Back to Guides
Content Creators

Your Next Video Idea Is Hiding in the Comments


Most creators scroll their comments, remember the loudest or most recent ones, and treat those as a sense of what the audience wants. That's not the same as knowing what's actually being asked most often, and it means the same real request can sit unanswered for months because it was never the loudest one. A creator who scrolls through 200 comments after posting is far more likely to remember one sharply worded criticism than the fact that fifteen different people quietly asked the same practical question in fifteen different ways.

Warning

A comment you remember and a comment that recurs are different things. The one you remember might just be the one that annoyed or flattered you most, not the one the most people are actually asking.

The comment you remember isn't always the one worth answering. The comment that keeps showing up is.

Why scrolling doesn't surface the real pattern

Reading comments in order gives you an impression, not a count. A question asked by five different people in five different ways doesn't look like a pattern when you're reading one comment at a time, it just looks like five separate comments. The human eye is good at spotting one comment that stands out emotionally and bad at counting how many differently worded comments are secretly asking the same thing.

Take a baking channel where, across a video's comments, one person asks "why did my starter collapse," another asks "is it normal for it to deflate overnight," and a third asks "how do I know if I overproofed." Scrolled one at a time, these read as three unrelated questions from three different people with three different problems. Grouped by the underlying question they're actually asking, they're the exact same question asked three different ways, and that's a much stronger signal for what the next video should cover than any single comment on its own.

Before: scroll the comments after a video goes up, reply to a few, move on with a vague sense of what people liked.

After: paste a batch of comments into Claude, ask it to group them by recurring question or request, and get a short list of actual topics with more than one person asking.


A process for turning comments into topics

  1. 1

    Collect a real batch, not a handful

    Pull comments, DMs, or survey replies from more than one video if you have them. A pattern across multiple videos is a stronger signal than a pattern within just one.

  2. 2

    Ask Claude to group by recurring question, not summarize

    A summary just compresses what was said. Ask specifically for comments grouped into the question or request each one is actually making, so you can see how many people are asking the same thing in different words.

  3. 3

    Ask for a representative quote per group

    A quote grounds the group in something real, and it's useful later when you want to reference "a lot of you asked about this" without making the claim up.

  4. 4

    Turn each group into one follow-up topic

    Ask for a specific video or post idea per group, not just the topic label. "Sourdough starter feeding schedule" is a topic. A concrete video idea is something you could put on a calendar.

Prompt

Here are 40 comments from my last video. Group them into the three most common questions viewers are asking, and suggest one follow-up video topic for each.

If you've built up comments from several past videos, it's worth running the same grouping across all of them together, since a question that shows up in three different videos' comments is a stronger, more durable signal than one that only appeared once.

Prompt

Here are comments pulled from my last four videos, not just one. Group them into recurring questions across the whole set, and note which questions showed up in more than one video's comments specifically.

Inside Claude Tutorial

Finding the real pattern in scattered feedback is transferable.

Grouping a pile of individual reactions into the few that actually recur, instead of reacting to whichever one you saw last, applies well beyond comments. The app has a full lesson on it, with practice that carries over to any feedback you collect.

Coming Soon
Download on theApp Store
GET IT ONGoogle Play

When the loudest comment isn't the most common one

Sometimes the process surfaces a group that's smaller than the comment that's been bothering you all week. That's useful information on its own, since it tells you the loud one is probably not worth building a whole video around. A single sharply critical comment can occupy far more mental space than its actual weight in the audience's real priorities deserves.

Tip

Ask Claude to flag which group has the most individual commenters behind it, not just which group has the most total comments. A handful of people commenting five times each isn't the same as five different people asking once.

Prompt

Of the three groups you found, which one has the most distinct people behind it, not just the most total comments? I want to prioritize based on reach, not volume from a few repeat commenters.

It's also worth checking whether the loud comment that's been on your mind actually belongs to any of the recurring groups at all, or if it's a genuine one-off, since that distinction is what tells you whether it deserves a video or just a reply.

Prompt

This specific comment has been bothering me all week: [paste comment]. Does it match any of the recurring groups you found, or is it a genuine outlier that doesn't represent a broader pattern?

Continue reading