Blog
Notes and ideas from the Tukey team.
- · 11 min read
YouTube Algorithm 2026: 7 Ranking Signals Every Creator Must Know
The 2026 shift from watch time to satisfaction changes everything. Here are the specific signals the algorithm now rewards, and how to structure your content around them. 500 hours of video get uploaded to YouTube every single minute. There are 65 million active creators competing for the same impressions. The platform has 2.7 billion monthly users watching 1 billion hours of content per day. The algorithm is not trying to find the best video. It is trying to predict which video will leave a s
- · 9 min read
YouTube Watch Time by Niche 2026: Average Retention Benchmarks
The benchmark table for 13 YouTube niches: healthy AVD%, danger zone thresholds, and optimal video length, so you stop measuring your retention against a number that was never meant for your content. A 40% average view duration on a gaming channel means your video is working. The same 40% on a tutorial channel means you have a serious problem. YouTube watch time by niche does not follow a single standard, and most creators are benchmarking against the wrong number. Here is the full data, broke
- · 8 min read
ChatGPT YouTube Scripts: 5 Things It Gets Wrong
How to stop burning your audience in the first 30 seconds and write AI scripts that the algorithm actually promotes. The average YouTube video retains 23.7% of its viewers. That number has gotten worse every year since 2022. And AI-generated scripts, written without any understanding of how YouTube retention actually works, are a significant part of why. If you have asked an AI tool to write your YouTube script and felt something was off, you were right. Not because AI cannot help. Because mos
- · 10 min read
YouTube Hook Formula: Stop the Scroll in 3 Seconds
How to engineer the first 3 seconds of any YouTube video to stop the scroll, command attention, and send exactly the right signals to the algorithm. 74% more viewers stay past the first 3 seconds when a video opens with a deliberate hook instead of a slow intro. Most creators spend hours on their thumbnails. They run keyword research for their titles. They re-edit the middle of their video three times. Then they open with "Hey guys, welcome back to the channel." That sentence is where the vid
- · 12 min read
Why YouTube Scripts Lose 60% of Viewers (4 Fixes)
The exact script problems that kill your audience retention before your video even gets started, with data from 10,000+ videos and the specific fixes for each. The average YouTube video loses more than 40% of its viewers in the first 30 seconds. For channels with slow intros targeting casual viewers, that number jumps to 60%. That is not a topic problem. It is not a niche problem. It is almost always a script problem. And it keeps happening because most creators diagnose it wrong. They look a
- · 11 min read
YouTube RPM by Niche 2026: Finance vs. Gaming Data
The complete 2026 RPM breakdown by niche, plus the scripting shifts that separate $3-per-thousand channels from $35-per-thousand channels. Finance YouTube channels average $35 RPM. Gaming channels average $3. That is an 11x difference from the same platform, the same ad system, and often the same view count. Most creators never do anything about this gap. They pick a niche they like, build a channel, and accept whatever RPM YouTube assigns them. The ones who actually close the gap understand s
- · 10 min read
The Faceless YouTube AI Stack 2026 (Under $150/Mo)
The full pipeline, stage by stage: research, script, voice, visuals, editing, thumbnail, and publish. Total monthly cost: under $150. 38% of new YouTube monetization attempts in 2026 are faceless channels. That number was 12% in 2022. Most of them are running on a stack of 6 to 8 tools. Here is what that stack looks like. The One Mistake That Kills Faceless Channels Before They Start Most people think faceless YouTube is a single-tool problem. They find one app that promises to "automate y
- · 12 min read
Why 97% of YouTube Scripts Plateau (And the Fix)
Retention-informed scripting is the framework the other 3% use to write videos the algorithm actually pushes. Here is exactly how it works, and how to apply it starting with your next upload. The Stat That Explains Your Channel 97 out of 100 YouTube channels never reach 1,000 subscribers. That is not a guess. Industry data consistently shows only about 10% of all YouTube channels ever cross the basic monetization threshold. Roughly 4% of active creators earn meaningful income from the YouTub
- · 19 min read
23 Claude Prompts for YouTube Creators (2026)
The complete prompt library for research, hooks, scripting, and titles, built around how the 2026 YouTube algorithm actually evaluates content. 87% of creators now use AI in their workflows. Most of them type "write me a YouTube script about X" and wonder why the output sounds like a corporate press release. The problem isn't Claude. The problem is the prompt. Claude is a reasoning engine. Give it context, constraints, and a specific output format and it produces work that a professional writ
- · 9 min read
How to Find YouTube Outlier Video Ideas That Work
The exact research process top creators use to identify concepts with proven viral demand, before they ever hit record. 500 hours of video are uploaded to YouTube every single minute. That is 82 years of new content every day. Your next video competes against 30,000 hours of fresh uploads created in the same hour it goes live. In that environment, guessing what to make is not a strategy. It is a coin flip with a very bad payout. The creators who consistently break through are not more creativ
- · 9 min read
YouTube Monetization 2026: Rules, Revenue & $10K/Mo
Marcus had been uploading for seven months. Forty-two videos. Consistent schedule. Decent production. He had done everything the tutorials told him to do. His watch hours: 1,200. His subscriber count: 340. He was not even close to monetization. And he had no idea why. The answer was not his editing. It was not his thumbnails. It was not his mic quality. He was making videos on topics nobody was searching for and the algorithm had no reason to push. The entire production stack was built on t
- · 6 min read
4 YouTube Tools: 90 Min/Video, $11,400 in Month 4
It was a Tuesday in January. Kelly had just uploaded her 19th video. Four months in. No studio, no team, no editor. Views that week: 600. Revenue: $0. She had a script workflow. She had a voiceover. She had a decent thumbnail. What she did not have was any idea whether anyone actually wanted to watch the topics she was choosing. Month four ended at $11,400. One thing changed. Here is what it was. The Mistake Killing 90% of AI YouTube Channels Every AI YouTube tutorial covers the same g
- · 7 min read
YouTube Algorithm 2026: You're Optimizing the Wrong 30%
Most creators treat YouTube like a search engine. They spend hours on keywords, tags, descriptions. They upload on schedule. They watch the subscriber count like it means something. 70% of all YouTube watch time comes from the algorithm. Not from search. Not from subscriptions. You've been optimizing for the other 30%. Here is what is actually running your reach in 2026. The Numbers Most Creators Never Look At Before anything else, here is the data that should change how you think about th
- · 3 min read
X Algorithm Explained: What Nobody Tells You
A like is worth 1 point. A reply is worth 27. If you reply back to someone who replied to you: 150 points. That's from the actual source code. Not my guess. The Algorithm Is Public. Nobody Reads It. X published the code powering your For You feed to GitHub in 2025. It's called xai-org/x-algorithm. 16,500 stars, sitting there, free to read. I went through it. Here's what actually matters. The Weight Table (This Is What You're Missing) The algorithm scores every post by predicting what a
- · 4 min read
Why We Built Tukey: The AI We're Afraid Of
What is the secret recipe for a viral account? How do you get a million views from one single video? Those are the questions that followed me around when I was working at ByteDance, where creators were desperately trying to find a path to growth and always believed that TikTok internal people like me might hold the answer. I spent months working with PMs from ByteDance's Ocean Engine team digging for better insights, trying to give creators something real. What I kept coming back to was this: it