YouTube keyword research is how you find the exact phrases people type into YouTube (and into Google, where video results sit on most pages), then map those phrases to titles, descriptions, chapters and whole content lines so your videos surface in Search, Suggested and Browse. It is a demand-discovery loop: you run it before you script, then validate it after you publish. Below is the way we run it: opportunity first, intent second, and vanity volume last.
YouTube Keyword Research
YouTube keyword research is the process of identifying, scoring, and prioritizing the search phrases real viewers use on YouTube so you can align video topics, titles, and metadata with proven demand.
What YouTube keyword research involves
Five steps, in this order:
- Collect candidate phrases from YouTube autosuggest, your channel’s search terms report, competitor titles and chapters, and a keyword tool set to YouTube.
- Check which engine owns each phrase. Search it on YouTube and on Google. If Google shows a video block for it, a video can rank in both places; if Google shows only text results, a page will usually serve the searcher better.
- Score the survivors on demand, competition, intent fit and whether the topic works as something people would watch.
- Map the winners to metadata: the primary phrase in the title and the first line of the description, related phrases in chapters and the transcript.
- Validate after publishing by reading which search terms actually brought viewers, then feed those back into step 1.
The sections below cover each step.
Why YouTube search is its own game
Treating YouTube like a smaller Google is the most common mistake we see. The two engines reward different things, and a keyword that’s golden on one can be dead weight on the other.
YouTube says its search ranking weighs three things: relevance (how well the title, description and video content match the query), engagement (signals such as watch time for that query) and quality (signals of expertise and trustworthiness on the topic). Google web search ranks pages on relevance and on signals like links. A query like “how to fix a leaking tap” is huge on YouTube because people want to watch hands doing it; “plumber near me” is a Google query that rarely sustains a video. Your research has to read which engine a phrase belongs to before you commit a production budget to it.
Google is the third surface. Video shows up on most results pages: in our study of 30 Google results pages, 29 carried a video feature. Google can also jump straight to a moment inside a YouTube video when the video has chapters, which rewards clean, query-matched chapter titles. If you’re mapping demand across formats, our keyword research primer covers the text side, and this page is its video counterpart.
Rule of thumb we hold clients to: if you can’t picture the thumbnail and the first 15 seconds for a keyword, it belongs on a web page. Don’t film it.
The sources that actually surface demand
You don’t need a $300/month tool to start. You need to mine the signals YouTube and its viewers already broadcast.
- YouTube autosuggest: type your seed phrase in the search bar and read the dropdown. The completions reflect what people actually search. Prefix tricks (“best ”, “how to ”, ” for beginners”, ” vs”) and the A to Z method (append each letter) expand a single seed into 50+ live phrases.
- Search terms in YouTube Studio Analytics: the YouTube search traffic source lists the exact queries that already brought viewers to your channel. It’s the highest-signal source you own, because it’s demand you’re already partly winning. Many channels also get a Research tab in Analytics that shows what viewers search across YouTube.
- Suggested-video rails: these reveal the topical neighborhood YouTube has clustered you into, and the adjacent intents worth targeting.
- Competitor titles and chapters: top-ranking videos for your seed are a free read on what satisfies the query. Note the angle as well as the words.
- Google’s own surfaces: “People also ask”, related searches and video results tell you which YouTube videos already appear in web search.
- Keyword tools: TubeBuddy, vidIQ, Keyword Tool, and Ahrefs Keywords Explorer with the search engine set to YouTube give estimated volume and competition. Treat the numbers as directional; the relative ranking between phrases matters more than the absolute figure.
Scoring: opportunity over vanity volume
A keyword with 40,000 monthly searches you’ll never crack is worth less than a 600-search phrase you can own this quarter. We score every candidate on four axes before it earns a slot on the calendar.
| Factor | What to check | Why it matters |
|---|---|---|
| Search demand | Monthly volume + trend direction | Confirms people actually want this; rising trends beat flat ones |
| Competition | Strength of top five results, channel authority | A small channel can’t out-rank a 2M-subscriber incumbent head-on |
| Intent fit | How-to, review, comparison, entertainment | Mismatched intent kills retention even when CTR is strong |
| Format viability | Can it be a watchable video? | Filters out text-only Google queries masquerading as keywords |
Lean into long-tail. Phrases like “youtube keyword research for small channels” turn intent into watch time far better than a bare head term like “keyword research”. They’re less contested, the viewer’s need is sharper, and they cluster naturally into series. The same low-contention logic we use for low-competition keywords in web SEO applies here: early-stage channels win on the edges.
Separate Search keywords from Browse keywords. Some phrases earn impressions in Search results; others mainly surface through Suggested and Browse, where the algorithm decides who sees you based on what viewers watch and how they respond. Both are valid; just know which lever a given keyword pulls so you set realistic expectations.
From keyword to metadata that ranks
Research is only worth the views it produces. Here’s how the winning candidates flow into a published video.
Title
Front-load the primary keyword in the first 40 to 60 characters, the part that survives truncation on mobile and in Suggested rails (YouTube allows up to 100). Then add a differentiator: a number, a timeframe, a result, or a contrast.
- How-to:
[Keyword]: How to [Outcome] in [Timeframe] - Problem to solution:
[Keyword]: Fix [Problem] Without [Pain Point] - Comparison:
[Keyword] vs [Alternative]: Which Wins for [Use Case]
Match the title to the intent exactly. Promising a tutorial and delivering a review tanks average view duration, and on YouTube that is one of the engagement signals that decides whether the algorithm keeps showing the video. The same click-then-satisfy discipline behind a strong organic CTR governs YouTube thumbnails and titles.
Description
Put the primary keyword in the first one or two lines, because only those show before the “…more” fold and in search results. Follow with a natural 200 to 300 word summary that weaves in two or three secondary phrases, then a timestamped chapter list. Chapters need a first timestamp at 0:00, at least three timestamps in order, and segments of at least 10 seconds. Each chapter title is a mini-title that YouTube and Google can match against a query, and Google uses chapters to link straight to a moment in the video.
Tags, captions, and chapters
Tags are minor. YouTube’s own help says they matter mainly when a topic is commonly misspelled, so use the primary keyword and a few close variations, then stop. The heavier work is an accurate transcript (upload your own caption file; treat auto-captions as a fallback) and keyworded chapter titles. Captions are an accessibility win and give engines readable text; YouTube has not said how much they weigh in ranking, so treat any ranking lift as a bonus.
For the publishing side (thumbnails, embeds, structured data, retention edits), see our video SEO guide.
Validate, then iterate
Publishing starts the data. The work continues from there.
- Check the funnel in order: impressions, then impression CTR, then average view duration, then relative retention. Low CTR is a title or thumbnail problem; low retention is a content or intro problem. Fix the one the numbers point at.
- Re-read the search terms 7-14 days after publishing to catch queries you’re ranking for by accident; those are your next videos.
- Revisit metadata on evergreen videos when demand shifts. Updating a title or chapter list to match how people now phrase the query costs far less than re-filming.
This loop compounds. Cross-platform, the same intent research feeds short-form video too, and the conversational, question-shaped phrasing that wins on video overlaps with voice search SEO and featured snippets. One demand map, several surfaces.
When YouTube keyword research becomes a business problem
A creator picking next week’s video does not need an agency: autosuggest, the search terms report and an afternoon cover it. If the channel itself is the business, hire a YouTube specialist. We don’t produce videos or run channels.
It becomes a search problem when video is one part of a company’s organic plan. The symptoms: the team films topics nobody searches for, a video and a blog post chase the same query, the how-to searches your buyers type show a video block your brand never appears in, and nobody can say which topics deserve a video, a page or both.
On a fractional SEO engagement, the two founders build one demand map across Google and YouTube, then assign each query the format its results page rewards, so production budget goes where search demand is.
Frequently Asked Questions
What does YouTube keyword research involve?
Five steps: collect candidate phrases (autosuggest, your channel’s search terms, competitor titles, a keyword tool set to YouTube), check whether each phrase belongs on YouTube or on a web page, score the survivors on demand, competition, intent and watchability, map the winners to the title, description and chapters, and re-read the search terms after publishing to find the next topics.
Is YouTube keyword research different from Google keyword research?
Yes. YouTube ranks videos on relevance, engagement (such as watch time for the query) and quality, while Google ranks pages on relevance and signals like links. A query can thrive on one engine and flop on the other, so always test whether a keyword justifies a watchable video before producing it.
Can you do YouTube keyword research in Ahrefs?
Yes. In Ahrefs Keywords Explorer, switch the search engine from Google to YouTube to see estimated YouTube search volume and related phrases. Like every third-party tool, its YouTube numbers are estimates, so use them to compare phrases against each other and confirm with autosuggest and your own search terms report.
What are the best free tools for YouTube keyword research?
Start with YouTube autosuggest (read the dropdown), the YouTube search traffic source in YouTube Studio Analytics, and competitor titles and chapters for your seed phrase. Google’s “People also ask” and related searches add web-search overlap. These free signals reflect real demand before you ever pay for a tool.
How many keywords should one video target?
One primary keyword plus three to five tightly related secondary phrases. A focused video satisfies intent better than a broad one, which protects retention. Spread the extra phrases across the description, chapters, and transcript, and keep the title to the primary one.
Do YouTube tags still matter for ranking?
Barely. YouTube says tags play a minimal role and help mainly with commonly misspelled topics. The title, the first lines of the description, keyworded chapters and how viewers respond to the video matter far more. Add the primary keyword and a few variations as tags, then put your effort into the rest.
How do Google video results affect YouTube keyword research?
Google shows video on most results pages and can link straight to a chapter inside a YouTube video. That rewards videos with query-matched chapter titles and accurate captions, which is exactly what disciplined keyword research produces. Map demand across both YouTube Search and Google’s video results so one piece earns impressions in both places.