This is the exact three-step system we use at our YouTube agency, where we’ve generated billions of views for clients.
TL;DR
- Audience Signals — Mine the data you already own (sales calls, support tickets, comments) with AI to surface topics your audience is asking for.
- Market Validation — Confirm real demand and find room to win using search data and competition analysis.
- Packaging & Execution — Lock a proven format and clickworthy packaging, then make sure you can produce the best video on the topic.
Get all three right and you’ve got a validated idea with outlier potential — not a guess.
Partner with the leading YouTube Agency for brands. 
Why coming up with ideas isn't the problem
Coming up with ideas has never been easier. You can fire up an AI chat right now and generate 100 video topics in under a minute, and the list will look great.
But a list of ideas is only a starting point, and a weak one if you don’t prompt AI properly. So the real unlock isn’t generating ideas. It’s validating them: looking at a list of 100 topics and being able to say, “this one is worth making.”
Here’s what our three-step video validation process does, and why you need all three:
- Step 1 gives you ideas grounded in what your audience already asks for.
- Step 2 proves the market actually wants them.
- Step 3 confirms you can position and package the video to get views.
Miss one and the system doesn’t work. Let’s start with step one.
How do you come up with YouTube video ideas?
Step 1 - Audience Signals
The first mistake people make when coming up with video ideas, is starting from scratch and treating it like a creative exercise: “What could we make videos about?” Wrong question.
The right question is: “What does our audience already tell us they care about?” And you already have the answer sitting inside your business.
Your sales team talks to prospects every day. On those calls, people ask questions, raise objections, and describe problems. Every one of those moments is a potential video topic. If ten different people ask the same question, that’s not a coincidence — that’s a demand signal.
The same is true across the rest of your data:
- Sales call transcripts — recurring questions and objections
- Support tickets and CRM notes — pain points that keep surfacing
- Marketing emails — which ones triggered the most replies
- Comments on your videos and social posts — what your existing audience reacts to
- Existing content performance — your highest-engagement blog posts, your outlier videos, and just as importantly, what flopped
Your audience is already telling you what they want. You just need to listen.
Loop in AI to do the listening
You’re not going to read through hundreds of sales call transcripts yourself. So gather the raw material and feed it to AI along with context about your business.
Inputs to feed your AI:
- Sales call transcripts
- Customer emails
- Support tickets
- CRM data and deal notes
- Comments on your videos and social posts
- Business context: target audience, the problems you solve, your UVP, your website
Then prompt it with something like this:
Based on the audience data I've provided, identify the most common problems, questions, and pain points that keep coming up. Then generate YouTube video ideas around them.
The output you get when AI has real context like this versus little to no context isn’t even comparable. One gives you good-sounding ideas. The other gives you topics that read like they came from someone who’s operated inside your business for years.
This post is proof the system works
This is a live example. We didn’t brainstorm this topic out of thin air. We uploaded hundreds of sales call transcripts into NotebookLM and asked it to surface the challenges and questions that come up most.
We use NotebookLM a lot because it can handle massive amounts of context at once and only pulls from the sources you give it (not its general training data) so the answers are grounded in reality, not hallucinations.
Of the results it returned, “we don’t know what videos to make” landed in the top five. So this topic isn’t a guess. It’s a pattern AI surfaced from real conversations with our actual audience. That’s how we knew it was worth pursuing.
What you want out of Step 1 is a pool of ideas grounded in data. Not guesses. Now you need to know which ones can actually outperform.
Ready to level up on YouTube? Get started with a discovery call.
How do you know if a YouTube video idea has demand?
Step 2 — Market Validation
You’ve got a good idea. Now prove there’s enough demand for it.
There’s two things to look at: search data and competition.
1. Search data: are people actually searching for this?
Start with YouTube autocomplete. Type your topic into the YouTube search bar — but don’t hit enter. The autocomplete suggestions are phrases YouTube already knows have demand. If YouTube is finishing your sentence, that’s your first signal.
Then check Google Trends. Search both broad and long-tail terms, and look at the top and rising queries. That’s where you’ll find specific angles gaining momentum. For this video we looked at terms like youtube content strategy, video ideas for youtube, and ai for youtube ideas, all of which showed solid demand.
Toggle between the YouTube Search and Web Search filters while you’re in there. YouTube Search shows platform-specific trends; Web Search shows broader demand.
This matters more than it used to, because YouTube content now shows up in AI search overviews as one of the top cited domains. (If you want to go deeper on this, here’s our full guide on how to dominate AI search with YouTube.)
Then get real volume numbers with a keyword tool like vidIQ. Search your broad terms and longer-tail variations, and look at two things: volume and competition. Ideally you want high volume and low competition. Demand outpacing the supply of good content, which leaves room to win.
But you don’t always need that. Context matters. A highly specific, bottom-of-funnel video (like a product comparison) will have lower volume by nature, but the intent behind those searches is far stronger.
What you want over time is balance: some broad, high-volume topics for reach, and some lower-volume, high-intent topics for business results.
2. Competition: is there room to win?
This is arguably the most important data you’ll uncover. Search your topic on YouTube and see what’s going on.
- Are there a lot of videos on this topic?
- Are they old or new?
- Are they getting views?
Use the search filters to sort by relevance vs. popularity and old vs. new.
- Strong, recent videos everywhere? Harder fight — the supply already exists. Not impossible, but you need a clear "better or different" angle.
- Weak competition (few videos, outdated, low quality)? Potential opportunity — unless it's weak because nobody cares about the topic.
Then look for outliers — videos that pulled far more views than the channel’s subscriber count or typical average. If a channel that usually gets a few thousand views suddenly hits 300,000, that’s a strong signal: demand for the topic far exceeds what’s being supplied. That’s where you want to be.
Using this post as an example: there are lots of videos on idea validation, and some are strong, but most are a few years old — so a newer take with current information is valuable. The recent uploads are mostly from small creators and under eight minutes, which tells me there’s room to win with a more in-depth, higher-value piece.
That insight about value is exactly what carries into Step 3.
How do you package a YouTube video to win?
Step 3 — Packaging & Execution
The idea is validated: your audience cares, the market confirms demand. Now, how do you position it to actually win?
1. Find a winning format
A format is the structure or style of the video, not the topic. The same topic could become a listicle, a myth-busting video, a comparison, an expert interview, or a tier list — each feels like completely different content even with the same core idea.
So don’t reinvent the wheel. Find formats already proven to work and apply them to your niche. We use the ClickPilot extension to save winning videos, so we always have a bank of proven formats to pull from.
And don’t only look in your own niche. Searching adjacent or completely different niches is one of the best hacks.
We did exactly this: tier-list videos have dominated gaming and pop culture for years, so we borrowed the format for YouTube strategy, ranking tactics from S-tier to F-tier, and put our own spin on it with a physical tier list instead of the usual digital one.
2. Create clickworthy packaging
Format and packaging go hand in hand. Start thinking about a thumbnail and title that win the click through curiosity, confidence, and clarity.
Look at what’s already working and lean on proven patterns while staying different enough to stand out.
Two examples:
- Say you're making a video on financial advice. Search an adjacent niche — legal — and you find "legal myths that will land you in prison." Adapt it: "financial myths that will make you bankrupt."
- The channel Lofi Cinema made a video essay on creativity that hit 2.7 million views — a massive outlier. Joanna Wiebe recognized the packaging was working and adapted it for her own video, which did over 100x her typical performance. She didn't invent new packaging. She built on something already proven.
You don’t need final packaging before you produce, but you should have a few strong ideas you know have potential. Because if the packaging doesn’t land, it doesn’t matter how strong the idea is — nobody will ever see it.
3. Run the "best video" test
Ask yourself: could your video be the best video someone watches on this topic? You don’t need the biggest channel or the highest production value.
You need the most valuable video — more depth, a more novel angle, and ideally information only you have access to. If your business holds data or knowledge no one else does, that’s what makes your video rise above the rest.
If you look at what already ranks and can’t meaningfully improve on it, sharpen the angle or move on.
When an idea is worth making
That’s the system. Three levels:
- Your audience cares about it.
- The market wants it.
- You know how to execute it to win the click and keep viewers watching.
When an idea clears all three, that’s your green light — a validated idea with outlier potential. You’re not guessing anymore. You’re executing proven ideas grounded in research and data.
