Y Combinator accepts roughly 1.5% of applicants. Here is what the other 98.5% mostly get wrong.
Figuring out how to get into Y Combinator is one of the most Googled questions in startup culture, and most of the answers are wrong. Not wrong in a subtle way, wrong in the way that gets you rejected at the application stage while you're convinced you're doing everything right. The official guide tells you to be concise. Every forum thread tells you to show traction. None of that is false, but treating the application as a test to pass rather than a conversation misses what YC partners are actually doing when they read it. They've read tens of thousands of these applications over the years, and they can spot a rehearsed answer in the first paragraph.
Start with what YC is actually selecting for. Garry Tan, YC's current president, has said this plainly in public: they're betting on founders, not ideas. The idea can change. Dropbox started as a different kind of file-sharing concept. Airbnb's original pitch was renting air mattresses in a San Francisco apartment, which most investors passed on because it sounded too small and too strange. What YC is reading for in every section of the application is whether the people behind it are the type who will figure things out when the original plan stops working.
The YC application has evolved over dozens of batches but its core questions stay consistent: what have you built, why this idea, what's your progress, why you. Each question is testing something specific, and the most common mistake is answering the surface version rather than the one underneath it. The application is deliberately short. There's no word count to hide behind. The only thing that actually works is an answer that makes a partner want to keep reading.
"Describe what you've built" is not a product description field. It's a test of how clearly you understand your own work. If you write about features, you're writing brochure copy. If you write about the problem you found and exactly what you did to solve it, you're writing for a partner who has ten minutes and no patience for vagueness. The clearest applications read like someone explaining something they genuinely figured out, not pitching a concept they think investors want to hear.
"Why did you pick this idea?" is where most founders blow it. Answers like "because the market is large" or "because we saw a gap" signal that you found the opportunity from the outside, the way a consultant maps whitespace. The answers that work come from founders who couldn't find a solution that already existed, or who spent years inside an industry and know exactly where the friction is. Brex is a clear illustration of what this looks like in an actual application. Henrique Dubugras and Pedro Franceschi had already built Pagar.me, a payments company in Brazil, before coming to the US and running directly into the wall that is getting corporate credit as a startup. Their application didn't need to claim founder-market fit. Their own history demonstrated it, without them having to say so.
On traction: what the numbers actually signal
The 1.5% acceptance rate sounds brutal, and it is. But the filter isn't purely about traction. YC funds teams at pre-seed and seed stage, which means some accepted founders have almost nothing built yet. What the application needs to show at early stages is momentum in the direction of evidence: a handful of real users beats a polished pitch with none, every time. Revenue beats users. But the deeper question underneath all of it is whether you've been doing anything at all, or whether the application is the first serious step you've taken on the idea.
This trips up technical founders in particular. A technically impressive product with no users reads worse than a rougher product with twenty paying customers. When Brian Chesky and Joe Gebbia applied for what became Airbnb, they had already rented their own apartment to conference attendees to make their rent. Not a product, not a startup, just a real transaction with real strangers. It meant they knew something about their customer that no founder pitching from a spreadsheet could know. The application question about progress is hunting for exactly that: have you left the building yet?
The video, the interview, and what partners are watching for
YC asks for a one-minute video introducing the founders. Most applicants treat it as another pitch opportunity, but it's really a preview of what it would be like to spend ten minutes with you in an interview room, which is exactly what happens if you're invited to the next stage. The best videos are unremarkable in production quality and remarkable in how clearly the founders talk. No script visible behind the camera. No cuts. Just founders who sound like they know what they're building and why they're the right people to do it.
Solo founders face a harder road statistically, not because YC is categorically against them, but because the combination of skills needed to build, sell, and hold things together through the first genuinely hard month is a lot to demonstrate in a single application. That doesn't mean don't apply. It means the solo founder application needs to cover ground that two co-founders would naturally split between them: technical depth, customer insight, and enough evidence of staying power that a partner reading it isn't left wondering if you'll quit.
YC interviews run ten minutes. The partners ask fast questions and they're not interested in a structured walkthrough of your deck. They want to know whether you understand your users better than anyone else, whether you know the number that matters most in your business right now, and whether you can handle an unexpected question without retreating to a rehearsed talking point. Partners who have done hundreds of these know the difference between a founder who is reasoning through an answer and one who is reciting from a script they rehearsed the night before.
The most common interview failure is a slow answer. If a partner asks how many active users you have and you say "well, it depends on how you define active," that's a red flag. Know four numbers cold before you walk in: weekly active users, revenue if any, month-on-month growth rate, and what conversion looks like from signup to retained user. Those four figures tell a partner more in thirty seconds than twenty minutes of narrative would.
What the official prep guides don't say clearly enough: YC is also watching founder dynamics in the room. If one founder does all the talking and the other looks down at the table, that gets noticed. Partners are thinking about who they'll be advising for the next three months of the batch, and possibly funding again later. The dynamic between founders in a ten-minute interview tends to be a fairly accurate preview of how they handle pressure when things aren't going well.
Frankly, the single biggest mistake founders make is treating the whole exercise as persuasion. The ones who get into Y Combinator are usually the ones who wrote an honest account of where they actually are, with enough real evidence under it that the partners don't need to take anything on faith. That's a different task than most applicants think they're doing, and it starts well before you open the form.
Also read: Build Your Startup Hiring Plan Before the Runway Runs Out • How to Negotiate a Term Sheet Before You Sign Away More Than You Know • How to cold email an investor and actually get a reply
