Blog

Some thoughts on writing a good Y Combinator application

Jonathan Parisot
Jonathan Parisot
Mar 09, 2020 · 7 min read
Conference

At Actiondesk, we'd applied 4 times to Y Combinator and got in the 4th time and participated to YC S19. Yes, that means we failed 3 times. You might think that disqualifies us from giving advice and you might be right 🤷. I've also helped ~20 teams with their YC application last batch. YC was a game changer for us. Here's a few things we've learnt writing 4 applications that could help you write yours.

1) Be extremely concise

YC partners review tons of applications per day. Imagine having read 100 applications, reading a new one and seeing 10 lines of response to each question. You already have a negative bias toward that application.

On the other hand, if you see short responses, you'll want to read it more.

As an applicant, you want to make it easy for the reviewer. My concrete advice:

  • Most responses should be less than 1-2 sentences.
  • You can have 2-3 wildcards. Put more content / longer responses when that content really differentiates you.

You need to be so concise that it makes you uncomfortable. You need to feel that there are so many things you could have put that you didn't.

2) Do not waste words and energy on useless points

The way I see it is as follows. Imagine the reviewer has a limited quantity of energy to spend on each application.

The longer a response, the more energy it takes.

You want to save that energy for the points that make you strong. See next point

3) Reviewers look for a reason to choose you rather than a reason to eliminate you - Have this in mind.

A YC application is not where you want to be average at everything, it's where you want to be exceptional at least at one thing. eg

  • Being a solo founder or not having known your cofounder for long is not a problem if you have tremendous growth
  • Not having growth (or even a product) might be okay if you have a tremendous team
  • etc

The point is to highlight what makes you good rather than being defensive on the points where you're okish.

Concrete example: On the question on how long the founders had known each other. In Actiondesk case, it was not the case. We thought it was a problem so we were defensive and tried to explain "we haven't known each other for very long but we get along super well and are very complementary, bla bla bla"

Don't do that, just state the fact "we've known each other for 2 months" and move on. You'll make the difference on other questions.

4) Identify what makes you different and make sure it's highlighted

Related to the point before, you need only one thing for the reviewer to think "humm, we should talk to those people". Here's a non exhaustive list of what it could be (no particular order):

  • Strong growth (that's the best to have)
  • Strong usage of your product
  • Founders who've built something meaningful together
  • Exceptional academic or professional background of founders
  • Founders who have a particularly strong insight related to the problem they set out to solve
  • In the case of no crazy growth / usage, impressive product
  • etc

Find what's yours and make sure it's extra clear in your application. That's the one thing you absolutely want the reviewer to remember.

5) Launch before

YC likes founders who try. If you haven't launched yet, instead of putting the usual "we'll launch in x weeks", launch before. Maybe your product is really not ready, in that case, still launch and start gathering a waitlist of potential users.

  • "We have 20 users who filled a 3min typeform and want to use the product" is better than "we'll launch in x weeks"

Ideally, though, it'd be great to have users, even 1 can make the difference. Think about how you can achieve this. This is probably the highest leverage activity you can work on to improve your chances of getting in.

Plus, if you don't get in, you'll still be better off having users than not.

I sometimes review YC applications, feel free to hit me up on twitter.

Keep reading

X
Please wait...
Oops! Something went wrong while submitting the form.