Reposition the channel you already run when your recent videos in the new direction still hold their audience. Start fresh only when the old subscribers, the old name, and the old comment culture all fight the change. I use three numbers to make this call, and they settle it faster than a month of second-guessing.
The first is how your last five uploads in the new direction performed against your old baseline. The second is how many comments asked for the format you are leaving. The third is whether the channel name still describes something you plan to make. If the first number is close and the other two are mild, you reposition. If the first is far behind and the other two are loud, you relaunch.
A reposition is cheaper than a relaunch whenever the people you already have still want where you are going.
Keep the channel when your new-direction videos land within 20 percent of your old median views. That gap usually means the audience followed the shift without being told. A woodworking channel I follow slid into tool reviews last year and held about 80 percent of its watch time, because everyone who watched the builds already cared about the tools.
Subscriber count is a weak signal on its own. A dormant list of 200,000 people who subscribed for a format you are dropping will drag your click-through rate down for months. What matters is whether the viewers who open your notifications still open them after the change. Check the returning-viewer share in YouTube Studio, not the subscriber badge.
Start a new channel when the reposition changes the promise so much that the old name, thumbnail style, and comment section all contradict it. If people subscribed for fitness transformations and you now want to teach software, the overlap is close to zero, and the old channel becomes a weight you carry for free.
A fresh start also clears the archive problem. YouTube keeps routing old videos to old viewers, and those viewers keep signaling the system that your channel is about the thing you left. I watched a creator double their first thirty-day retention on a new channel for one reason: there was no conflicting history for the algorithm to read.
My preferred move sits between keep and relaunch. Keep the channel, unlist the videos that no longer represent you, and publish the new direction for ninety days before you decide anything permanent. Unlisting is reversible. Deleting is not. I leave the old uploads unlisted rather than removed so the watch time stays attached and I can restore them if the pivot stalls.
Attach one number to the checkpoint. Mine is median views per upload, not total views, because a single outlier can flatter a failing channel for months. If the median is climbing by day ninety, you keep going. If it is flat, you now have data instead of a feeling.
Write the decision as a single sentence before you touch the channel. A reposition you cannot state in one line will not survive month three.
A series gives viewers a reason to come back and gives the recommendation system a reason to keep suggesting you. Here is how to design one that actually grows.
Your first 100 subscribers come from people who already trust you, not the algorithm. A concrete plan for earning each one without gimmicks or paid ads.
Niche selection fails when it is theoretical. Choose a lane by testing three topics for two weeks each and letting real responses pick the winner.
Analyze a channel, extract its structural DNA, and generate scripts engineered from patterns that actually work.
Analyze a Channel