---
title: 'How to Get and Test Startup Ideas - Michael Seibel'
source: 'https://youtube.com/watch?v=vDXkpJw16os'
video_id: 'vDXkpJw16os'
date: 2026-08-16
duration_sec: 381
channel: 'Y Combinator'
---

# How to Get and Test Startup Ideas - Michael Seibel

> Source: [How to Get and Test Startup Ideas - Michael Seibel](https://youtube.com/watch?v=vDXkpJw16os)

## Summary

Michael Seibel, a co-founder of Justin.tv (which became Twitch and was sold to Amazon for nearly $1 billion), debunks the myth that a startup must begin with a great idea. Instead, he advises starting with a problem you personally care about, validating your unique qualifications, building a Minimum Viable Product (MVP) quickly, and carefully selecting initial users to test whether your solution truly solves the problem.

### Key Points

- **Destroying the 'Great Idea' Myth** [00:02] — Seibel argues that a startup doesn't need a great initial idea. He cites his own experience with Justin.tv, which started as an online reality TV show but evolved into Twitch, a live video platform for gamers, showing that ideas can pivot dramatically.
- **Start with a Problem, Not an Idea** [00:44] — Starting with an idea leads to grading it, but starting with a problem is easier to evaluate. Ask yourself: Do I have a personal connection to this problem? Is it a problem I, my friends, my family, or my community face? A personal connection helps you gauge if a solution is viable and keeps you motivated during tough times.
- **Keep a Problem Book and Brainstorm with Friends** [01:49] — Instead of idea books, keep 'problem books' to jot down complaints from friends and family. Brainstorming with friends about problems is also a great way to test co-founder chemistry and see if you can build ideas together.
- **Why Are You Uniquely Qualified?** [02:31] — Once you find a problem, ask why you are uniquely qualified to solve it. This isn't about a resume; it's about having a unique angle, perspective, or personal experience that others don't. Look at past attempts to solve the problem to see if your insight is different.
- **Build and Release Your MVP Quickly** [03:39] — Your MVP (Minimum Viable Product) is the first solution to test with users. It will likely be poor, but that's okay—it's the first step in learning if the problem is solvable. Fall in love with the problem, not the product, and be ready to iterate. Release it in days or weeks, even if it's not perfect.
- **Hand-Pick Your Initial Users** [04:52] — Your initial users should be carefully selected. They should be willing to try a startup and have a real, important problem you're solving. The goal isn't to see how many people want your product, but whether it actually solves the problem for any group. Best startups heavily filter their early users.

### Conclusion

The key to starting a startup is not a brilliant idea but a deep connection to a real problem, a unique angle, and a willingness to iterate quickly with a focused group of early users. This approach turns vague ideas into testable hypotheses and sets the foundation for a successful venture.

## Transcript

idea has to be great in order start a company and the first thing I want to do is destroy that misconception personally I was one of the cofounders of a company called justin.tv it later became a company
called twitch and sold to Amazon for almost a billion dollars our original idea was to create an online reality TV show clearly it's hard to draw a line
between that and creating a live video site where people watch video gamers and chat with their friends but that's exactly what it became and so don't fall
initial idea has to be great by any definition this is what you should think about instead my first advice is to start with a problem I think the
starting with ideas is tricky because immediately people want to grade your idea it's a lot easier to start with a problem and to think about how do you grade a problem when you think about a problem you should be asking yourself do
I have any relationship with this problem personally is this a problem that I have is a problem that my friends have it's a problem - ma'am family has is it a problem that exists through the work is it a problem that's in my
community oftentimes having a personal connection to the problem is really connection to the problem is really helpful for two reasons one you can tell whether this solution is even within the ballpark of solving the problem because
problem and - when you're feeling discouraged and the solutions that you're building are working you still problem that sees you through until you find a
solution that does work so that's the first place I would start thinking about keep little idea books with start-up success if they kept little problem books and wrote down notes on the
basis or that their friends complain about their family complain about you're starting to think about problems is to brainstorm with friends this is the perfect time to start thinking how do you recruit co-founders and a lot
having really good brainstorming sessions sessions we're like you talk about a problem they elaborate on it and you kind of realize that you trade ideas well and that you can talk to each other well you can build ideas together and so
I really advise when you're doing these thinking about problem do it with some friends friends that you might want to create a start-up with one day once you find a problem that's interesting enough to start pursuing further the next
question I ask myself is why are you uniquely qualified to work on this problem now if I qualified I don't mean in the classical resume sense like oh you've had 25 years experience in this area or something like that but what I
do mean is that like is there some angle on the problem that you understand that you don't think others understand is there some way of attacking the problem is there's some perspective on the problem is some personal experience for
the problem that you have you don't feel as generally used or understood in other solutions to the problem in this face it's often helpful if you're solving a common problem to look at some of the people who tried to solve the problem
that are trying to solve the problem or have failed to solve the problem just so you can get a basic understanding of what other people thought their unique insight was and whether it was the same or different than yours this is kind of
your gut check hey is this a right fit for me the next step is to start thinking about how do you want to build your MVP your Minimum Viable Product first solution you're going to build and release to your users to see if maybe
you can try to help them solve this problem the most important thing to love with it your MVP is most likely not going to the problem very poorly and really it's the first step in the learning process
of whether the saw problem is solvable at all let alone whether you can solve at all let alone whether you can solve it a lot of people fall in love with their product and are not in love with their problem or their customer I advise
problem be in love with the cause tomorrow and treat your product in a way that it can change it can develop it can improve don't be too in love with your initial product the most important thing to not fall too in love with initial
product is to release it quickly oftentimes an MVP can be built and released in days or weeks it's not going to be perfect it might not be something you're proud of but it can be the beginning of learning about whether or
beginning of learning about whether or not you can actually solve the problem the last thing I would consider once you have an MVP is be very careful here initial users are I actually believe that your initial users should almost be
hand-picked but certainly you should have a strong opinion about who your initial customer is and you shouldn't try to solve this problem for someone who one isn't willing to try a start-up or two the problem you're trying to
solve is not a really important problem in their life so your goal here with your MVP is not to see how many people want to use your product is to see whether for if for any group of people does this actually solve the problem for
does this actually solve the problem that you're trying to solve and so the best startups actually very heavily filter the people who are able to use filter the people who are able to use the initial product and make sure that
they're the right type of initial customer so those are the steps that I would take from thinking about of problems of getting feedback on it that's how I would think about this problem thank you