How I Built A Startup on Twitch in 24 Hours

Pat Walls

•

Published: September 27th, 2018

I built and launched a startup in 24 hours and streamed the whole process on Twitch to thousands of viewers.

Some numbers from the launch:

#1 on Product Hunt
#1 on Show HN (Hacker News)
15k uniques
9,000 Twitch views
1,000 registered users
18 hours of continuous streaming
12 La Croix’s

I started the day with ZERO lines of code, and ended the day with a submission to Product Hunt.

During the live stream, I decided on an idea, bought a domain, planned features, wrote code, got feedback, tested, and ultimately launched it - all in the same day.

After it all went down, there was a ton of buzz on Twitter about it, so I wanted to write an in-depth post about the whole process. I also wanted to write about a couple of things I believe made the stream so successful.

UPDATE: I'm doing it again! I'm creating another startup on Twitch! See bottom of post for more details.

What sparked this crazy idea

First of all, I just want to say that I had NO IDEA this was going to blow up - I just wanted to build something.

I came up with the idea to build a startup in 24 hours while I was on a run last Tuesday. Sidebar - a couple months ago, I set a goal to go for a run every day for 100 days.

embed:tweet

Running is a great way to come up with and challenge your own ideas.

During my runs I think a lot. I come up with little ideas on how to improve my life and my projects.

Why build a startup in 24 hours?

I have a full time job and I run StarterStory.com on the side. My time is super limited. But I have a lot of ideas, and no time to implement them.

Even though I’m busy, I get bored sometimes. I want to BUILD STUFF.

This idea I had - to create a blog “powered by Google docs” had been in the back of my mind for a while, and it kept coming back every so often. I was getting antsy to implement it.

But how could I justify spending a few weeks building it? I couldn’t.

The only way I could justify it was getting it done AS QUICKLY AS POSSIBLE. I realized I had a free weekend and I was already ahead of my work on StarterStory.com (shoutout to @WalterDom_).

So I thought.. how can I actually build this as quickly as possible? Also, I need to hold myself accountable! How can I hold myself accountable?

By streaming the whole thing!

Inspiration and influence

Obviously, the idea to stream is not new.

I had heard before of people building startups while streaming. For example, Pieter Level’s talks about it on his book Make Book. And also Andrey Azimov streamed his launch for Make OS X Great Again.

I was also recently inspired by an initiative from Matt Henderson and a few friends, who are launching 24 startups in 12 months.

I think it's really badass to have the confidence to work on tons of different projects, fail quickly, and be able to worry less about the outcome (and the ego).

My impulsive tweet

embed:tweet

So I’ll be honest, I wrote this tweet on a complete impulse. I wanted to get this project done and this idea out of my mind.

The only way I could commit to finishing this project was by tweeting that I would do it.

I’m not sure if it’s a good thing, but I kind of have a knack for saying I’m going to do something before I actually do it. I think it’s my way of committing.

Then, the tweet totally blew up.

Pieter Levels liked it, then it hit Ryan Hoover’s feed.

Crap, now I actually have to do this..... 😲😁
— Pat Walls (@thepatwalls) September 21, 2018

Once Ryan Hoover commented, the tweet went semi-viral.

My reply to Ryan Hoover was actually how I felt at the time.

Before the tweet went viral, I was thinking I would probably just have a couple of people and I didn’t have to really commit to this. But, that was clearly not that case anymore.

I didn’t even have a Twitch account. I had no idea if my computer could handle streaming. I had (and still don’t have) any external microphone and camera setup - just my laptop webcam and internal mic.

Commit to something and let the world know! Do crazy stuff and figure it out later.

The night before / the morning of

Now that this tweet was going viral, I was getting excited but also nervous. I was thinking - What if my computer died tomorrow? I guess I would have to go buy a new one. I can’t back out now!

I realized I had to switch to Twitch because YouTube has a 24 hour holding period for new accounts. This was actually a blessing in disguise which I’ll get into later.

This is Friday night. I went to a bar by myself and had a couple beers to calm the nerves. Then I went to the grocery store to stock up on La Croix’s and food.

The night before:

how-i-built-a-startup-on-twitch-in-24-hours

My friend Leeron also texted me after seeing the tweet:

I was also nervous about my workstation setup.

Right now, I’m in between apartments and I don’t even have a desk.

I did this whole thing from an extremely uncomfortable futon with a makeshift desk-chair.

The worst workstation of all time.

The stream

I woke up the next morning, got a matcha latte and a bagel, and got mentally prepared.

I started the stream and instantly 20-30 people joined.

I think people were waiting for me? This is crazy. I heard some people even woke up early in their time zone to catch the beginning. This. Is. Nuts.

Spreading the word

The most important thing I did, other than my original tweet, was to immediately share the stream across the internet.

And I actually did this all on stream, too.

I explained to the folks watching the importance of sharing what you’re working on and building your audience. I literally typed up tweets and reddit posts on stream.

I replied to my original tweet, made a new tweet, posted in a few subreddits, posted on WIP.chat, and I even posted on my own personal Facebook.

The idea

The idea for the product was a blogging platform powered by Google Docs.

As I mentioned earlier, I had an idea of what I wanted to build, but I was still hazy on how I planned to implement.

A little background on the idea:

This idea came from the fact that I use Google Docs religiously for StarterStory.com. It’s excellent for feedback, commenting, revisions, sharing, etc.

Since all my interviewees write out their interviews on Google Docs, I use the Google Drive API to extract that text, convert it to markdown, and make it a blog post on Starter Story.

But I thought maybe I could take it one step further... and use Google Docs as the CMS?

I have also struggled with learning WordPress. I tried using it as the backend for Starter Story, but I found it painful to learn how to customize.

A few months ago, I also saw someone do something similar with Dropbox.

Coming up with a name

As I explained on stream, I wanted to have a name that was catchy and controversial.

I knew that my product was going to be limited in features and not very polished, so I needed something to catch people’s eye.

This is a play right out of Pieter Levels or Andrey Azimov’s style, for example Pieter’s Go Fucking Do It uses a cuss word in the name. Or Andrey’s Make OS X Great Again is a pun on Donald Trump’s slogan.

I chose the name You Don’t Need WordPress because I knew that people would have mixed opinions on it. I anticipated WordPress advocates getting triggered by the name, but others getting excited about the utter simplicity of it.

It was at this point that I had around 100 concurrent viewers.

Holy shit… I’m not fully aware or thinking about how many people that is, but it seems kind of crazy.

But I was just very stressed about getting it done. I knew I had a lot of work ahead so I wasn’t really thinking too much about the amount of people watching.

I also started getting more “used to” streaming, being more of myself and forgetting that so many people were watching.

It’s funny, because I would be a nervous wreck if I had to speak in front of 100 people in person…

A couple of streaming tips:

Click Popout on the chat to have a permanent chat window in sight at all times.
Put links below your stream on what you’re building and other common questions you will get. When people ask in the chat, you can tell them to look below.

A lot of people will be joining and asking the same questions. You don’t need to answer everything. Over time, my loyal streamers would help field many of the newcomer questions in the chat.

Planning features

I wanted to plan features at the beginning so I could take viewers through that process.

We had a working google doc (which I linked to below the video).

We started out by planning out these features:

As you can see, we planned out the features at a very high level. These features were simple, mostly CRUD stuff so they didn’t need much thought on design.

Know what technology you’re going to use but don’t think about technical implementation.

Landing page

For me, this part was easy. I pay for premium Launchaco - a landing page builder - so I was able to mock up a landing page with ease.

Then, I just exported the HTML and put it in my app.

I tried to spend as little time on this as possible, as your idea isn’t even built yet - I just describe the main points and got a general layout.

I figured I could polish it later. I wanted to save the final copy and assets once I was close to launching - plus, I wanted to get right into the coding.

Here’s what my original homepage looked like:

As you can see the copy and images changed with my final product, but layout and color scheme are the same.

Coding

This was by far the longest, most tiring, and hardest part.

The ups and downs

I was confident at times:

And I was slow, pathetic, and embarrassing at times, and I forgot basic shit like how media queries work in rails and CSS !important tags.

Deploy in the beginning

I do this with all my projects regardless if I stream them - but I make sure to get my production environment set up and deployed before I get started.

This also helps because you can deploy your landing page before you even start coding and viewers can more easily understand what you’re building.

People will help you!

This was was a pleasant surprise.

When I was struggling, people on the stream helped me so much. Especially with CSS.

I also struggled a bunch with the google API and hooking up oAuth. I almost gave up like 10 times. I would not have been able to implement it without a few of you guys in the chat.

When @alecofjones saved my ass:

There were so many people that I can’t name everyone, but special thanks to @alecofjones @chrisevans1001 @zachinatx @cahouser @grantkiely @stephenfjohnson Tim Hwang and so many others. Please let me know if I forgot you! I will add you.

People want to hear you THINK!

I think talking while you code is super important.

I tried my best to explain all of my decisions. Showing people your thought process is why they are watching. Also, most people won’t actually understand your code in the context of a stream, so talk!

You are sharing viewers your mental models on how to get things done, how to research and solve problems and how to communicate your goals and progress. This is what will keep them here.

Explain everything. Talk about your other projects and businesses, tell people who you are.

I think some of the most important skills in the world are the ability to teach, and to tell a story.

Instead of telling people how to do something, show them, teach them, explain different points of view, and let them have input and ownership.

Pick a language you know

If you’re going to do something like building a startup in one day, you should be very confident with your framework/language.

For me, that’s Ruby on Rails, and although I’m not a great engineer, I know the ins and outs of the framework from working with it for many years in my full time job and side project.

Commit your code

Just commit your code. I was so frantic that I barely did it.

I actually lost some important code for being stupid and had to rewrite it.

Take breaks

Streaming for 18 hours was mentally exhausting. Like nothing I’ve done before.

There were many times where the lines of code on the screen became blurred. That’s when I needed to take a break.

I took a lot of breaks.

During this time, I answered tons of questions from the chatters, checked Twitter, Reddit, email, etc. I did this all in front of the audience. I told stories about myself and talked about my other businesses. Be an open startup!

Taking breaks and shooting the shit with the viewers was the best part.

I also made sure to continously tweet as I "checked things off" the list.

embed:tweet

Feedback and test

This was one of those moments where I remember being awestruck.

Once I deployed the code for the first version of the MVP, I had like 10 people actually testing it immediately.

They were able to find most of the bugs within 15 minutes. Shoutout to everyone who made accounts and tested.

Preparing to Launch

At this point I’m at like hour 15 or 16, or like 1:30AM EST.

I know it will be perfect timing to launch on Product Hunt, as their submissions open up at 3AM EST. So just one more hour and then I can sleep.

I was close but still needed to re-do the creative and copy on the landing page and make a logo and images for Product Hunt and all that.

It’s getting late, so many people are starting to drop off. But they all wish me good luck and that they will check and upvote in the morning.

Writing at 2AM sucks!

This was one of the worst parts.

Writing the Product Hunt description and intro comment (you can see how tired I am at this point):

I was so brain dead.

Finally, we launched.

Once I hit submit on Product Hunt, I checked back a few minutes later and it had nearly 30 upvotes. Nuts!

There’s something to say about having this audience rooting for you.

It had been 17 hours of continuous streaming at that point.

I shared on Twitter that it was done and linked to Product Hunt. I was so tired I barely remembered writing this tweet.

embed:tweet

Little did I know, that tweet would go viral.

Before I started the stream, I also created a quick newsletter. I had a link for that below my video. Anytime people let me know they were going to leave the chat, I let them know I would email them when the product was launched.

Sleep! & the next day

To be honest, I had a few bad dreams: like thinking that I got so many early upvotes that I tricked Product Hunt’s algorithm into thinking I was cheating it, and as a result didn’t even make the front page.

Luckily, that wasn’t the case!

I woke up at 11AM on Sunday to #1 on Product Hunt and hundreds of notifications, messages, and emails.

My Twitter was also going bonkers.

It honestly felt surreal. I spent the whole day on Twitter and email trying to respond to everything.

I inspired people?

I still don’t understand this…

I received loads of tweets and messages about this live stream inspiring people, including from people that I really admire:

This was the best part for me.

The product is cool, but inspiring tons of people to build their own products is a much bigger thing.

And if I can do that by just showing the world the process isn’t as crazy or hard as it sounds, then that’s fucking amazing!

I also inspired a website to be created?

Check out it out at shipstreams.com.

But of course, I got some haters too:

I’ve learned from posting on reddit that having haters is a good thing.

Having haters means that you are polarizing and being yourself. It also means people are taking notice of what you’re doing.

My takeaways

Commit now, figure it out later. You’ll always find a way.
Doing crazy stuff has big rewards.
Streaming is not easy!
Have a stated goal for your stream - something your viewers can stay excited about.
Tell the world about your stream.
Play some music.
Twitch is probably better than YouTube since they promote your stream to other Twitch users.
Talk and think out loud while streaming.
Get anyone to watch - share on your personal Facebook and with friends.
Engage the audience.
Your viewers are your lifeline when it comes to bugs and getting through tough times.
Take breaks and just talk to the audience.
Push live and commit early and often.
Be open and show everything.

I don’t think that streaming your afternoon of coding is particularly valuable. There’s plenty of videos on YouTube if people want to learn to code.

It’s important to have a well planned out stream and do something that will get people excited.

I’m doing it again!

Yep, you heard that right. For those that missed it or want to see it again, I will be launching another startup in 24 hours this weekend! Two weekends in a row!

Come hang out, code alongside me, and watch it happen again. This time, it will be a SaaS app, and the idea was actually inspired by something that happened after my stream!

The stream will start this Saturday, 9/29 at 10 AM EST. It will be streamed here.

Follow me on Twitch and Twitter. If you want to be notified via email right before I go live, subscribe here. I will never spam you :)

Special thanks to @ramoncacho, @WalterDom_, @StephenFJohnson, and @jjsee for helping me with this blog post.

Add Text Section Add Story Add Idea Add Business Add Tool