Twice a year Y Combinator takes applications for funding. I thought it might helpapplicants if I explained what we look for when weread them.
Probably the biggest thing people don』t understand about theprocess is the importance of expressing yourself clearly. Everyyear we get some applications that are obviously good, some thatare obviously bad, and in the middle a huge number where we justcan』t tell. The idea seems kind of promising, but it』s notexplained well enough for us to understand it. The founders seemlike they might be good, but we don』t get a clear enough picture ofthem to say for sure.
I suspect for every group we invite to interviews, there are oneor two more that are just as good but that we pass over becausethey don』t manage to convey how good they are. If that』s true,another way to say it is that, of groups good enough to make it tointerviews, more than half blow the application.
If we get 1000 applications and have 10 days to read them, wehave to read about 100 a day. That means a YC partner who readsyour application will on average have already read 50 that day andhave 50 more to go. Yours has to stand out. So you have to beexceptionally clear and concise. Whatever you have to say, give itto us right in the first sentence, in the simplestpossible terms.
All the YC partners read applications. We each do it separately,to avoid groupthink, so I』m not sure exactly what the others do,but it』s probably similar to what I do.
The first thing I notice when I look at an application is theusername it was submitted under. If it』s one I recognize for makingthoughtful comments on Hacker News, I give theapplication extra attention. I don』t do this in a systematic way;there』s no software that ranks applications by karma or anythinglike that; it』s more that I can』t help noticing namesI know.
The other YC partners are less influenced by this, because theydon』t read HN as much. But in borderline cases I often find myselfsaying 「We have to interview so-and-so. I know him from Hacker Newsand he』s clearly smart.」 (I have a good deal of anecdotal evidencethis heuristic works, but I』ve never tried to measure it.)
The first question I look at is, 「What is your company going tomake?」 This isn』t the question I care most about, but I look at itfirst because I need something to hang the application on inmy mind.
The best answers are the most matter of fact. It』s a mistake touse marketing-speak to make your idea sound moreexciting. We』re immune to marketing-speak; to us it』s justnoise. 1. So don』t begin your answer withsomething like
We are going to transform the relationship between individualsand information.
That sounds impressive, but it conveys nothing. It could be adescription of any technology company. Are you going to build asearch engine? Database software? A router? I haveno idea.
One test of whether you』re explaining your idea effectively isto ask how close the reader is to reproducing it. After readingthat sentence I』m no closer than I was before, so its content iseffectively zero. Another mistake is to begin with a sweepingintroductory paragraph about how important the problem is:
Information is the lifeblood of the modern organization. Theability to channel information quickly and efficiently to those whoneed it is critical to a company』s success. A company that achievesan edge in the efficient use of information will, all other thingsbeing equal, have a significant edge over competitors.
Again, zero content; after reading this, the reader is no closerto reproducing your project than before. A good answer would besomething like:
A database with a wiki-like interface, combined with a graphicalUI for controlling who can see and edit what.
I』m not convinced yet that this will be the next Google, but atleast I』m starting to engage with it. I』m thinking what such athing would be like.
One reason founders resist giving matter-of-fact descriptions isthat they seem to constrain your potential. 「But it』s so much morethan a database with a wiki UI!」 The problem is, the lessconstraining your description, the less you』re saying. So it』sbetter to err on the side of matter-of-factness.
We advise startups presenting at Demo Day to do the same. Betterto start with an overly narrow description of your project than tryto describe it in its full generality and lose the audiencecompletely. If there』s a simple one-sentence description of whatyou』re doing that only conveys half your potential, that』s actuallypretty good. You』re halfway to your destination in just thefirst sentence.
One good trick for describing a project concisely is to explainit as a variant of something the audience already knows. It』s likeWikipedia, but within an organization. It』s like an answeringservice, but for email. It』s eBay for jobs. This form ofdescription is wonderfully efficient. Don』t worry that it will makeyour idea seem 「derivative.」 Some of the best ideas in historybegan by sticking together two existing ideas no one realized couldbe combined.
After spending 20 seconds or so trying to understand the idea, Iskip down to look at the founders. My initial goal is to figure outwhat kind of group I』m dealing with.
Three friends about to graduate from college? Two colleagues whowork together at a big company and want to jump ship? Are they allprogrammers? A mix of programmers and business people? There aremaybe 20 or 30 different configurations of founders.
Once I know what type of group I have, I try to figure out howgood an instance of that type it is. The most important questionfor deciding that is
Please tell us in one or two sentences about somethingimpressive that each founder has built or achieved.
To me this is the most important question on the application.It』s deliberately open-ended; there』s no one type of answer we』relooking for. It could be that you did really well in school, orthat you wrote a highly-regarded piece of software, or that youpaid your own way through college after leaving home at 16. It』snot the type of achievement that matters so much as the magnitude.Succeeding in a startup is, in the most literal sense,extraordinary, so we』re looking for people able to doextraordinary things.
As with all questions on the application, the best answers arethe most specific. A surprising number of people answer withsomething like:
Jordan is an exceptionally dedicated person who gives 100%effort to every project he undertakes.
This kind of generic claim carries no weight. A single, specificexample would be much more convincing. You probably shouldn』t listthe startup itself as your most impressive achievement. We alreadyknow you』ve created that. Why waste the opportunity to brag aboutsomething else?
If there』s no one thing about you that you feel stands out, whatshould you list? I』d go with whatever you』ve done that was thehardest—-preferably (though not necessarily) the hardestintellectually. It doesn』t matter if it』s not the sort of thingyou』d put on a resume. We』re not looking for the same things asHR departments.
If the founders seem promising, I』ll now spend more time tryingto understand the idea. I care more about the founders than theidea, because most of the startups we fund will change theiridea significantly.
If a group of founders seemed impressive enough, I』d fund themwith no idea. But a really good idea will also get ourattention—-not because of the idea per se, but because it』sevidence the founders are smart.
Just as what we look for in founders is not the type ofachievement but the magnitude, what we look for in ideas is not thetype of idea but the level of insight you have about it. You』regoing to start an auction site? That could be a good idea or a badidea. What matters is how you』re going to hold your own againsteBay. What』s going to be distinctive about your solution?
It』s a common mistake to say the distinctive thing about yoursolution will be that it』s well-designed and easy to use. That isnot an insight. You』re just claiming you』re going to execute well.Whoever wrote the current software was presumably also trying to.So you have to be more specific. Exactly what are you going to dothat will make your software easier to use? And will that beenough? The reason a lot of big companies』 software sucks is thatthey have some kind of natural monopoly. Unless you have a plan forcracking it, it won』t make any difference if yoursis better.
We don』t mind if you』re doing something that will face seriousobstacles. In fact, we like that. The best startup ideas aregenerally outliers that seem crazy to most people initially. But wewant to see that you』re aware of the obstacles, and have at least atheory about how to overcome them. We』d be delighted to get anapplication that answered the question 「What are you going tomake?」 with
A new search engine to compete with Google.
so long as this was followed by
We know that sounds impossible, but we think we can get atoehold initially by…
Wouldn』t you be interested at this point? Even if the plan hadonly a 1% chance of working, it would be worthbacking. 2.
Whereas if we can see obstacles to your idea that you don』t seemto have considered, that』s a bad sign. This is your idea. You』vehad days, at least, to think about it, and we』ve only had a coupleminutes. We shouldn』t be able to come up with objections youhaven』t thought of.
Paradoxically, it is for this reason better to disclose all theflaws in your idea than to try to conceal them. If we think of aproblem you don』t mention, we』ll assume it』s because you haven』tthought of it. And since we care more about you than the idea, it』sa mistake to risk sacrificing yourself to make the ideaseem better.
If the founders seem promising and the idea is interesting, I』llnow spend a lot more time on the application.
I』ll take a look at the video, if there isone. (Statistically we』re much more likely to interview people whosubmit a video.) I』ll check out the demo. And I』ll look at answersto some of the more mundane questions, like thestock allocation.
If the founders seem promising but the idea doesn』t, I check thequestion near the end that asks what other ideas the founders had.It』s quite common for us to fund groups to work on ideas theylisted as alternates.
There』s one question that acts like a wildcard, at leastfor me:
Please tell us about the time you most successfully hacked some(non-computer) system to your advantage.
If this wasn』t already clear, we』re not looking for the sort ofobedient, middle-of-the-road people that big companies tend tohire. We』re looking for people who like to beat the system. So ifthe answer to this question is good enough, it will make me go backand take a second look at an application that otherwise seemedunpromising. In fact, I think there are people we』ve invited tointerviews mainly on the strength of their answer tothis question.
Generally, the advice I』d give to applicants is: help us out.Investors are optimists. We want to believe you』re great. Mostpeople you meet in everyday life don』t.
If you go around saying you』re going to start the next Google,most people』s initial reaction will be skepticism. Partly becausethe odds of succeeding are low, so skepticism is the safe bet, butalso because most people are threatened by ambition: you seem to betrying to put yourself above them, even if that isn』tyour intention.
Investors are different—-not because they』re more generousspirited than other people, but because they get equity. Tellinvestors you』re going to start the next Google and theyimmediately perk up. They don』t default to skepticism, because theylike risky bets. And they don』t feel like you』re trying to putyourself above them, because they hope to be drawn upwith you.
Like all investors, we want to believe. So help us believe. Ifthere』s something about you that stands out, or some specialinsight you have into the problem you plan to work on, make sure wesee it.
The best way to do that is simply to be concise. You don』t haveto sell us on you. We』ll sell ourselves, if we can just understandyou. But every unnecessary word in your application subtracts fromthe effect of the necessary ones. So before submitting yourapplication, print it out and take a red pen and cross out everyword you don』t need. And in what』s left be as specific and asmatter-of-fact as you can.