How to pick YouTube video ideas that actually get watched

A bad video idea filmed beautifully is still a bad video. That's the expensive lesson on YouTube, because every video costs you a weekend and you only find out afterwards.

The good news is you can find out most of it beforehand.

The idea matters more than everything else combined

Your thumbnail and title decide whether someone clicks. Your idea decides whether the title could ever have been interesting. You cannot write a compelling title for a topic nobody wanted.

So the order is idea, then title, then thumbnail, then film, whatever length you are making. Most people do it backwards and wonder why the good camera didn't help.

Look for the videos that beat their channel's normal

Go to a handful of channels near your size in your subject. Look for videos that did far better than that channel usually does. Not their biggest ever, the ones that broke their pattern.

Those are proof that a specific topic pulled harder than the channel's baseline, with the same subscribers and the same face. That's about as close to a controlled test as you'll get for free.

Check whether it's a topic or a moment

Some standout videos did well because of timing, a news event, or a collaboration. Those don't repeat. What you want is a topic that would still be interesting in six months.

A quick test. Would somebody search for this next year? If yes, it's a topic. If it only made sense that week, leave it.

Make it yours before you film

Take the topic, then find the angle only you can do. Your own numbers, your own mistake, your own client story. The topic gets the click and your angle is why they stay and subscribe.

A video that could have been made by anyone gets watched by nobody twice.

Keep a running shortlist

The people who never run out aren't more creative, they're just collecting all week instead of deciding on Saturday morning. Keep a list of standout videos from your space with a line on why each one worked, and film from the list.

Kadenza keeps that list current for you, including what the video actually said, so you're choosing from real evidence rather than a title and a hunch.