Somebody handed you a YouTube search volume number and you have already made a decision with it. It came out of a tool, it had two significant figures, and nobody in the room could tell you where it originated. That is not an attack on the tool. Google has never published search volume for YouTube, and Phil Nottingham said it plainly on Moz's Whiteboard Friday in June 2026: there is no way of actually knowing the search volume on YouTube, and the products that claim to know are approximating it from other signals. Which raises a question you can answer for yourself: if the number under your title is an estimate, what are you actually optimising for? The method that does not depend on it costs nothing, and most of it you can run this afternoon.
Here is what the guides get wrong before they reach any tool. They treat YouTube search as Google with videos attached, so they hand you Google keyword volumes and tell you to build a title around the biggest one. YouTube does not rank the way Google ranks, and it does not search the way Google searches. It runs its own index, its own ranking weights, and a watching session that decides what to show next instead of a query that decides what to show now. That difference quietly moves your target. The video that would have won a Google query can lose on YouTube to a shorter one that held viewers three seconds longer in the first minute. So which of those two outcomes are you optimising for right now?
How YouTube Search Actually Works
YouTube handles more than 3 billion searches a month, according to Tubular data compiled by Searchlab for 2026, and that headline hides the more useful number underneath it: roughly 70% of watch time arrives through the recommendation system rather than through search. Both paths matter, but they reward different things, and a keyword strategy copied from Google optimises for the wrong one.
Google answers a query. YouTube fills a session. A Google search ends when the user has the answer. A YouTube search ends when the user stops watching, which is usually several videos later. That structural difference is why retention outweighs relevance in close calls, and why an optimised title is worth more than a stuffed description. Searchlab's 2026 compilation puts the ranking weights roughly in this order: watch time very high, click-through rate high, engagement high, relevance from title, description and tags medium to high, then upload frequency and thumbnail quality.
Relevance still gates everything. If your title and description do not match the query, no amount of retention gets you shown. But relevance is the entry ticket, not the score. Two videos can both be relevant to "how to edit a podcast", and the one that holds viewers past the first thirty seconds takes the session.
One more mechanic matters when you pick keywords. YouTube's autocomplete is not reading a query log the way Google's does. It is shaped by what people search and then watch, and partly by what the platform wants to surface. That makes an autocomplete phrase two signals at once: demand, and a hint about whether there is already inventory the platform likes. A suggestion with thin, dated results behind it is a gap. A suggestion with three multi-million-view videos from this year is a wall.
A Five-Step Method for YouTube Keyword Search
The method below works with no subscription. Steps two through five need nothing but a browser and your own channel data.
1. Start from topics, not keywords. Write down five topics your viewer would search for if they already knew the category. Not "podcast editing software", but "my podcast audio sounds bad". YouTube search skews toward the problem the viewer can name, and problem phrasing is where the long tail actually lives. For scale on that skew, consider the most-searched term on YouTube in the US: "asmr", at roughly 38 million monthly searches per Semrush's YouTube data, against the 10 to 100 searches a month that new channels are usually told to target. The gap between those two numbers is the entire argument for building on tighter phrases instead of chasing head terms.
2. Read autocomplete properly. Type your seed into the YouTube search bar and stop. Do not press enter. Now type the seed followed by each letter of the alphabet, one at a time, and record every suggestion. Then use the underscore wildcard: your seed_ forces YouTube to fill in the words that follow, and _your seed fills in what precedes it. This is the highest-yield free technique in the list, and it is what the strongest competing guides lead with for good reason.
3. Mine competitor channels for outliers, not averages. Open three channels that already rank for your topic and sort their videos by popularity. You are looking for the mismatch: a video with far more views than that channel's typical video gets. The mismatch is the signal, because it means the topic travelled beyond the existing subscribers. That is the closest thing to demand data you can get without a tool, and it works on any channel size.
4. Search inside YouTube, then check your own gaps. Run the phrase in YouTube search and read the results the way a viewer would. Note the age of the top results, how many views they carry, and whether the titles all say the same thing. Then open YouTube Studio, go to Research, and read the two tabs available to you: what your own viewers search for, and the content gaps YouTube reports for your channel. If you have never opened that tab, it is the most underused free research surface on the platform.
5. Validate with numbers you can compute yourself. This is the step that replaces a fake volume number. Take the top 100 videos for your phrase, then calculate three things: the median monthly view velocity (total views divided by the video's age in months, which is the demand signal), the median subscriber count of the channels behind those videos (the competition signal), and the median age of the videos (the decay signal). Moz's Phil Nottingham proposed exactly this set in June 2026 as a substitute for search volume, and it holds up. If the median velocity is healthy, the median channel is small and the median age is measured in years, you have found something worth making. Nobody can sell you this number, and nobody can lie to you about it.
Free Methods Worth Doing Before You Pay for Anything
Three more free surfaces belong in the same session, and one warning belongs with them.
Google Trends with the YouTube Search property. Set the property to YouTube Search, not web search, and compare two or three candidate phrases. Trends will not give you volume, but it will show you direction and seasonality, which is what a keyword decision actually needs most of the time.
Filters inside YouTube search. Sort by upload date to see what the platform is currently rewarding for that phrase, and sort by view count to see what it has historically rewarded. If the recent results look nothing like the all-time results, the topic has shifted and the older videos are riding authority rather than fit.
Transcripts of the top three videos. Read them. This is competitor analysis that no keyword tool performs, and it tells you whether the top results are answering the question or circling it. The gap you find there is your script outline.
The warning: any "free YouTube keyword research tool" that returns a volume number is returning an estimate, and the number to distrust most is the one with three significant figures. Ask what the estimate is derived from. Most tools will not say, which is itself the answer.
YouTube Keyword Research Tools Compared
Tool guides tend to describe features. What you need to know is which tool solves which problem, and what each one cannot do.
Tool | Best for | Honest limitation |
|---|---|---|
vidIQ | Channel-level scoring, competitor tracking, title suggestions in the browser | Scores are proprietary, so treat them as one input, not a verdict. Volume figures are estimates. |
TubeBuddy | Keyword explorer plus A/B title testing on your own channel | Heavier on channel management than pure research. Paid tiers gate the more useful research features. |
Keyword Tool (keywordtool.io) | Fast YouTube autocomplete export at scale | Autocomplete aggregation, not measured search demand. Depth behind a subscription. |
Ahrefs YouTube Keyword Tool | Free browser-based keyword and autocomplete lookup with traffic estimates | The real value sits behind a paid Ahrefs plan, starting around $129 a month in the current pricing and reaching $449 a month. Overkill if YouTube is your only channel. |
Google Trends and YouTube Studio | Direction, seasonality, and your own channel's content gaps | No volume, no competition score, and Studio data only covers your own channel. |
Pricing on the three YouTube-native tools is deliberately absent above. None of them publishes a durable rate card, and the roundups ranking for this term quote the same product at two different monthly prices in the same article. If you need a current figure, take it from the vendor's own checkout page on the day you buy, because that is the only version that will still be true next quarter.
There is a separate set of queries that signal buying intent and deserve their own treatment: "youtube keyword planner" is one of them. It is searched by people who have already decided to pay for a tool and are looking for the closest thing to Google's Keyword Planner for video. There is no official YouTube equivalent. The honest answer to that search is that the closest you get is a third-party explorer, that all of them estimate demand, and that the validation step above beats all of them for a channel below a few thousand subscribers.
Where tools do earn their money: tracking your competitors' upload patterns over months, testing titles on your own videos, and saving you the manual work of step two. Pay for one of those, not for a volume column.
What AI Search Changes About YouTube Discovery
YouTube is no longer only a place people go to watch. It is a source that AI answers cite, and that changes the value of the transcript.
The numbers are worth reading carefully. YouTube overtook Reddit as the most-cited social platform in AI-generated answers in early 2026, driving 38.1% of all social citations across Google AI Overviews, ChatGPT and Perplexity, per an OtterlyAI study reported by Adweek in January 2026. The platform split is uneven and worth knowing: 38.7% on Perplexity, 36.6% on Google AI Overviews, and 4.4% on ChatGPT. A separate SE Ranking study of 50,807 health queries in January 2026 found AI Overviews appearing on more than 82% of them, with YouTube supplying 20,621 of 465,823 citations, or 4.43%, ahead of every medical institution in the set.
Why YouTube specifically? Because Google owns the platform. AI Overviews treats it as a first-party source with unrestricted access to the full transcript and metadata of every video. That is a structural advantage no other social platform can copy, and it is the reason transcript quality now matters outside YouTube search.
Practical consequence for your keyword work: the phrases you research are now also the phrases an AI answer may pull your video's spoken content from. A title that matches a real question plus a first sixty seconds that states the answer in plain sentences is doing keyword research and citation work at the same time. This is the same reasoning that sits behind AI search monitoring, applied to video instead of web pages.
Where Allable Fits, and Where It Does Not
Let me be direct about the second half of that heading, because most tool articles in this space will not be. Allable is not a YouTube tool. There is no YouTube keyword feature, no video analytics, and nothing in the product that reads your channel. If a comparison table in a roundup places us next to vidIQ, that table is wrong.
Where it does fit is the layer either side of the video. A YouTube keyword pass and a web keyword pass on the same topic are usually done twice by two people, with two different outcomes, and the two assets then compete for the same search demand instead of supporting each other. One research pass fixes that. The topic gets validated once, by volume, difficulty and intent, clustered against what you already publish, checked for gaps against competitors, and then turned into a brief. That brief feeds the video script notes, and the same topic becomes the companion article that the video sits above. Our keyword research tool is built for that web and content side, which is the only side we can honestly claim, and the methodology overlap with what you just read is the point rather than an accident. If you want to see how the web-side research is assessed properly, Ahrefs keyword research is the cluster pillar I would start with.
The claim I will defend: the video and the article should share one research file. The one I will not make: that any of this requires a YouTube integration.
Bottom line
Do the free version first, and do it properly. Autocomplete plus competitor outliers plus computed view velocity will tell you more about real YouTube demand than any volume column, and it will cost you an afternoon instead of a subscription. Pay for a tool only when it does something you cannot: tracking rival upload patterns over time, or testing titles on your own catalogue. Then write the topic down once and use it twice, in the script and in the article, so the two assets reinforce each other instead of splitting the same audience. The number you were shown was never the point. What your viewer typed before they found you is.
Frequently Asked Questions
How do I find YouTube keywords without paying for a tool?
Use autocomplete with the alphabet and the underscore wildcard, mine competitor channels for outlier videos, read the Research tab in YouTube Studio, and validate candidates by calculating median view velocity, median subscriber count and median video age across the top 100 results. That combination produces better decisions than a paid volume column for channels below a few thousand subscribers.
Is there a free YouTube keyword generator that shows real volume?
No. Google does not publish YouTube search volume, so every generator producing a volume figure is estimating it from other signals. Free tools are genuinely useful for autocomplete aggregation, which is a real demand signal. The moment one of them prints a precise number, you are looking at a model output, not a measurement.
What is the best YouTube keyword analysis tool?
It depends on which job you want done. vidIQ and TubeBuddy are strongest for channel-level tracking and title testing. Keyword Tool is fastest for bulk autocomplete. Ahrefs runs a free browser-based YouTube Keyword Tool, with deeper features on paid plans starting around $129 a month. If you want one tool and no subscription, YouTube's own autocomplete plus Studio Research covers more ground than most people expect.
How do I check YouTube keyword volume?
You cannot check it directly, and that is the honest answer. What you can check is whether demand exists: pull the top 100 videos for the phrase and divide each video's total views by its age in months to get view velocity. Reliable views and reliable age are both public. That median tells you more about real demand than any third-party volume estimate, and you can reproduce it in a spreadsheet.
Does YouTube keyword research still matter now that AI answers dominate search?
More than it did, for a different reason. YouTube became the most-cited social platform in AI-generated answers in early 2026, at 38.1% of social citations, partly because Google owns the platform and can read every transcript. The phrases you research now also determine what an AI answer may quote from your video, so research that targets real questions is doing double duty.
How is YouTube keyword research different from Google keyword research?
Three ways that change your decisions. YouTube ranks on watch time and click-through more than relevance, so retention is the tiebreaker. Autocomplete reflects what people watch after they search, not only what they type. And YouTube is a session platform, where one search leads to several videos, which makes the first thirty seconds of your video a ranking factor in a way a web page's first paragraph never is.
Get the research done once
Bring the topic into Allable, run the keyword research on the web side, and get the brief, the companion article and the script notes from the same source. The trial starts when you connect.
