The 600

The problem

I wanted to find out whether a content niche was worth a multi-year commitment, without making that commitment first. So I ran the niche for a short, fixed window and decided from the numbers whether to carry on.

The constraints: a solo operation, one external editor on long-form post-production only, 10 to 15 hours a week of my time, no producer and no team. The plan was ten videos across four content categories over eight to ten weeks.

What I built

A ten-document operating manual I wrote and kept updating: a master ruleset, seven pillars covering identity, audience, long-form, shorts, newsletter, voice and growth stage, plus two performance logs that fed real data back into the rules. Every output passed a five-step check before production: does it fit the channel, is it for the right viewer, is it the right format, does it sound like me, and is it the right move for this growth stage.

I used the performance data from every output to change how I made the next ones.

The four-lane test

I tested four categories in parallel, so I didn't have to assume I already knew what would work:

What worked

One lane was ahead during the test. Lived experience produced the only outliers. The Bulgaria video did a 7.5% click-through rate on 134,653 impressions: 14,763 views, 92 new subscribers and 760 watch hours. The Australia video confirmed the pattern with a 6.6% click-through rate, and viewers watched 41.8% of a video that runs over twelve minutes. The other six videos across the other three lanes averaged 176 views and a 3.4% click-through rate during the test.

The newsletter went out weekly without missing an issue: seven issues, a 57% average open rate and almost no clicks.

Why I stopped at eight, not ten

By videos six and seven the picture was clear: one lane was producing outliers and the other three weren't close. YouTube data needs about 72 hours to settle, so I read patterns across several uploads rather than reacting to single videos. Video eight was already filmed and covered the fourth lane, so I let it run to get complete data across all four, then stopped.

Newer uploads in the weaker lanes weren't pulling in new viewers. They were sending existing viewers back to the Bulgaria and Australia videos, which backed up the decision.

Problems, and what I did about them

Three of four lanes underperformed. I captured complete data on all four before stopping.

Newsletter opens strong, clicks near zero. The newsletter wasn't built for list growth at that stage, so sharper writing wouldn't have fixed it.

One video lost half its viewers at 30 seconds. The packaging didn't match the opening. I logged the cause, wrote a rule into the system so it couldn't happen again, and kept the script for later with corrected packaging.

What I did, and what the editor did

The editor handled long-form cutting and assembly, on briefs and feedback from me. Everything else was mine: the scripts, Shorts, newsletters and Substack Notes, thumbnails, titles and packaging, research, performance review, and the pipeline, rules and logs behind it.

Tools

• Claude Projects for the pillar system, idea checks, scripts and performance review
• Claude Code, Remotion and HeyGen for programmatic animated intros on the best-performing videos
• A self-hosted AI agent on AWS, reached through Telegram, for quick operational questions
• YouTube Studio analytics feeding the logs, Substack, Premiere Pro and DaVinci Resolve, Canva, Kling and ElevenLabs

After the test

The strongest videos kept working after I stopped. As of October 2026, the channel has passed 100,000 lifetime views. In the last 28 days it picked up over 12,000 views with nothing new uploaded.

Bulgaria is now past 52,000 views and Australia is at about 8,500. Georgia, the video I let run to finish the test, was one of the six that averaged 176 views. It has since reached about 7,400 views, with a click-through rate of 8.9%. All three hold around 40% of the video on average.

The output

The channel is now dormant, with the archive intact. The two strongest videos are public:

• Living in Bulgaria 2026: The Honest Reality · 52,000+ views to date, 7.5% click-through rate during the test
• I Lived in Australia for a Year: The Honest Reality · about 8,500 views to date, 41.8% of the video watched on average during the test
• Newsletter archive · 7 issues, 57% average open rate