YouTube Keyword Research: Finding Search Terms Your Channel Can Actually Rank For

Most keyword research advice hands you a tool and a search volume column, and stops there. That is why so much of it fails. Volume tells you how many people are looking; it says nothing about whether your channel has any chance of being the one they find. For a channel under a few thousand subscribers, winnability is the only column that matters.

This guide covers where to actually get YouTube keyword data, how to score a keyword for competition without paying for anything, and how to turn a research session into a publishing calendar that lasts six months.

Start with intent, not volume

Every YouTube search falls into one of four intents, and each one demands a different video. Getting this wrong is the most expensive research error, because it produces a technically well-optimised video that satisfies nobody.

Read the query out loud and ask what the person wants to be true five minutes from now. That answer is your format.

Seven free sources of real keyword data

1. YouTube's own search suggestions. Type a seed phrase into the YouTube search bar and read the dropdown. These suggestions come from real search behaviour on the platform, which makes them more relevant than any third-party estimate. Then run the alphabet trick: type your seed followed by each letter — "obs studio a," "obs studio b" — and harvest what appears.

2. The Research tab in YouTube Studio. This is the single most underused feature in Studio. It shows what your own viewers and the wider YouTube audience are searching for, flags searches with low result coverage, and lets you save queries to a list. Because it draws on YouTube's data rather than a scraped estimate, it is the closest thing to a free first-party keyword research tool.

3. Your existing traffic source report. Studio shows the exact search terms bringing people to each of your videos. If a term appears repeatedly against a video that only partly addresses it, that is a proven-demand topic waiting for a dedicated video.

4. Google Trends. Switch the search type to YouTube Search — this is a distinct dataset from web search and most people never change the dropdown. Trends will not give absolute volume, but it answers the two questions that matter: is this rising or dying, and which of two phrasings is more common?

5. Comments on competitor videos. Unanswered questions in the comments of the current top-ranking video are pre-validated topics, phrased in your audience's own language.

6. Google's autocomplete and "People also ask." Web search intent overlaps heavily with video intent for how-to and comparison topics, and the People also ask box is a ready-made list of subtopics.

7. Community and forum threads. Niche subreddits and forums surface the questions people ask when they cannot find an answer on YouTube — which is precisely where the gaps are.

The competition test

Once you have a candidate keyword, spend three minutes on this test before committing a week of production.

1. Search it and read the first six results. Not the first twenty — nobody scrolls that far.
2. Check subscriber counts. If every result comes from a channel more than ten times your size, and the videos are thorough, choose a narrower variant.
3. Check dates. Results all older than two or three years in a fast-moving topic is an open door. Freshness alone can outrank authority when software versions have changed.
4. Check relevance. If the top results are only loosely related, YouTube is showing them because nothing better exists. That is the strongest possible signal to make the video.
5. Check view distribution. If one video has 2 million views and the rest have 4,000, the query is dominated by a single answer. If results cluster between 20,000 and 80,000, the query is winnable and stable.

Why long-tail keywords win for small channels

A long-tail keyword is a longer, more specific phrase with lower volume and much lower competition. "Camera settings" is a head term. "Best camera settings for indoor video without a light" is long-tail — and it is the one a new channel can own.

Three advantages compound. Ranking is achievable, because large channels rarely make something that narrow. Intent is stronger, so retention and satisfaction signals are better. And long-tail videos accumulate: twenty videos each owning a small query produce more reliable monthly traffic than one video that briefly touched a big one.

Long-tail also tends to attract higher-value viewers. Someone searching a precise, problem-shaped query is closer to a decision than someone browsing a broad category, which is one reason narrow videos often carry better ad rates and better affiliate conversion than their view counts suggest.

When a paid keyword research tool earns its cost

Be honest about the sequence. For your first thirty videos, free sources cover everything you need; the constraint is your production capacity, not your data. A paid tool starts earning its keep when you publish weekly and your bottleneck becomes triage — scoring fifty candidates quickly rather than finding five.

What paid tools genuinely add: estimated volume, a competition score, historical trend data, and bulk export. What they do not add is judgment. Every tool's difficulty score is a model, and the three-minute manual test above beats it more often than vendors would like to admit.

Building a keyword map

Do not research video by video. Research once per quarter, in a spreadsheet, with these columns: query, intent, format, competition verdict, and a pillar it belongs to.

Group the results into three or four pillars — the sub-topics your channel intends to own. Under each pillar, aim for one broad "hub" video and six to ten narrow long-tail videos that link to it and to each other through end screens, playlists and pinned comments. This structure is what turns individual uploads into a library: a viewer who arrives at any one video has an obvious next step, and YouTube learns that your channel covers the topic comprehensively rather than incidentally.

Fill twenty to thirty rows and you have a six-month calendar. The relief of never opening a blank publishing schedule again is worth the afternoon on its own.

Five mistakes that waste research time

• Chasing volume you cannot rank for. Position 19 on a huge term earns nothing.
• Researching after scripting. Keyword research is a pre-production activity. Retrofitting a keyword to a finished video produces a title that matches nothing well.
• Ignoring your own analytics. The queries already sending you traffic are the highest-confidence data you will ever have.
• Optimising for tools instead of people. Google's guidance on people-first content applies just as well here: content built to satisfy a metric rather than a person tends to underperform on both.
• Treating research as one-off. Queries shift with software updates, seasons and news. A quarterly refresh keeps the map alive.

Frequently asked questions