How we got to 10,000 stars (+ why you should be launching as often as you can)

How we got to 10,000 stars (+ why you should be launching as often as you can)

Why does YC keep talking about "launch early, launch often"?

By default.. people don't know that you exist. It's your job to be the squeaky wheel, and let people know "HEY I EXIST", "HEY I CAN SOLVE YOUR PROBLEM"

You may have seen my many posts talking about Skyvern hitting star count numbers

Those didn't happen out of the blue. We kept launching, and launching, and launching. Trying to spread the word about us. Trying to learn about which use-cases we're best suited for. Trying to solve anyone's problem!

Not all of the launches were effective.. but if you're curious about all the launches we've done, get ready.

1. (Pre-idea) November: I cold emailed ~100 different founders that had talked about web-automation in random online forums -> Led to our first customers + 3 pilots (and tons of lessons)

2. (First launch) January: Launched with a barely functional API and a landing page to the YC Community -> Led to our second customer and a very important disqualification criteria: we don't automate tasks on LinkedIn

3. (Second launch) February: Did our first public launch via Launch YC with a slightly more featured API + landing page -> Led to a lot of prospects we couldn't serve (healthtech companies) -> Gave us conviction that we should open source

4. (Third launch) March: Open sourced + did a ShowHN (anyone can do a ShowHN) -> This is when we first felt market pull. Our product was still just an API (no UI) with a demo that disqualified anyone who didn't want to work with startups. This went viral – we got to 3K Github stars + we got a handful of new customers

5. (Fourth launch) June: We shipped a MVP of Skyvern Cloud (with a UI) and launched on Product hunt -> We got a handful of inbound leads but ended up #10 on the day and didn't generate much interest

6. (Fifth launch) October: Launched Skyvern Cloud on Hackernews -> went viral again and went from 6K -> 9K github stars in 4 days. Made it to #1 on Github trending daily + weekly.

You'll notice that as our product got more mature, we started launching less frequently. This is because the gaps in our product became more and more clear as we talked to more users. This meant that we had market fit, but not product market fit. We still needed to iterate on the product

This was obvious before the launches though.. so why launch early? Well.. more often than not, our customers ended up caring about features we thought were unimportant (ie cancelling tasks), or some features (instant live-streaming) ended up being way more important than others (debugging workflows).

How would we know precisely what to build without this feedback?