TL;DR

  • A 31-year-old in Columbus, Ohio built a faceless YouTube channel using Claude AI for scripts, an $11 voice tool, and stock footage
  • He published one video per day for 30 days and earned $37,380
  • The money came from affiliate links and a digital product, not AdSense
  • Total tool cost: $43 per month. Total time: about 25 hours across the month
  • This case study is based on Mark Poiler’s X article (@mpoilerfx), which got 722,500 views and 902 bookmarks

Thirty-seven thousand dollars in 30 days. No camera. No crew. No face on screen. Just a laptop, a few cheap tools, and a system that a 31-year-old in Columbus, Ohio built one Tuesday night while his coffee was getting cold.

The story came from Mark Poiler (@mpoilerfx) on X, and it went viral for a reason. Not because it promised overnight riches, but because it showed the actual steps. No fluff. No “just be consistent.” Ten concrete moves that turned an empty YouTube channel into a revenue machine in 30 days.

This post breaks down exactly how it worked, what each step looks like in practice, and what you can take from it to build your own faceless YouTube channel with AI.

The Story Behind the $37,380

The guy behind this channel was not a marketing guru. He was not a tech bro with a trust fund. He had a script-reading job that he could do half asleep, and he wanted a side project that did not require him to be on camera.

One Tuesday night, he opened Claude, typed in a topic, and watched it write an 8-minute video script before his coffee went cold. That moment changed his month. By the end of 30 days, he had 30 videos live, a folder called “youtube-machine” on his desktop, and a payout screenshot he kept reopening to check it was real.

The whole operation ran from his laptop. No camera purchases. No lighting equipment. No filming schedule. Just a repeatable process that he ran every single day for a month.

What makes this story worth studying is not the dollar amount. It is the system. Every step is something you can replicate, and most of it costs almost nothing to start.

How It Works: The 10-Step Faceless YouTube System

This is where the case study gets practical. Each of these 10 steps is a piece of a machine. Skip one, and the whole thing breaks. Follow all of them, and you have a channel that produces revenue on a predictable schedule.

Step 1: Pick a Niche YouTube Already Pays to Fill

Not all YouTube niches pay the same. RPM (revenue per mille, or what advertisers pay per 1,000 views) sets the ceiling on how much you can earn. Cooking vlogs pay $2 to $4 per 1,000 views. Software tutorials, personal finance, and B2B how-to content pay $15 to $40.

That difference is massive. A video that gets 100,000 views in a cooking niche earns $200 to $400 from ads. The same view count in a software tutorial niche earns $1,500 to $4,000. Same effort, completely different payout.

He spent one evening searching “how to” queries inside those high-RPM topics. He logged every query that had under 50,000 views and a weak thumbnail into a file called rpm-gaps.csv. He found 40 gaps. Each one was a video that nobody had made well.

This research step is the foundation. If you pick a low-RPM niche, you are capping your earnings before you even start. If you pick topics that already have hundreds of great videos, you are fighting an uphill battle. The sweet spot is high-RPM topics with underserved queries.

If you are thinking about starting a digital project from scratch, our guide on starting digital marketing as a beginner in 2026 covers how to pick a profitable niche from day one.

Step 2: Turn Claude Into a Script Engine

This is where AI does the heavy lifting. He wrote one prompt and saved it as script-engine.md. The prompt takes a topic and returns a 1,200-word script with specific structural requirements: a hook in the first 15 seconds, a payoff promise up front, and a retention beat every 30 seconds.

Input a gap from the csv file, get a finished script in about 8 minutes. Over the course of 30 days, he produced 30 scripts. Total time spent on scripting: 4 hours across the entire month.

The key here is that he did not just ask Claude to “write a YouTube script about X.” He built a structured prompt that enforced the format YouTube rewards. The hook grabs attention in the first 15 seconds, which is where most viewers decide to stay or leave. The payoff promise tells them why they should keep watching. The retention beats every 30 seconds keep the pacing tight enough that viewers do not click away.

Claude is particularly good at this kind of structured output. You give it a framework, and it fills in the content while respecting the constraints. If you have used Claude for other content tasks, you know the quality difference between a generic prompt and a well-engineered one.

For a deeper look at how Claude can power an entire content operation, check out our case study on using Claude to build an Instagram money machine. The same principles apply to YouTube.

Step 3: Give the Script a Voice

He pasted each script into a text-to-speech tool, picked one calm voice, and rendered it as an audio file. Cost: $11 per month. The voice never cracks. It never asks for a day off. It records at midnight or noon without complaint.

Text-to-speech has gotten remarkably good in the last two years. The best tools now produce voices that are nearly indistinguishable from human narration, especially for informational content where the tone is calm and conversational. Viewers watching tutorials and how-to videos care about the information, not whether the voice has perfect emotional range.

The trick is picking one voice and sticking with it. Consistency builds familiarity. If your channel sounds the same in every video, viewers start to recognize it. That recognition is part of your brand, even without a face.

Steps 4 and 5: Build the Visuals and Assemble in One Pass

For tutorial videos, he screen-recorded the software itself. For everything else, he pulled free stock clips and a few AI images into a b-roll folder. No lighting. No set. No filming of any kind.

An 8-minute video needs about 25 clips. Once he built his reusable b-roll library, later videos cost zero new footage. The same clips get used across multiple videos, and because the voiceover is different each time, viewers do not notice.

This is the part that trips most people up. They think a YouTube video needs original footage. It does not. Faceless channels have been running on stock footage, screen recordings, and simple graphics for years. The content is in the script and the voice. The visuals are just something to look at while the information lands.

Step 5: Assemble in One Pass

He dropped the voice track into a free video editor, laid the b-roll over it, and ran one saved channel template for the intro, captions, and end screen. The first edit took 2 hours. The thirtieth edit took 18 minutes.

That speed improvement is the whole point of a system. You build the template once, and then every video after that is just a matter of dragging in new files and hitting render. The captions, the intro style, the end screen layout, the font choices. All of it is locked in from the start.

If you are spending more than 30 minutes editing a faceless YouTube video after your first week, your template needs work. The goal is assembly, not creation. You already did the creative work in the scripting step. Editing is just putting the pieces together.

Our post on the content machine system you can run in one hour covers how to build repeatable content workflows that cut production time dramatically.

Thumbnails, Titles, Cadence, and Revenue

The content is only half the game. If nobody clicks, nobody watches. And if you publish inconsistently, the algorithm never learns what your channel is about.

Engineer the thumbnail and title. He asked Claude for 10 title options per video and picked the sharpest curiosity gap. Thumbnails went together in a free design tool: big text, one face-free focal image, and 3 versions of each. He swapped the loser after 48 hours if click-through stayed under 4 percent.

This is where most creators phone it in. They write one title, make one thumbnail, and move on. The problem is that a great video with a bad title and thumbnail gets zero views. The title and thumbnail are the ad. The video is the product.

Three versions of each thumbnail is non-negotiable. YouTube lets you A/B test thumbnails now, and even without that feature, you can manually swap after 48 hours by looking at your click-through rate. If CTR is under 4 percent, the thumbnail is not working. Replace it.

Claude generating 10 title options is a smart move because titles are hard. You need to balance curiosity, clarity, and searchability. Having 10 options lets you pick the one that creates the strongest curiosity gap without crossing into clickbait territory.

Ship on a fixed cadence. One video a day for 30 days. He scheduled each video to publish at 9 a.m. Eastern. He fed the algorithm one signal a day, and it learned the channel faster than a single viral hit would have.

This is the unsexy part that actually matters. Publishing daily tells YouTube that your channel is active, that it has a consistent upload pattern, and that it should be recommended to more people. The algorithm rewards consistency above almost everything else, especially for new channels.

Scheduling each video at the same time every day also trains your audience. Even if they do not consciously know you publish at 9 a.m., their viewing habits start to align with your upload schedule.

If you are managing content across multiple platforms, tools like InstantDM let you schedule and automate publishing so you are not manually uploading every single day. Setting up a week of content in one sitting is far easier than logging in every morning to hit publish.

Stack the revenue behind the video. This is the part of the story that most people get wrong. They hear “$37,380 in 30 days” and assume it all came from YouTube AdSense. It did not. In fact, AdSense was the smallest piece of the puzzle.

AdSense pays slow. A brand-new channel sits below the YouTube Partner Program bar for weeks. You need 1,000 subscribers and 4,000 watch hours before you even qualify. So day-30 cash never comes from ad revenue.

The money came from the description. Every video carried two affiliate links to the exact tools it demonstrated, plus one link to his own $39 setup guide. The pinned comment under each video repeated the offer.

Viewers who searched a problem, found the video, and got the fix bought the shortcut. That is the entire sales mechanism. No complex funnel. No email sequence. Just a helpful video that shows people how to solve a problem, and a link in the description to a product that helps them do it faster.

The $39 setup guide was his own digital product. It took him a few hours to create, and every sale was pure profit. The affiliate links paid commissions on tools he was already demonstrating in the videos. Both revenue streams were embedded directly into the content, not bolted on as an afterthought.

This is the same approach we covered in our AI tools business autopilot guide. The tools generate the content, the content drives the traffic, and the revenue comes from products and affiliate offers layered underneath.

The Realistic 30-Day Ramp and the Math

Week 1 paid nothing. The videos were indexing. YouTube was figuring out what the channel was about and who to show it to. No views, no clicks, no revenue.

Week 2 brought the first affiliate clicks and 11 sales of the setup guide. That is $429 from the guide alone, plus a small amount from affiliate commissions. Not life-changing, but proof that the system worked.

Week 3 is where the spike happened. One video caught a 90,000-view wave, and it pulled the rest of the channel up with it. When one video blows up, YouTube starts recommending your other videos to the same audience. That is the compounding effect of a library of content on the same topic.

Week 4 stacked affiliate commissions, guide sales, and a thin AdSense trickle into the $37,380 total.

Here is the critical lesson: most months will not look like week 3. The spike was real, but it was not the baseline. Plan for the slope, not the spike. If your system works at the week 2 level, the occasional spike is a bonus, not a requirement.

The math. Tools ran $43 for the month. That covers the voice tool ($11), a design upgrade, and a stock footage subscription. Time ran about 25 hours total across the entire month. That is less than an hour per day.

The revenue split was roughly 55 percent from the setup guide, 35 percent from affiliate commissions, and 10 percent from AdSense. One laptop. One folder called youtube-machine. That is the entire operation.

If you want to build a personal brand alongside a faceless channel, the same research and scripting systems apply. But the faceless route is faster to start because you skip the filming, lighting, and on-camera confidence curve entirely.

What You Can Learn From This Case Study

The numbers are impressive, but the lessons are more valuable than the dollar amount. Here is what stands out when you look at this system as a whole.

Revenue does not come from views. It comes from having something to sell. This channel earned more from a $39 digital product and affiliate links than it did from AdSense. Views are the traffic source. The product is the engine. If you build a faceless channel with no offer in the description, you are leaving most of the money on the table.

Consistency beats virality. One video a day for 30 days taught the algorithm what the channel was about. A single viral hit might get views, but a consistent library of content builds a compounding recommendation engine. The algorithm learned this channel faster because it had a new signal every single day.

AI is the production tool, not the strategy. Claude wrote the scripts, but a human picked the niche, engineered the prompts, and built the revenue model. The AI saved time on production. The strategy was entirely human. If you want to see how this plays out on other platforms, our viral content strategy guide breaks down the same principle for Instagram and TikTok.

Speed compounds. The first edit took 2 hours. The thirtieth took 18 minutes. The first script took 8 minutes, and by the end of the month he had a library of prompts and templates that made each new video faster than the last. Building a system means every unit of effort gets more efficient over time.

The niche matters more than the tool. A faceless cooking channel with the same system would have earned a fraction of the revenue because cooking RPM is $2 to $4. Software tutorials and personal finance pay $15 to $40. The tool does not fix a low-paying niche. Pick the niche first, then build the system.

Automation removes friction, and friction kills momentum. When every step requires manual work, you start skipping steps. When the system is templated, each video is just filling in the blanks. Scheduling tools like InstantDM fit into this same philosophy for social media. The less manual work between you and a published piece of content, the more content you actually ship.

The folder is the framework. He called it youtube-machine. Everything lived in one place. Scripts, voice files, b-roll clips, templates, thumbnails, and the csv file with his topic gaps. When the whole operation fits in one folder, there is no friction to starting a new video. You open the folder, pick a topic from the list, and the next step is already waiting. That physical organization mirrors the mental model of a system that runs the same way every time.

For creators who are also thinking about their broader content strategy across platforms, our guide to Instagram content types in 2026 covers which formats are working now and how to repurpose content across channels.

Building Your Own Faceless YouTube Channel

If this case study has you thinking about starting your own faceless channel, here is the honest version of what you are signing up for.

The first week will feel slow. You will publish videos that get 47 views and wonder if the system is broken. It is not. YouTube needs time to index your content, figure out your topic, and start testing your videos with small audiences. The system is designed to compound, not to spike on day one.

The scripting step is where Claude saves you the most time. Writing 30 scripts from scratch would take 60 to 90 hours. With a well-engineered prompt, it takes 4 hours. That time savings is the difference between a side project that burns you out and one you can sustain for months.

The monetization step is where most people leave money on the table. They build the channel, get some views, and wait for AdSense checks that take months to arrive. Meanwhile, every video is an opportunity to link to a relevant product, tool, or digital download. Stack those opportunities from day one.

If you want to automate the publishing side, InstantDM handles multi-platform scheduling so you can batch your uploads and let the tool handle the timing. When you are publishing daily, anything that removes a manual step is worth the investment.

Conclusion

The 31-year-old in Columbus did not have a secret. He had a system. Ten steps, repeated daily, with tools that cost less than a streaming subscription. Claude wrote the scripts. A text-to-speech tool gave them a voice. Stock footage and screen recordings provided the visuals. A simple digital product and affiliate links turned views into revenue.

The faceless YouTube model works because it removes every barrier that stops people from starting. You do not need a camera. You do not need to be comfortable on screen. You do not need to spend money on equipment. You need a niche with high RPM, a script prompt that works, and a product to sell in the description.

Credit goes to Mark Poiler (@mpoilerfx) for originally sharing this case study on X, where it earned 722,500 views and 902 bookmarks. The original post is worth reading for its direct, no-nonsense breakdown of the 10 steps.

The tools are cheap. The system is simple. The only variable is whether you actually ship 30 videos in 30 days. Everything else is already solved.