds — almost impossibly hard in fact — because business guyscan't tell which are the good programmers. They don't even get ashot at the best ones, because no one really good wants a jobimplementing the vision of a business guy.In practice what happens is that the business guys choose peoplethey think are good programmers (it says here on his resume thathe's a Microsoft Certified Developer) but who aren't. Then they'remystified to find that their startup lumbers along like a World WarII bomber while their competitors scream past like jet fighters.This kind of startup is in the same position as a big company,but without the advantages.So how do you pick good programmers if you're not a programmer? Idon't think there's an answer. I was about to say you'd have tofind a good programmer to help you hire people. But if you can'trecognize good programmers, how would you even do that?7. Choosing the Wrong PlatformA related problem (since it tends to be done by bad programmers)is choosing the wrong platform. For example, I think a lot ofstartups during the Bubble killed themselves by deciding to buildserver-based applications on Windows. Hotmail was still runningon FreeBSD for years after Microsoft bought it, presumably becauseWindows couldn't handle the load. If Hotmail's foundershad chosen to use Windows, they would have been swamped.PayPal only just dodged this bullet. After they merged with X.com,the new CEO wanted to switch to Windows — even after PayPal cofounderMax Levchin showed that their software scaled only 1% as well onWindows as Unix. Fortunately for PayPal they switched CEOs instead.Platform is a vague word. It could mean an operating system, or aprogramming language, or a "framework" built on top of a programminglanguage. It implies something that both supports and limits, likethe foundation of a house.The scary thing about platforms is that there are always some thatseem to outsiders to be fine, responsible choices and yet, likeWindows in the 90s, will destroy you if you choose them. Javaapplets were probably the most spectacular example. This wassupposed to be the new way of delivering applications. Presumablyit killed just about 100% of the startups who believed that.How do you pick the right platforms? The usual way is to hire goodprogrammers and let them choose. But there is a trick you coulduse if you're not a programmer: visit a top computer sciencedepartment and see what they use in research projects.8. Slowness in LaunchingCompanies of all sizes have a hard time getting software done. It'sintrinsic to the medium; software is always 85% done. It takes aneffort of will to push through this and get something released tousers.[3]Startups make all kinds of excuses for delaying their launch. Mostare equivalent to the ones people use for procrastinating in everydaylife. There's something that needs to happen first. Maybe. Butif the software were 100% finished and ready to launch at the pushof a button, would they still be waiting?One reason to launch quickly is that it forces you to actuallyfinish some quantum of work. Nothing is truly finished till it'sreleased; you can see that from the rush of work that's alwaysinvolved in releasing anything, no matter how finished you thoughtit was. The other reason you need to launch is that it's only bybouncing your idea off users that you fully understand it.Several distinct problems manifest themselves as delays in launching:working too slowly; not truly understanding the problem; fear ofhaving to deal with users; fear of being judged; working on toomany different things; excessive perfectionism. Fortunately youcan combat all of them by the simple expedient of forcing yourselfto launch something fairly quickly.9. Launching Too EarlyLaunching too slowly has probably killed a hundred times morestartups than launching too fast, but it is possible to launch toofast. The danger here is that you ruin your reputation. You launchsomething, the early adopters try it out, and if it's no good theymay never come back.So what's the minimum you need to launch? We suggest startups thinkabout what they plan to do, identify a core that's both (a) usefulon its own and (b) something that can be incrementally expandedinto the whole project, and then get that done as soon as possible.This is the same approach I (and many other programmers) use forwriting software. Think about the overall goal, then start bywriting the smallest subset of it that does anything useful. Ifit's a subset, you'll have to write it anyway, so in the worst caseyou won't be wasting your time. But more likely you'll find thatimplementing a working subset is both good for morale and helps yousee more clearly what the rest should do.The early adopters you need to impress are fairly tolerant. Theydon't expect a newly launched product to do everything; it just hasto do something.10. Having No Specific User in MindYou can't build things users like without understanding them. Imentioned earlier that the most successful startups seem to havebegun by trying to solve a problem their founders had. Perhapsthere's a rule here: perhaps you create wealth in proportion to howwell you understand the problem you're solving, and the problemsyou understand best are your own. [4]That's just a theory. What's not a theory is the converse: ifyou're trying to solve problems you don't understand, you're hosed.And yet a surprising number of founders seem willing toassume that someone, they're not sure exactly who, will want whatthey're building. Do the founders want it? No, they're not thetarget market. Who is? Teenagers. People interested in localevents (that one is a perennial tarpit). Or "business" users. Whatbusiness users? Gas stations? Movie studios? Defense contractors?You can of course build something for users other than yourself.We did. But you should realize you're stepping into dangerousterritory. You're flying on instruments, in effect, so you should(a) consciously shift gears, instead of assuming you can rely onyour intuitions as you ordinarily would, and (b) look at theinstruments.In this case the instruments are the users. When designing forother people you have to be empirical. You can no longer guesswhat will work; you have to find users and measure their responses.So if you're going to make something for teenagers or "business"users or some other group that doesn't include you, you have to beable to talk some specific ones into using what you're making. Ifyou can't, you're on the wrong track.11. Raising Too Little MoneyMost successful startups take funding at some point. Like havingmore than one founder, it seems a good bet statistically. How muchshould you take, though?Startup funding is measured in time. Every startup that isn'tprofitable (meaning nearly all of them, initially) has a certainamount of time left before the money runs out and they have to stop.This is sometimes referred to as runway, as in "How much runway doyou have left?" It's a good metaphor because it reminds you thatwhen the money runs out you're going to be airborne or dead.Too little money means not enough to get airborne. What airbornemeans depends on the situation. Usually you have to advance to avisibly higher level: if all you have is an idea, a working prototype;if you have a prototype, launching; if you're launched, significantgrowth. It depends on investors, because until you're profitablethat's who you have to convince.So if you take money from investors, you have to take enough to getto the next step, whatever that is.[5]Fortunately you have somecontrol over both how much you spend and what the next step is. Weadvise startups to set both low, initially: spend practicallynothing, and make your initial goal simply to build a solid prototype.This gives you maximum flexibility.12. Spending Too MuchIt's hard to distinguish spending too much from raising too little.If you run out of money, you could say either was the cause. Theonly way to decide which to call it is by comparison with otherstartups. If you raised five million and ran out of money, youprobably spent too much.Burning through too much money is not as common as it used to be.Founders seem to have learned that lesson. Plus it keeps gettingcheaper to start a startup. So as of this writing few startupsspend too much. None of the ones we've funded have. (And not justbecause we make small investments; many have gone on to raise furtherrounds.)The classic way to burn through cash is by hiring a lot of people.This bites you twice: in addition to increasing your costs, it slowsyou down—so money that's getting consumed faster has to lastlonger. Most hackers understand why that happens; Fred Brooksexplained it in The Mythical Man-Month.We have three general suggestions about hiring: (a) don't do it ifyou can avoid it, (b) pay people with equity rather than salary,not just to save money, but because you want the kind of people whoare committed enough to prefer that, and (c) only hire people whoare either going to write code or go out and get users, becausethose are the only things you need at first.13. Raising Too Much MoneyIt's obvious how too little money could kill you, but is there sucha thing as having too much?Yes and no. The problem is not so much the money itself as whatcomes with it. As one VC who spoke at Y Combinator said, "Once youtake several million dollars of my money, the clock is ticking."If VCs fund you, they're not going to let you just put the moneyin the bank and keep operating as two guys living on ramen. Theywant that money to go to work. [6]At the very least you'll moveinto proper office space and hire more people. That will changethe atmosphere, and not entirely for the better. Now most of yourpeople will be employees rather than founders. They won't be ascommitted; they'll need to be told what to do; they'll start toengage in office politics.When you raise a lot of money, your company moves to the suburbsand has kids.Perhaps more dangerously, once you take a lot of money it getsharder to change direction. Suppose your initial plan was to sellsomething to companies. After taking VC money you hire a salesforce to do that. What happens now if you realize you should bemaking this for consumers instead of businesses? That's a completelydifferent kind of selling. What happens, in practice, is that youdon't realize that. The more people you have, the more you staypointed in the same direction.Another drawback of large investments is the time they take. Thetime required to raise money grows with the amount.[7]When theamount rises into the millions, investors get very cautious. VCsnever quite say yes or no; they just engage you in an apparentlyendless conversation. Raising VC scale investments is thus a hugetime sink — more work, probably, than the startup itself. And youdon't want to be spending all your time talking to investors whileyour competitors are spending theirs building things.We advise founders who go on to seek VC money to take the firstreasonable deal they get. If you get an offer from a reputablefirm at a reasonable valuation with no unusually onerous terms,just take it and get on with building the company.[8]Who caresif you could get a 30% better deal elsewhere? Economically, startupsare an all-or-nothing game. Bargain-hunting among investors is awaste of time.14. Poor Investor ManagementAs a founder, you have to manage your investors. You shouldn'tignore them, because they may have useful insights. But neithershould you let them run the company. That's supposed to be yourjob. If investors had sufficient vision to run the companiesthey fund, why didn't they start them?Pissing off investors by ignoring them is probably less dangerousthan caving in to them. In our startup, we erred on the ignoringside. A lot of our energy got drainedaway in disputes with investors instead of going into the product.But this was less costly than giving in, which would probably havedestroyed the company. If the founders know what they're doing,it's better to have half their attention focused on the productthan the full attention of investors who don't.How hard you have to work on managing investors usually depends onhow much money you've taken. When you raise VC-scale money, theinvestors get a great deal of control. If they have a board majority,they're literally your bosses. In the more common case, wherefounders and investors are equally represented and the decidingvote is cast by neutral outside directors, all the investors haveto do is convince the outside directors and they control the company.If things go well, this shouldn't matter. So long as you seem tobe advancing rapidly, most investors will leave you alone. Butthings don't always go smoothly in startups. Investors have madetrouble even for the most successful companies. One of the mostfamous examples is Apple, whose board made a nearly fatal blunderin firing Steve Jobs. Apparently even Google got a lot of grieffrom their investors early on.15. Sacrificing Users to (Supposed) ProfitWhen I said at the beginning that if you make something users want,you'll be fine, you may have noticed I didn't mention anything abouthaving the right business model. That's not because making moneyis unimportant. I'm not suggesting that founders start companieswith no chance of making money in the hope of unloading them beforethey tank. The reason we tell founders not to worry about thebusiness model initially is that making something people want isso much harder.I don't know why it's so hard to make something people want. Itseems like it should be straightforward. But you can tell it mustbe hard by how few startups do it.Because making something people want is so much harder than makingmoney from it, you should leave business models for later, just asyou'd leave some trivial but messy feature for version 2. In version1, solve the core problem. And the core problem in a startup ishow to create wealth (= how much people want something x the numberwho want it), not how to convert that wealth into money.The companies that win are the ones that put users first. Google,for example. They made search work, then worried about how to makemoney from it. And yet some startup founders still think it'sirresponsible not to focus on the business model from the beginning.They're often encouraged in this by investors whose experience comesfrom less malleable industries.It is irresponsible not to think about business models. It'sjust ten times more irresponsible not to think about the product.16. Not Wanting to Get Your Hands DirtyNearly all programmers would rather spend their time writing codeand have someone else handle the messy business of extracting moneyfrom it. And not just the lazy ones. Larry and Sergey apparentlyfelt this way too at first. After developing their new searchalgorithm, the first thing they tried was to get some other companyto buy it.Start a company? Yech. Most hackers would rather just have ideas.But as Larry and Sergey found, there's not much of a market forideas. No one trusts an idea till you embody it in a product anduse that to grow a user base. Then they'll pay big time.Maybe this will change, but I doubt it will change much. There'snothing like users for convincing acquirers. It's not just thatthe risk is decreased. The acquirers are human, and they have ahard time paying a bunch of young guys millions of dollars just forbeing clever. When the idea is embodied in a company with a lotof users, they can tell themselves they're buying the users ratherthan the cleverness, and this is easier for them to swallow.[9]If you're going to attract users, you'll probably have to get upfrom your computer and go find some. It's unpleasant work, but ifyou can make yourself do it you have a much greater chance ofsucceeding. In the first batch of startups we funded, in the summerof 2005, most of the founders spent all their time building theirapplications. But there was one who was away half the time talkingto executives at cell phone companies, trying to arrange deals.Can you imagine anything more painful for a hacker?[10]But itpaid off, because this startup seems the most successful of thatgroup by an order of magnitude.If you want to start a startup, you have to face the fact that youcan't just hack. At least one hacker will have to spend some ofthe time doing business stuff.17. Fights Between FoundersFights between founders are surprisingly common. About 20% of thestartups we've funded have had a founder leave. It happens so oftenthat we've reversed our attitude to vesting. We still don't requireit, but now we advise founders to vest so there will be an orderlyway for people to quit.A founder leaving doesn't necessarily kill a startup, though. Plentyof successful startups have had that happen. [11]Fortunately it'susually the least committed founder who leaves. If there are threefounders and one who was lukewarm leaves, big deal. If you havetwo and one leaves, or a guy with critical technical skills leaves,that's more of a problem. But even that is survivable. Bloggergot down to one person, and they bounced back.Most of the disputes I've seen between founders could have beenavoided if they'd been more careful about who they started a companywith. Most disputes are not due to the situation but the people.Which means they're inevitable. And most founders who've beenburned by such disputes probably had misgivings, which they suppressed,when they started the company. Don't suppress misgivings. It'smuch easier to fix problems before the company is started thanafter. So don't include your housemate in your startup becausehe'd feel left out otherwise. Don't start a company with someoneyou dislike because they have some skill you need and you worry youwon't find anyone else. The people are the most important ingredientin a startup, so don't compromise there.18. A Half-Hearted EffortThe failed startups you hear most about are the spectacularflameouts. Those are actually the elite of failures. The mostcommon type is not the one that makes spectacular mistakes, but theone that doesn't do much of anything — the one we never even hearabout, because it was some project a couple guys started on theside while working on their day jobs, but which never got anywhereand was gradually abandoned.Statistically, if you want to avoid failure, it would seem like themost important thing is to quit your day job. Most founders offailed startups don't quit their day jobs, and most founders ofsuccessful ones do. If startup failure were a disease, the CDCwould be issuing bulletins warning people to avoid day jobs.Does that mean you should quit your day job? Not necessarily. I'mguessing here, but I'd guess that many of these would-be foundersmay not have the kind of determination it takes to start a company,and that in the back of their minds, they know it. The reason theydon't invest more time in their startup is that they know it's abad investment.[12]I'd also guess there's some band of people who could have succeededif they'd taken the leap and done it full-time, but didn't. I haveno idea how wide this band is, but if the winner/borderline/hopelessprogression has the sort of distribution you'd expect, the numberof people who could have made it, if they'd quit their day job, isprobably an order of magnitude larger than the number who do makeit.[13]If that's true, most startups that could succeed fail because thefounders don't devote their whole efforts to them. That certainlyaccords with what I see out in the world. Most startups fail becausethey don't make something people want, and the reason most don'tis that they don't try hard enough.In other words, starting startups is just like everything else.The biggest mistake you can make is not to try hard enough. To theextent there's a secret to success, it's not to be in denial aboutthat.Notes[1]This is not a complete list of the causes of failure,just those you can control. There are also several you can't,notably ineptitude and bad luck.[2]Ironically, one variant of the Facebook that might work is afacebook exclusively for college students.[3]Steve Jobs tried to motivate people by saying "Real artistsship." This is a fine sentence, but unfortunately not true. Manyfamous works of art are unfinished. It's true in fields that havehard deadlines, like architecture and filmmaking, but even therepeople tend to be tweaking stuff till it's yanked out of theirhands.[4]There's probably also a second factor: startup founders tendto be at the leading edge of technology, so problems they face areprobably especially valuable.[5]You should take more than you think you'll need, maybe 50% to100% more, because software takes longer to write and deals longerto close than you expect.[6]Since people sometimes call us VCs, I should add that we'renot. VCs invest large amounts of other people's money. We investsmall amounts of our own, like angel investors.[7]Not linearly of course, or it would take forever to raise fivemillion dollars. In practice it just feels like it takes forever.Though if you include the cases where VCs don't invest, it wouldliterally take forever in the median case. And maybe we should,because the danger of chasing large investments is not just thatthey take a long time. That's the best case. The real dangeris that you'll expend a lot of time and get nothing.[8]Some VCs will offer you an artificially low valuation to seeif you have the balls to ask for more. It's lame that VCs playsuch games, but some do. If you're dealing with one of those youshould push back on the valuation a bit.[9]Suppose YouTube's founders had gone to Google in 2005 and toldthem "Google Video is badly designed. Give us $10 million and we'lltell you all the mistakes you made." They would have gottenthe royal raspberry. Eighteen months later Google paid $1.6 billionfor the same lesson, partly because they could then tell themselvesthat they were buying a phenomenon, or a community, or some vaguething like that.I don't mean to be hard on Google. They did better than theircompetitors, who may have now missed the video boat entirely.[10]Yes, actually: dealing with the government. But phone companiesare up there.[11]Many more than most people realize, because companies don't advertisethis. Did you know Apple originally had three founders?[12]I'm not dissing these people. I don't have the determinationmyself. I've twice come close to starting startups since Viaweb,and both times I bailed because I realized that without the spurof poverty I just wasn't willing to endure the stress of a startup.[13]So how do you know whether you're in the category of peoplewho should quit their day job, or the presumably larger one whoshouldn't? I got to the point of saying that this was hard to judgefor yourself and that you should seek outside advice, before realizingthat that's what we do. We think of ourselves as investors, butviewed from the other direction Y Combinator is a service foradvising people whether or not to quit their day job. We could bemistaken, and no doubt often are, but we do at least bet money onour conclusions.Thanks to Sam Altman, Jessica Livingston, Greg McAdoo, and Robert Morris for reading drafts of this.
