The short version: YouTube keyword research isn’t about finding the biggest search volume, it’s about finding the keyword you can rank for this month, with the video length and format the top results already show you work. Start in the search bar, not a tool, then use a tool to sanity-check volume and competition before you film a single frame.
Why most people get this backwards
I’ve watched dozens of small business owners open a keyword tool, sort by highest search volume, and pick the top result. Then they wonder why a video with 90,000 monthly searches gets 340 views after three weeks.
Here’s what nobody tells you clearly enough: search volume is the least useful number on the screen. YouTube isn’t Google. It doesn’t just match your title to a search term, it watches what happens after someone clicks. If your video doesn’t hold attention for longer than the other results ranking for that term, YouTube stops showing it, no matter how well you keyword-stuffed the title. I’ve seen channels with 4,000 subscribers outrank channels with 400,000 because their video kept people watching for eleven minutes instead of two.
So the real question isn’t “what has volume?” It’s “what can I make that beats the current top five results on watch time, for a term small enough that I’m not fighting a channel with a production team?”
Start where the searches happen: the search bar itself
Open YouTube, not Google. Type your broad topic and stop. The autocomplete suggestions are real searches from real people, pulled straight from YouTube’s own database, and they update constantly.
- Type the seed phrase (say “email marketing”) and note every autocomplete suggestion.
- Add a letter after it (“email marketing f”, “email marketing h”) to force different suggestions.
- Search the term, then scroll to “People also watched” and the related searches YouTube shows partway down results and again at the bottom.
Do this for fifteen minutes and you’ll have 30 to 50 real phrases with zero guessing involved. This is the single most underused step in the whole process, and it’s free.
Then check volume and competition
Once you have a list, I run it through TubeBuddy or VidIQ (both have free tiers) to see two numbers side by side: search volume and competition score. Neither tool is perfectly accurate, YouTube doesn’t publish real numbers publicly, but they’re directionally right, which is all you need.
The rule I use with clients: look for a competition score under 40 out of 100 and a search volume you can see actual demand behind, even if that’s only a few hundred searches a month. A keyword with 300 searches and weak competition will outperform a keyword with 30,000 searches and channels like Ahrefs or Neil Patel already sitting on page one.
Google Trends is worth a quick check too, not for volume but for direction. If a topic is sliding downward over 12 months, I deprioritise it even if the raw numbers still look decent, because by the time you film, edit, and publish, that window may have closed.
A real example from a client channel
A client of mine runs a home organising channel, around 6,000 subscribers at the time. She wanted to target “how to declutter your kitchen”, 40,000 monthly searches according to the tool, and every result on page one was a channel with 200,000-plus subscribers and slick editing. We pulled it and went for “declutter a small kitchen with no pantry” instead, roughly 700 searches a month, competition score of 22.
Her video hit page one within nine days. Average view duration was 7 minutes 40 seconds on an 11-minute video, which is strong retention. That single video now brings in around 1,800 views a month, eighteen months later, and it’s still her third highest traffic source to the channel. The broad keyword would have buried her under channels with budgets she doesn’t have. The narrow one gave her a fight she could win.
Look at who’s already ranking, honestly
Before you commit to a keyword, search it and watch the top three results, not just skim titles. Ask yourself:
- How many subscribers do these channels have?
- How old are the ranking videos? Anything published in the last 60 days and still ranking well tells you the topic has active demand.
- What’s missing? Did they skip a step, use bad audio, film in poor light, or ignore a sub-question people are clearly asking in the comments?
Read the comments on the top two videos. This is where you find your actual angle. People type the exact frustration they came with, and half the time it’s a question the video never answered. That gap is your title.
Match the keyword to the right format and length
Every keyword implies a length, and ignoring that is a fast way to waste a week of filming. If the top ten results for a term are all under 60 seconds, YouTube is telling you this is a Shorts intent, not a long-form one. Chasing that same phrase with a 14-minute video usually flops, because you’re competing against a format YouTube isn’t matching it to.
If you’re working across both long-form and Shorts, it’s worth reading how hashtags work differently for YouTube Shorts before you plan the keyword, because the discovery mechanics aren’t the same as long-form search.
For long-form, check the average length of the top five ranking videos. If they’re all 8 to 12 minutes, don’t film a 3-minute video on the same term, you’ll get outranked by videos YouTube has already decided satisfy that search better through longer watch sessions.
Build a keyword cluster, not one keyword
The mistake I made early on with my own channel was treating each video as a one-off keyword bet. What works better is building five to eight videos around one topic cluster, so YouTube starts linking them together in suggested videos and your channel becomes the obvious authority on that specific subtopic.
For example, instead of one video on “LinkedIn lead generation”, build out:
- LinkedIn lead generation for consultants
- LinkedIn lead generation without paid ads
- LinkedIn lead generation message templates
- LinkedIn lead generation mistakes to avoid
Each one is narrow enough to rank on its own, and together they build a session chain where someone watches one, gets suggested the next, and your total watch time on the channel climbs. This is the actual mechanism behind channels that seem to “suddenly” grow after months of flat numbers. It isn’t one viral video, it’s a cluster that finally gave YouTube enough signal to trust the channel with a topic.
Put the keyword where it counts
Once you’ve settled on a keyword, use it in these four places, in this order of importance:
- The first 60 characters of the title, exact match if it reads naturally
- The video file name before you upload (yes, YouTube does read this)
- The first two sentences of the description, written for a human, not stuffed
- Spoken naturally in the first 15 seconds of the video itself, because YouTube’s speech-to-text transcription is part of how it understands relevance
Tags matter far less than they did in 2015. I still add them, mostly as a backup for misspellings and close variants, but I’ve stopped treating them as a ranking lever worth agonising over.
If you’re new to search generally and learning as you go, it’s worth knowing that a lot of what you’ll find online conflates YouTube SEO with website SEO, and they overlap less than people assume. This piece on whether video SEO deserves separate attention from your written content is a useful gut check before you sink a week into either one.
The uncomfortable part nobody wants to say out loud
Here’s the bit that annoys people: most small channels are targeting keywords that are already won. Not competitive, won. There’s a difference. Competitive means you could realistically break through with a better video. Won means a channel with a decade of authority, a subscriber base in the hundreds of thousands, and a video that’s been sitting at position one for two years with 40,000 comments is not moving for your upload, no matter how good it is.
I say this because I’ve watched clients pour three weeks into a single video chasing a keyword that was, frankly, never available to them at their current channel size. It’s not a lack of effort or talent. It’s picking a fight you can’t win with the resources you have right now. The honest fix isn’t a better thumbnail, it’s picking a smaller keyword, winning it, growing the channel’s authority, and coming back for the bigger term in eight or twelve months once you have the subscriber base and watch-time history to compete for it.
How often to redo this
I redo keyword research for a channel every quarter, not every video. Search behaviour shifts, new terms emerge, and a keyword that was tight competition in January can be wide open by June if a couple of the previous top-ranking channels have gone quiet. I keep a simple spreadsheet: keyword, search volume, competition score, date checked, video published, current rank. It takes ten minutes to update and saves hours of guessing later.
If keyword research feels like one more thing on a pile that already includes content planning, filming, and everything else that comes under what falls under content marketing for a small business, you’re not imagining it. It’s a real time cost, which is exactly why doing it (checking real search terms and real competition) up front saves you from filming videos nobody was searching for in the first place.
And if you’re teaching yourself all of this by watching creators explain their own process, it’s worth reading how to learn SEO from YouTube tutorials without losing months to conflicting advice, because a lot of what gets taught online is out of date within a year of being filmed. YouTube’s own ranking signals shift more than most tutorials admit.
A simple checklist before you hit record
- Pulled 20 to 30 real phrases from YouTube’s own autocomplete and related searches
- Checked search volume and competition score in TubeBuddy or VidIQ
- Watched the top three ranking videos in full, not just the thumbnails
- Read the comments for the exact gap or frustration your video can fill
- Matched your planned length to what’s already ranking for that term
- Confirmed the channels ranking are a size you can realistically challenge
Frequently asked questions
How many searches a month is good enough for a YouTube keyword?
There’s no magic number, but for a channel under 20,000 subscribers, a keyword with 300 to 2,000 monthly searches and low competition will usually outperform a 30,000-search keyword that’s already dominated by large channels.
Should I use the same keyword research process for Shorts as long-form videos?
No. Shorts discovery leans more on hashtags and the swipe feed than on search intent, so check whether the top results for your term are Shorts or long-form before you plan the video, and treat hashtag choice for Shorts as a separate step.
Is TubeBuddy or VidIQ more accurate for YouTube keyword volume?
Neither pulls exact numbers from YouTube because YouTube doesn’t publish them publicly, so both are estimates. Use them for direction and comparison between keywords rather than treating either number as precise.
How long should I wait before deciding a keyword didn’t work?
Give a video 30 days before judging it. YouTube often takes two to four weeks to fully understand where a new video fits, and early view counts can look flat even on videos that end up ranking well long term.
