The workable methodology is a two-stage process: first, use YouTube’s own demand signals (autocomplete/search suggestions and, for your existing channel, the Search terms report in YouTube Analytics) to identify what people are actually typing, then manually audit the current top-ranking videos for each candidate keyword to judge whether the existing supply is actually good. A real opportunity exists where demand is evident but the top-ranking videos are outdated, thin, poorly produced, or don’t fully answer the query, i.e. a supply-side gap, not just a demand number in isolation.
Stage one: mapping demand with YouTube’s own signals
YouTube doesn’t publish exact search volume numbers the way Google Ads’ Keyword Planner does for web search; there’s no official YouTube tool that tells you “this phrase is searched X times per month.” Treating any third-party tool’s YouTube volume estimate as an authoritative, YouTube-disclosed number would be inaccurate, since YouTube itself doesn’t expose that data publicly. What YouTube does provide, directly and reliably:
- Autocomplete/search suggestions, worked systematically rather than casually. Typing a seed phrase into YouTube’s search bar surfaces suggested completions, which reflect actual common query patterns among YouTube users. A single seed term only surfaces a handful of suggestions, so the useful version of this exercise is systematic: take your seed term and append each letter of the alphabet one at a time (“[seed] a,” “[seed] b,” and so on, sometimes called the alphabet soup method), then repeat with numbers where relevant, then repeat again with common modifier words layered on the seed (how, why, best, vs, review, tutorial, for beginners, mistakes, cost). Recording every distinct suggestion this produces, rather than stopping after the first pass, is what turns autocomplete from a quick glance into an actual demand map.
- The Search terms report in YouTube Analytics. For an existing channel, this report (found in Studio Analytics, under the Research tab or the traffic-source breakdown for Search) shows the actual queries that led viewers to your channel’s content, which is first-party demand data specific to your niche and audience, not an estimate. This report is uniquely valuable precisely because it isn’t hypothetical: every query listed there already drove a real click to your channel, so it tells you what demand your content is currently capturing, and by extension what adjacent or related queries in the same report you might not yet have dedicated content for.
- Related/suggested video patterns and comment sections on top videos in your niche, which surface the specific sub-questions and phrasing real viewers use, supplementing the more structured autocomplete and Analytics data.
This stage produces a list of candidate keywords/phrases with reasonable confidence that real search demand exists, even without an exact volume figure attached to each one.
Stage two: auditing supply quality for each candidate
This is the step that’s frequently skipped, and it’s the one that actually determines whether a keyword is a genuine opportunity rather than just a popular query that’s already well served. For each candidate keyword:
- Search the term on YouTube as a real user would, and look at what’s actually ranking in the first page of results, not just the top one or two videos.
- Check publish dates first, before watching anything. A niche where the top-ranking results are several years old is an immediate, low-effort signal worth flagging for closer review, since a query with ongoing search demand but stale top content is exactly the kind of timing gap current supply may not have kept pace with.
- Watch enough of each top-ranking video to assess whether it actually, fully answers the query, or whether it’s thin, tangential, padded with unrelated content to hit a runtime, or technically on-topic but shallow relative to what the query implies the viewer actually needs. This means watching more than the first thirty seconds; a video can open strong and still fail to deliver on the specific query by the midpoint.
- Assess production quality relative to what the topic deserves, not against some abstract universal bar. A simple screen-recording tutorial style might be entirely appropriate for a software walkthrough query, while the same style would be a clear quality gap for a query where viewers are evaluating a physical product or comparing options, where visual demonstration matters more. Judge quality against the format the query implies, not a single fixed standard.
- Check comment sections on the top-ranking videos for viewers explicitly saying the video didn’t answer their question, was outdated (referencing old software versions, old prices, old procedures that have since changed), or asking a follow-up question that goes unaddressed in the video or in creator replies. This is direct, first-party evidence of unmet demand within an existing “answer,” and it’s often more reliable than your own judgment of whether a video is thorough, since it reflects what actual searchers felt was missing.
- Cross-reference against your own Search terms report over time, if you have channel history, to see whether a query keeps appearing as a traffic source despite you not having dedicated content that squarely answers it. A recurring query in that report, month over month, with no corresponding video on your channel that directly targets it, is a strong internal signal of a gap, arguably stronger than an external competitor audit, because it confirms real viewers in your specific niche are already looking for this and finding their way to you only tangentially.
Why this framing, not a single official tool or formula
This process draws on YouTube’s legitimate, documented data sources (autocomplete, Search terms reporting) but the overall methodology, the specific practice of manually auditing top-ranking supply for quality gaps, is practitioner process built up across the SEO/YouTube-optimization industry, not a single official framework YouTube has published as “how to do keyword research.” It’s worth being explicit about that distinction rather than implying YouTube has endorsed a specific opportunity-scoring method, since it hasn’t.
What to do with a validated opportunity
Once a keyword clears both stages (real demand signal, genuinely weak current supply), the practical next step is making sure your content actually closes the specific gap you identified, not just retreading the same angle as the existing top results. If the gap is that existing videos are outdated, your content needs to be genuinely current and should say so clearly in the title/description/opening, ideally naming what specifically changed since the older top-ranking videos were published. If the gap is that existing videos don’t fully answer the question (based on comment evidence), structure your video to explicitly address the sub-questions viewers were asking in those comments, and consider addressing them early rather than assuming viewers will watch to the end to find them. A keyword with confirmed demand and weak existing supply is still only an opportunity if your video is materially better at answering the query than what’s currently ranking; matching the weak existing supply doesn’t capture the opportunity, only clearly exceeding it does.
