There’s a lot of different ways to make content. And it’s not always easy to find ideas, validate them, and find good ones.

So here’s a trick I’ve been using recently. My latest big video did almost 12 times my normal views, and the idea came from this.

Here’s what you do step by step.

Start with demand

Step one is knowing why YouTube pushes some videos and ignores others. It ultimately comes down to demand.

There are two things that decide whether someone picks your video. Your topic, and your packaging. The topic is whether anyone wants it at all. Packaging decides if they notice it next to everything else on their homepage.

Great packaging can’t save a topic nobody really cares about. I’ve made this mistake many times and I'm sure I'll make it again.

Every step after this one is about finding demand before you film.

Find who’s winning right now

Step two is finding the outliers in your space. Those are videos that did way more views than their channel’s normal. They show you what people already want to watch.

You can do this on YouTube. You search what your viewer would type, then dig through channels one at a time. It just takes a lot of time.

So I’d highly recommend using an outlier tool. It pulls the outliers across your whole niche into one place, with the math already done. These are the three I’d look at:

• ​Velio​
• ​1of10​
• ​vidIQ​

I’m an affiliate for all three, so I get a small cut if you use my links. <3

Then keep it (relatively) recent. That’s step three. An outlier from three years ago shows you what people wanted three years ago. I stick to the last year. You know what’s appropriate for your niche.

You don’t need huge numbers either. An established channel doing 50% more views than normal is a pretty big deal.

On the other end, skip anything over 200 times normal. Those are lottery wins and they’re not worth recreating.

Pull out the topic and the title

Now look at what your outliers have in common. That’s step four. Group them by subject. When the same subject keeps showing up across different channels, that’s your topic.

When I did this, one group was hard to miss:

“If Nobody Watches Your Videos, Do This.”
“Why Your Views Die After 48 Hours.”
“If Your Videos Get Less Than 500 Views, Do This.”

It’s the same topic, over and over. People want to know why YouTube isn’t showing their videos.

Step five is finding a title framework that’s working. A title framework is the shape of the title, the words wrapped around your topic.

Here are a few that are working right now, each with the source outlier:

• Give me [time], I’ll make your [thing] [result]​
  “give me 9 minutes, I’ll make your thumbnails 81% better” (9 times its channel’s normal)
• I tested [common advice] for [time]. [Surprising result].​
  “I Tested YouTube Consistency for 90 Days. Hard Work Doesn’t Move the Algorithm.” (70 times normal)
• Don’t [obvious move], do THIS instead to [result]​
  “Don’t Teach on YouTube - Do THIS Instead to make your first $100K” (77 times normal)
• If you have [situation], [do the thing] ASAP​
  “If you have a Job, Start a YouTube Channel ASAP” (57 times normal)
• I reviewed [a pile of something] to find [the pattern]​
  “I Reviewed Small Channels to Find the Most Common Mistakes” (44 times normal, from a channel with 12,000 subscribers)
• Treat [your thing] like [something unexpected]​
  “Treat Social Media Like a Video Game and Get Rich” (42 times normal)

My 12x video used a version of that first one.

Look outside your niche for title frameworks too. What people click comes down to psychology, and psychology works the same everywhere. If you talk about quitting alcohol, a title that worked for quitting smoking can work for you too.

I know that sounds like copying. And it is, as long as you copy like an artist.

Take a title framework from one video, a word from another, and a thumbnail idea from a third. Mix them and make it yours. Just don’t be the jerk who rips off someone’s whole video.

Make the idea yours

Step six is putting the two together. Drop your topic into the title framework. Write a few versions until one feels right.

This is a part AI is really good at. Give it your topic and the title frameworks you found, and ask for a bunch of versions. It’ll give you a lot of options fast.

You still make the call. I almost always end up tweaking whatever it gives me, but it gets me to a strong starting point in a fraction of the time.

Step seven is finishing your angle. Your topic already tells you what people want. The last piece is what your video gives them that the others don’t.

That last piece is sitting in the comments of the outliers you found. Read what people loved, and what they kept asking for that the video never gave them.

When I did this for my latest video, one outlier showed a spreadsheet of its research and never explained it. The comments kept asking, “Show us the research.”

So the video I just made walks through the whole process, start to finish.

Reading comments by hand is slow. I built a free Chrome extension called Boundless Insight to do it. It'll do this faster and better than you can.

One button breaks down a video and its comments. You can see what people liked and what they wanted more of. Get Boundless Insight here​

Bonus tip: views per hour

Once you’re using an outlier tool, there’s one more number worth checking. Views per hour. It’s how fast a video is getting views right now.

An outlier tells you what people wanted when it was posted. It can’t tell you if they still want it today. Views per hour can.

When I sorted my outliers by score, one of the top videos was “How to Actually Edit Shorts That Go Viral Every Time.” It did 140.9 times its channel’s normal. But it wasn’t getting any views per hour. The people who wanted it already watched it.

When I sorted by views per hour instead, a Kallaway video sat near the top. It did 28 times his normal, which is a lot lower than 140.

But six months after it came out, it was still getting over 400 views an hour. That’s close to 10,000 views a day. People want that one right now.

One heads up. Brand new uploads float to the top, because they just came out. Look past them. The older videos still pulling views every hour are the ones worth studying.

YouTube doesn’t show views per hour anywhere. All three of the tools I mentioned do. You can take this a step further and plug Velio or vidIQ into Claude with their MCP and it'll be able to filter the views per hour results so you don't see the recent uploads.

Try it on your next idea

Pick the next video you’re planning. Before you script or film anything, run it through all seven steps. If you can, check the views per hour on what you find.

If you don’t have an idea yet, start at step two and let the outliers hand you one.

None of these steps are hard. The first time through takes a while, but once you’ve done it a couple of times, it goes pretty fast. And the more you do it, the more patterns you start to see.

It won’t work every time. You might make two or three videos that don’t hit. But you’re not leaving it to luck anymore.

Thanks,
Dave