I Couldn't Find Startup-focused Youtubers. So I Turned Myself into One

Hello! What's your background, and what are you working on?

My name is Will Kwan and I run the self-titled YouTube channel Will Kwan. I’m a software engineer and indie hacker, and I make videos to share my startups in an entertaining and creative format.

My audience is mainly guys in their early 20’s, but why they watch differs from viewer to viewer. I try to avoid too much coding jargon to focus on showing the results and telling a memorable story, so I suspect a lot of people are watching for inspiration for their own projects. Or maybe they want to learn how to code, find my personality relatable, are sexually attracted to me, etc.

I started making these videos a year ago when I decided to pivot from my channel from Dota Auto Chess videos (250 subscribers!) and talk about something I’m actually good at. Since then, I’ve made 21 videos and grown to 90k subscribers.

What motivated you to get started on YouTube?

I follow way too many YouTubers, including a variety of programmers like Jon Blow, carykh, and Joma Tech, but none of them are focused on startups. So I filled that niche. I was confident that there was an audience for this because these are the kind of people I hang out with in real life. For the first six months, I had fewer than a thousand subscribers, but I kept putting out one or two videos a month. In those early days, Reddit was my main validation. My videos would repeatedly get upvoted to the top of popular subreddits. Reddit traffic is extremely low and short-lived compared to what you can get through YouTube recommendations (more on that later), but it motivated me to keep going.

At the time, I was making about $2k/month in passive income from 7D2DServerHosting, so that became the topic of my first video. I had also raised $70k for another startup with two co-founders. That startup flopped, but I had some funding left to explore different ideas. The plan was to use my channel to document serious, profitable ventures but now I find myself (purposefully or not) coding stuff with the memes in mind. Please don’t judge me.

I never considered any other platform besides YouTube to share my startup content. (I did want to stream Auto Chess on Twitch at one point, but I find livestream coding too boring to watch and hard to follow if you’re popping in and out of a stream like most viewers do.) I’m addicted to video, I find short-form video ala TikTok/Instagram unsatisfying, and I don’t think I’m getting a Netflix show anytime soon.

What was your first video?

My actual first video was a Unity tutorial on how to create the perspective illusion from Monument Valley. I had just quit my job and I had a lot of free time, so I decided to start my channel since I had been thinking about it for a while. A few videos later I quit because it took a lot of time, not many people were watching, and I discovered Indie Hackers (from Hacker News!) so I decided to focus on building web apps.

Two years later, I started making Dota Auto Chess videos because I was addicted to the game. That got boring after a couple months, so I started making tech videos again but in a more informal style.

The first one was How Much Money I Made from My SaaS Web App. It was the obvious choice for the first video topic because it was my main income source. It actually had less than 2k views until six months after uploading when my channel blew up. Now it has 281k views.

I don’t remember the other video topics I considered at the time, but this was the obvious choice for the first video because it was my main income source.

I wrote the script in a day, and I filmed and edited the next day. I borrowed my roommate’s DSLR camera, which I didn’t know how to use so my face was out of focus. I didn’t have a boom or lavalier mic so I used the built-in mic on the camera which was very echoey, and everyone complained about in the comments (which ironically might have given the video a boost for the YouTube algorithm). I edited with a pirated version of Adobe Premiere (I bought it afterwards).

How have you attracted users and grown your channel?

I used to spam my videos everywhere (Reddit, Facebook groups, Slack channels, etc). This was useless in the big picture, though the Reddit upvotes fed my ego, which was actually important. 94% of my views are from the YouTube recommendations (those videos you see on your homepage, or on the sidebar when watching a video). For my most popular video, it’s 99%, and this video contributed to half of my total channel views.

So now I don’t put any effort in self-promotion (not counting random Tweets that nobody reads) and focus on making the best videos I can. YouTube is insanely good at surfacing your content to the right viewers if your metrics are good. Click-through rate of the title/thumbnail and average watch percentage/duration of the video seem to matter the most, but I’m sure Google is tracking and taking into account everything and anything. I’m never happy with any of my videos, I want each one to be better than the last, and YouTube provides the best ecosystem on the Internet to support that mindset. Their neural network is too smart to outsmart, so you actually have to to make good videos.

What's your business model, and how have you grown your revenue?

AdSense and sponsorships. AdSense revenue is unstable because it depends on your views and your CPM. CPM depends heavily on the niche and length of the video (longer videos usually have higher CPM). My CPM is relatively high because techies and entrepreneurs like to buy expensive stuff.

Also, you only get paid if the viewer isn’t using AdBlock, so most of the revenue comes from mobile viewers. It’s unfortunate, but I always watch YouTube with AdBlock so I can’t complain.

Here’s my AdSense revenue since I started monetizing in October 2019 when I reached the 1k subscribers threshold required to enable ads. As you can see, my AdSense revenue is dying and I need another viral video.

Sponsors started reaching out to me through email in October after my video How My Dumb Mobile Game Got 400k Downloads started getting recommended to everybody. Startups usually reach out to me directly, and bigger companies usually work with middleman influencer marketing companies that find relevant content creators and reach out to them. Since I don’t post much, I can get sponsors for all my videos though I sometimes choose not to if it detracts from the video. I started off charging as low as $900 for a 45-90s integration, though now I charge around $1500. I don’t do cold outreach, except for one sponsor (Kite) because I saw them sponsoring another coding YouTuber and I thought their product was awesome and relevant.

I rarely push my own products because I haven’t built anything that would be a good fit for a broad audience like my viewers. My indie hacker instinct of building very niche products conflicts heavily with the marketing potential of my channel right now, but that was my background before YouTube and sharing those experiences is what grew my channel in the first place. Right now, I’m happy just making whatever video I feel like (which is usually some AI project because I learn a lot from building those) and getting paid with ads and sponsors. But I’m not against self-promotion. You can bet I’ll be plugging my startup when I build the next Uber or TikTok.

Do you still like your original videos?

My big mistake when I started my channel two years ago was making coding tutorials. I rarely watch coding tutorials. My order of preference for learning is:

running code → reading code → watching people code

So naturally I didn’t even like the videos I made even as I was making them. But I was really awkward back then and I didn’t know what else I could make.

In the last year since I started taking YouTube seriously, I’ve improved a lot as a filmmaker but in terms of content, I still like the videos from early last year.

Has anything you've made gone viral? What's your "viral recipe?"

My breakout video was How My Dumb Mobile Game Got 400k Downloads. After this video went viral, a bunch of my older videos started getting recommended to people as well.

As you can see, 97.6% of the impressions (the number of times the title/thumbnail was seen) came from the recommendations and this is typical for viral videos, so it doesn’t make sense to spend much time on things like external promotion or search optimization if your goal is to grow your channel.

I’m not surprised that this video was the first to blow up because the game itself was a notable achievement and it’s a more relatable experience than most of my other projects since everybody plays games. But at the same time, I got lucky because I messed up the recording (I filmed in 4k mode on my camera for the first time and the crop factor was too high so it was too zoomed in) and apparently my eyes are creepy so the comments exploded. I’ll take it :3

My recipe for creating viral videos is to always keep the video engaging. Average watch percentage is the most important metric, so every second should make the viewer either think about or feel something, and I’m still far away from being that good. I’ve heard YouTubers talk about making longer videos to boost average watch time as opposed to percentage (freeCodeCamp is a great example) but I haven’t seen any correlation between video length and views on my channel, and I don’t like wasting people’s time.

You have over 90k subscribers and more than 4 million views. Did you see yourself getting to this point?

I’m unjustifiably cocky and I wouldn’t have started the channel if I didn’t think I could get here. But I’m very grateful for it and I want to grow even faster. I know it’s dangerous to care too much about the numbers because fame doesn’t bring long-term happiness or even that much money like in my case. But right now it’s fun for me, so I’m going to keep pushing.

What offline opportunities have come from running your channel?

A lot of people ask me to help or join their startup but I’ve got enough on my plate right now. I’ve bumped into a few viewers randomly in real life before, which is cool since I don’t go outside much. The biggest thing for me is the flexibility to work on whatever I want and not have to worry about making lots of sales right away or finding investors, because as long as I make a good video out of it I can keep bootstrapping.

What’s your logic behind choosing the titles and descriptions you do?

I spend a lot of time thinking about the title. I try to concisely convey the value of the video while keeping it broadly appealing. For example, one of title formats I like to use is doing something in some time frame, such as:

• 7 Years of Coding Startups in 7 Minutes
• Building a Web Startup in 57 Hours
• Coding an AI Startup in One Day

A long time frame implies that the video is packed with information and a short time frame implies that you’re going to see how things get done fast.

To keep the title broadly appealing, I use generic terms like “Web Startup” instead of something like “Esports SaaS Startup”. You don’t know who YouTube is going to recommend your video to, and in my videos I try to explain things in simple and easy to understand terms so I’m selling myself short if I restrict my audience from the get-go with a niche title.

I don’t care about optimizing keywords in the title. I doubt keyword stuffing leads to any significant benefits in this day and age, so I focus on making the titles human-readable with conversational tone and simple vocabulary.

YouTube also truncates the title differently depending on the display location, so I try to keep it short or put the important information in the first 70 characters if it needs to be longer.

I don’t think the description matters much since it doesn’t show up when the video is recommended. However, one thing I’ve been doing recently is putting links to relevant videos. I doubt many viewers will expand the description and watch those other videos, but if they do, I assume YouTube is tracking that my video is helping keep viewers on the site for longer. No idea if this gives my video a boost, but either way, at least I can help promote some other channels.

What’s your process for creating successful thumbnails?

I don’t have a consistent thumbnail style. I’m not convinced that having a consistent style like many YouTubers do helps your existing subscribers recognize your videos when they can already see who published it. Even if it did help, I’m focused on reaching new viewers and I’d rather make the most appropriate thumbnail for each video. However, I do like to keep my thumbnails simple and make sure all text is legible when shrunk down.

I’ve improved my thumbnails a lot in the last few months. I used to just take a frame from the video and use Photoshop/Lightroom to make it prettier and add some text, but I’ve since realized how critical the thumbnail is to grab people’s attention.

Recently, I’ve been doing a lot of AI videos and I think AI generated imagery like deepfakes is excellent clickbait because it’s weird and gets people curious about how it was made. Also, I’m not the best artist so the fact that I can train a model to generate a thumbnail like this saves me a lot of time and/or money.

The original image is Ava from Ex Machina, and I trained a deepfake model on 7000+ pictures of Billie Eilish to faceswap her onto Ava’s head. I did this purely for the thumbnail. (The title of the video is “I Cloned Billie Eilish’s Face and Voice with AI”, but obviously I didn’t actually build a humanoid robot with Billie’s face.) If I were to manually make this in Photoshop, I would have to find the perfect pictures of Ava and Billie with their heads at the same angle and then painstakingly get the skin tone, lighting, and edges to match up. Much easier for me to spend a couple hours gathering/cleaning data and setting up a model in DeepFaceLab, and then just let it train on the cloud for a day and half. The viewer probably thinks that this is just a Photoshopped image, but it doesn’t matter because it looks good.

Most YouTubers aren’t going to be making deepfake thumbnails but the general principles are the same. You have to make your thumbnail stand out while keeping it relevant to the video (a misleading thumbnail will result in low watch-time which will kill your viewership, so don’t do that). There’s no formulaic way to do it.

What does your creative process or workflow look like?

It’s all over the place. Early last year, the videos would be a combination of talking to the camera, acting out random skits, unnecessary b-roll, and screencaptures while I’m working. So I would come up with the general structure of the video in a Google Doc, shoot lots of footage/record lots of screencaptures, and try to edit it together.

Not surprisingly, my three most popular videos were videos that I planned out from start to end. I like doing spontaneous vlogs, but I don’t have the most thrilling or relatable lifestyle so it hasn’t worked well for me. So now I focus on two main aspects:

• What I want to build
• What story I want to tell (a Google Doc script + shotlist)

If I can get some interesting spontaneous footage then that’s great, but I don’t rely on it.

I always jump back and forth between the tech and the story. The script/shot list is much easier to change than, for example, the specifications of a model I want to train, so oftentimes, I’ll have an idea for a story but then I’ll change it to make it easier to build. I’m fine with scrapping lots of footage or code because my goal is to make the best videos I can and do interesting things in the process, not to pump out the most videos.

When I have all the footage, I start editing. By this point, I have a clear idea of what I want so it’s pretty straightforward, though I’m a perfectionist and will often spend a couple days making it look exactly how I want. Considering that my videos take weeks from planning to completion, I’m fine with spending a bit of extra time in Premiere though I wish it wasn’t so buggy.

What YouTubers inspire you?

Tech is more than just tutorials, and seeing Devon Crawford, TechLead, and Joma Tech blow up convinced me that I should play to my strengths and carve my own niche in tech startups instead of making gaming videos.

In terms of specific video ideas, my inspiration comes mainly from what I feel like working on or learning about, but I’m sure I’m subconsciously influenced by what I watch. carykh and Code Bullet are my favorite AI YouTubers, and I’m also a big fan of this milk-chugging gamedev named Dani.

Outside of tech, I used to follow a lot of gamers but I stopped since it was wasting too much time. I like YouTubers with very unique niches. Recently, I’ve been watching a lot of HealthyGamerGG, Nathaniel Drew, and Awkward Since Birth.

What are your goals for the future?

In the short-term, my top priority is grow my channel by making the best videos I possibly can and becoming a better filmmaker. My second priority is to use my videos to motivate myself to work on machine learning projects. I feel seriously incompetent when I talking to real machine learning engineers, and that bugs me a lot because it’s the most interesting field in tech in my opinion.

In the long-term, even though I make fun of Silicon Valley stereotypes a lot, I do want to use my channel to kickstart a unicorn startup. The type-A personality in me just can’t let go of the idea of converting all those viewers into product users. Don’t let your memes be dreams!

What are the biggest challenges you've faced and obstacles you've overcome?

As a YouTuber in general, the biggest challenge is making good videos, and there’s no formula for it. Personally, I prioritize having an engrossing story with meaningful shot selection. This means I have to spend a lot of time in Google Docs. A 10 minute video takes me weeks to plan out, but filming and editing usually only takes a few days.

For my videos specifically, the biggest challenge is using open-source machine learning projects. Academic researchers write terrible code. I always get excited about some model (e.g. for AI music generation), start coming up with lots of ideas for the video, and then I bang my head against my keyboard because I can’t install some library or the trained model performs poorly. Often, I’ll start out with an idea for a video, try to build it, and then come up with a different idea based on what I’m able to build with the tools I’m messing around with. Sometimes, it ends up even better than the original idea. For example, in my AI Generates Beautiful Instagram People video, I was originally planning to train a model to identify faces that would get lots of likes on Instagram, but I ended up generating “Instagrammable” faces from scratch.

Have you found anything particularly helpful or advantageous?

Watching movies and Netflix shows! Been watching a lot of sci-fi and documentaries recently, though I don’t think the genre matters that much. High-quality inspiration and reference material is important no matter what you’re working on, but content creators in particular like to copy other creators on the platforms they’re on. I love YouTube, but there’s usually more to learn by analyzing movies and shows that took a lot of time and expertise to make.

What's your advice for indie hackers or YouTubers who are just starting out?

Read Google’s research paper on the YouTube neural network architecture. It’s a few years old and doesn’t have all the specific details, but it’s the best resource I’ve found on the most important topic for new YouTubers.

TLDR if you’re not a programmer or are too lazy to read it: Think carefully about your title and thumbnail and try to keep people watching for a long time.

When I was starting out, I didn’t know about this paper. If I could go back a year and change one thing, I would read this paper. I thought if I promoted myself in a lot of places, people would like and share my videos and that would make me famous. People liked and shared my videos, and it didn’t make me famous. Social media influencers these days are the by-product of neural networks. It’s a beautiful time to be alive.

Where can we go to learn more?

Feel free to DM me @_willkwan on Twitter or Instagram, though all my good content is on YouTube.

Do you want to be a YouTuber? Pitch your channel idea in one sentence and I’ll give you my thoughts in the comments!