Half of YC is events in which all the startups participate. The other half happens in individual conversations with us.
There are 16 full-time partners: Sam Altman, Tim Brady, Paul Buchheit, Dalton Caldwell, Jared Friedman, Kevin Hale, Aaron Harris, Justin Kan, Carolynn Levy, Jon Levy, Jessica Livingston, Kat Mañalac, Kirsty Nathoo, Geoff Ralston, Michael Seibel, and Qasar Younis.
16位全职合伙人：Sam Altman, Tim Brady, Paul Buchheit, Dalton Caldwell, Jared Friedman, Kevin Hale, Aaron Harris, Justin Kan, Carolynn Levy, Jon Levy, Jessica Livingston, Kat Mañalac, Kirsty Nathoo, Geoff Ralston, Michael Seibel, and Qasar Younis.
There are also 10 part-time partners: Joe Gebbia, Elizabeth Iorns, Andrew Mason, Rick Morrison, Yuri Sagalov, Emmett Shear, Ben Silbermann, Ilya Sukhar, Peter Thiel, and Anne Wojcicki.
The individual conversations with the startups happen during office hours they book online. There’s no limit to the number of office hours they can book; startups talk to us as often as they want, which in practice varies enormously depending on the startup and where we are in the cycle. The office hour software is designed not just to book slots, but to show us if startups are getting enough time with us. The office hour request queue shows us how many founders need to talk to us; when it gets long, we schedule more blocks of office hours.
What we talk about at office hours also depends on the startup and where we are in the cycle. Usually we talk about whatever is the most urgent question right now. Sometimes, especially early on, the most urgent question is to figure out what the most urgent question should be. That’s less trivial than it sounds; we spend a lot of time telling founders what not to worry about.
(About 10% of the time we talk not about immediate problems but about the big vision for the company. You don’t have to be bound by this, but it’s good to have one. Some startups arrive with a big vision already, but most don’t. It’s a useful exercise to spend some time thinking about what the path would be from what a startup is doing now to a giant company, even if that’s not the current goal of the founders. Helping founders come up with these big visions is one of our strengths, because we’ve explored so much of the space of startup ideas that we know what’s over each hill.)
If the startup either hasn’t decided what to work on or wants to change their idea, then we talk about what the company should do. That usually means satisfying two constraints: something (1) users would want, that (2) the founders of this startup would be good at building. We have these types of conversations surprisingly often. Startups modify or even replace their ideas much more than outsiders realize. By Demo Day perhaps 15% are working on new ideas that grew out of conversations during office hours. Some of the most successful startups we’ve funded have.
If the question of what the company should do is settled, the most urgent question tends to be what to build first. Usually we advise startups to launch fast and iterate. This doesn’t apply to all startups (Clustrix, for example), but it works for most. The reason we advise startups to launch fast is that till you launch you’re designing for hypothetical or at best tame users, instead of actual ones. Once you launch you begin the conversation with real users, whose often surprising reactions to your product teach you what you should have been building.
Usually we advise startups to launch when they’ve built something with a quantum of utility—when they’ve built something sufficiently better than existing options that at least some users would say “I’m glad this appeared, because now I can finally do x.” If what you’ve built is a subset of existing technology at the same price, then users have no reason to try it, which means you don’t get to start the conversation with them. You need a quantum of utility to get a toehold.
Since there are a large number of points on the perimeter of most existing technologies at which one could push outward to create a quantum blister, what to build first is one of the most important questions we talk about. The general answer is to pick something where the product of how fast it can be built and how excited users would be about it is high. (How excited is distinct from how many would be excited.) But in practice the answer tends to be very specific to each startup. The goal is for office hours to end with the founders having a clear direction and saying “OK, we’ll go build x.” That goal is almost always achieved, though in practice about 10% of the time x turns out to be a bad idea and we have to go back to the drawing board.
Deciding what to build is not simply a technical question, because a startup is a company, not just a piece of software. The larval product should also have a larval business model. Like the product, this can change, but there should usually be an initial hypothesis about who to charge, how much to charge them, and how to reach them.
If a startup has built something that’s ready to launch, then the conversations at office hours tend to be about how to launch it. That’s two questions: how to present it to users, and how to present it to the press and investors. If the startup has a web site, we try to ensure it’s at least above a basic threshold of usability. Startup web sites tend to be confusing to users who show up for the first time with no idea what the company is doing (and not caring that much either), so it’s critical the site convey clearly and briefly what it is. Two partners, Garry Tan and Kevin Hale, are designers who can give advice about site design. As a writer, Paul Graham can often help startups come up with a concise explanation of what they do.
通常，创业团队的网站对第一次访问的用户（往往不知道也不关心这家公司是干啥的）总是充满迷惑，所以一个简单明了的总结非常重要。YC的两位合伙人，Garry Tan和Kevin Hale将作为设计师给网站提设计意见；而作为一名作家，Paul Graham常常帮团队网站写简明扼要的说明以解释其功能。
The second half of launching, explaining the company to the press, is also the beginning of explaining it to investors. When each startup is ready to launch, they stand in front of a whiteboard with a YC partner and figure out what the basic pitch should be. The founders usually photograph this and writes up an initial draft that we help revise. On the day they launch, we turn that into an email to whatever publication they’ve chosen (most often TechCrunch). They’re free to contact the publication themselves, of course, but the reasons we usually help them with the intial intro is because (a) we help them use plain language instead of the marketing-speak even technical founders often revert to when describing their own companies, and (b) journalists know we can’t bullshit them, or they’ll stop believing us.
The whiteboard pitch is useful for more than just launching. It becomes the basis of how the company will describe itself, particularly to investors. In fact it usually helps the founders themselves both clarify and agree with one another about the vision for the company.
If a company is launched and has users (about a quarter are before they come to YC), then the conversations at office hours tend to be about the various things that happen to actually launched companies. Most things that happen to newly launched startups are bad. But paradoxically, these disasters are precisely the reason to launch fast: they all represent problems you’re going to need to solve eventually, and the only way even to find out what they are is to launch. In practice they vary from technological bottlenecks to threats of lawsuits, but the most common problem is that users don’t like the product enough.
This is normal; it’s to learn where your initial hypothesis is wrong that you launch. Now at least you have some evidence to analyze. So for companies at this stage, the conversations at office hours tend to be about how to figure out from available evidence what users want, how to get more data about what they want, and how to reach more of them. It’s similar to the initial phase when we’re trying to figure out where on the perimeter of existing technology to break out, particularly in the sense that the search space is huge. Fortunately this is an area where our experience advising so many startups is especially useful. This is detective work, and we are experienced detectives.
As the 3 month cycle progresses, conversations at office hours start to be about investors as well as the company. Some questions all startups have, like what the fundraising strategy should be. How much should they raise, and from whom, and when should they start? The answers depend on the startup. Some startups are immediately attractive, and they’ll find it easy to raise money. Others are ugly ducklings, who will grow into swans in time, but initially should try to raise a comparatively small amount of angel money and use that to get the company to a more swanlike stage. 2.
It’s important for a startup to have an accurate estimate of their fundraising potential, because raising money is like choosing an angle of attack for a plane. If you try to climb too steeply you just stall. Similarly, if you try to raise too much for the stage you’re at, you’ll not only waste lots of time and end up with nothing to show for it, you’ll even jeopardize your chances of raising a smaller amount, because your initial leads will cool as you start to seem shelfworn. Fortunately this is another area where our experience helps. We can predict based on data about hundreds of previous startups what investors’ reactions to you will be and suggest an appropriate strategy.
Once fundraising starts, the conversations about it at office hours shift from the strategic to the tactical. Raising money is, unfortunately, usually a question of breaking deadlocks. Once you start to get hard commitments from investors, more investors want in. So the most important tactical question is usually who to close first, and how to convince them. This again is something where our experience helps, because we know a lot of the investors’ personalities from previous interactions. At this stage we do more than advise, though: we also talk directly to investors to help convince them. We can often do this more credibly than the founders themselves could, because we have to tell the truth; if we bullshit investors, they’ll stop believing us.
The goal of course is not just to get the startup money, but to get it from the investors who will help them most. Our knowledge of investors helps us identify those as well as convince them.
As Demo Day approaches, conversations at office hours start to be dominated by questions about presentations. This is the busiest part of the cycle for us. Since the startups are at different stages, they need to talk to us different amounts the rest of the cycle. But the week before Demo Day they all need to talk to us about what to say on Demo Day. So during that week office hours run together into one long rehearsal. We’re there most days well into the night as founders come in and rehearse successive versions of their presentations.
随着Demo Day越来越近，对话时间开始被与演讲相关的内容占据。这是YC最忙的一段时间。既然团队各自处在不同的阶段，他们在资助期之后也应该各自有各自的对话安排——然而在Demo Day之前的一周，他们关心的内容都是“我应该在Demo Day上演示什么”。于是这周的对话时间就连成了一次漫长的彩排——这段时间我们总是守到半夜，帮各个创始人一遍又一遍地迭代他们的演说。
Usually we begin with a whiteboard conversation like the one before launching, during which we figure out the most important points to convey. Then the startup makes a presentation based on that, we (and whatever other founders are around) watch it, and then they fix what’s broken and try again. These sessions make the presentations markedly better than they’d otherwise have been. Startups who come to YC a lot the week before Demo Day always have better presentations than the ones who for whatever reason don’t, even if their companies aren’t actually as good.
Office hours don’t stop after the 3 month cycle ends. In fact, the only thing that stops at the end of the 3 month cycle is the dinners. We have office hours year round, and startups from all previous cycles can book time whenever they want. (Except during interviews, and the first week of each cycle and the week before Demo Day, which we reserve for the startups in that cycle.) We still talk regularly with founders from the first YC batch in the summer of 2005.
Eventually we may hit some kind of limit on the number of people we can advise, but we haven’t hit it yet. There are only 4 weeks a year (2 each cycle) when so many people want office hours that we have to limit them to startups in the current batch. And the length of the office hour request queue shows us that, as of now at least, we’re never so overloaded that we can’t satisfy everyone who wants to talk to us.