After using [[ChatGPT]] to analyze the past two years of my channel analytics, I came away with two interesting insights. The first: **people come to my channel to either make a decision or learn how to do something.** The second: **opinion videos on their own will not work, they need to be framed as a decision or problem/solution.** With this in mind, let's use these as my repeatable formats: # Reviews / Comparisons (Decisions) Even if I don't want to be an e-reader review channel, this is still my biggest opportunity for growth. I can focus on occasional product reviews, alongside lots of comparisons. My unfair advantage: I can begin doing long-term reviews since I have so many products, and I can also do lots of comparisons between the different products. It also doesn't have to be e-readers. I can compare Kindle vs Phone Reading, Tracking Books vs Not Tracking Books, Reading One Book vs Multiple Books at a time, etc. # Systems and Workflows (How-to) This is the bucket I can use to lean a bit more into PKM and other worlds outside of just Book Tech. What are the systems and workflows I have in relation to reading? I imagine this will be a lot of notes or highlights related content. Teach people how to do something, step by step. # Opinions (Problem/Solution) Repurposing my newsletter into a problem/solution is a good opportunity for additional content. If I frame my opinion as "My thoughts on XYZ", it will not do well. But if I package the same idea as a big mistake or problem, and here's the solution, it will do much better. # First Impressions Timely content based on recent updates or Book Tech news tends to always do well. This can be low-effort content that I publish with a focus on getting it out the door to ride the trend while it's hot. These videos can be a mix of the other buckets, depending on the context (decision, how-to, problem/solution). --- Related Notes: - [[My YouTube channel]] - [[My strategy for YouTube]] This note was originally created on **<% moment(tp.frontmatter.created).format("MMMM D, YYYY") %>**.