Keyword Research for YouTube

I start with a broad seed topic related to my niche. For this example, say my channel is about video editing. The seed is "how to edit video in Premiere Pro". I open YouTube and type the seed into the search bar. YouTube shows autocomplete suggestions: "how to edit video in premiere pro for beginners", "how to edit video in premiere pro faster", "how to edit video in premiere pro 2025" and so on. I write down the first 10 suggestions. Then I use the alphabet trick: I type the seed followed by " a", " b", " c", and note each long-tail phrase that appears. This gives me another 20 [long-tail](/long-tail-keywords/) variations. I now have a list of around 30 potential [keywords](/keywords-youtube/). I remove duplicates. Next, I validate demand. I do not look at Google search volume — that is a mistake I made before. Instead, I search each phrase on YouTube and look at the top 5 results. I note the title, the view count, the channel subscriber count, and the upload date. For the phrase "speed ramp premiere pro", the top result has 8,000 views, channel has 2,000 subscribers. That tells me the competition is low. For "effects premiere pro", top result has 200,000 views and a channel with 500,000 subscribers. I skip that. I also check the suggested videos next to the search results — those are often related queries that YouTube thinks users search next. I add those to my list if they are relevant. I then use Google Trends set to YouTube Search to check if the interest over time is stable or declining. I prefer keywords that are trending up or flat. Seasonal keywords are fine if I publish at the right time. Finally, I pick one primary keyword per video. For this example, I choose "speed ramp premiere pro" because the [keyword difficulty](/keyword-difficulty/) is low — the top video has only 8k views. My decision rule: if the top 3 results have channels with under 50,000 subscribers, I go ahead. If any has over 500,000, I move on. I then write the title with the exact keyword at the beginning. This path has consistently got me into the top 5 for new channels.

The biggest trap on YouTube keyword research is chasing high-volume keywords without checking the actual ranking potential. I call this volume blindness. You see 50k monthly searches in a tool and think it is a goldmine. But when you search the phrase on YouTube, the first page is dominated by channels with millions of subscribers who have been uploading for years. Your 100-subscriber video has almost zero chance of appearing there, even with perfect optimisation. Instead, I look for keywords where the top result has under 20,000 views and the channel is small. Real example: "premiere pro colour grading tutorial for beginners" — top video 15k views, channel 8k subs. That is achievable. Another watch-out: using many loosely related keywords in a single video. YouTube's algorithm uses the video's topic graph, not just the tag field. If you target "premiere pro tutorial" and "after effects tutorial" in the same video, the algorithm cannot decide what the video is about, so it recommends it for neither. I now stick to one primary keyword and maybe two closely related long-tail phrases. I also avoid stuffing keywords into the description. YouTube reads the first two lines of the description heavily, so I put the primary keyword there naturally, not a list. A third watch-out: ignoring [search intent](/intent-seo/). A video titled "What is colour grading?" has a different intent than "How to colour grade in Premiere Pro" — one is informative, one is tutorial. If your video is a tutorial but the search results for that keyword are all explainer videos, you will get low CTR because users expect something else. I check the top 5 results and ask: is the format the same as mine? If not, I adjust the keyword or the video style. Finally, be careful with Google Trends. If you do not switch the filter to YouTube Search, you are looking at web volume which is often misleading for video content. I have an edge case: a keyword "video editing tips" shows high web volume but on YouTube it is saturated. I now always check YouTube-specific data.

I got keyword stuffing wrong for the first 40 videos. I believed that more keywords in tags and description would increase my chances of ranking for any of them. I wrote descriptions like "In this video editing tutorial for beginners, we cover video editing tips and tricks for Adobe Premiere Pro. Learn how to edit videos, how to keyframe, how to add effects..." It was a mess. After 6 months, none of those videos ranked for anything meaningful. I changed my approach completely. Now I write the title first with the exact primary keyword, then the description briefly explains what the video covers, using the primary keyword once naturally and maybe one related phrase. I limit tags to the primary keyword, a branded variation, and one long-tail phrase. That is it. The results improved immediately. Another thing I got wrong: ignoring the thumbnail as a ranking factor. I used to think thumbnails only affect CTR, and CTR is not a direct ranking signal. But YouTube's algorithm uses CTR as a proxy for satisfaction — if people click and then watch, the video gets a boost. So a bad thumbnail means even if you rank, no one clicks, and you drop. I now spend as much time on the thumbnail as on the [keyword ranking](/keyword-ranking/) research. I also misunderstood the role of tags. I thought tags were the main way YouTube understood a video. Actually, YouTube transcribes the audio and reads the description. Tags have little weight. I stopped stuffing and started focusing on the spoken content aligning with the keyword. Finally, I used to treat all keywords the same. I did not differentiate between informational, commercial, and navigational intent. I made a tutorial for a keyword that was clearly navigational — people searching for "Adobe Premiere Pro download" want the software, not a tutorial. My video had high impressions but zero clicks because the search result showed the download page. Now I analyse the SERP to understand what users expect before producing a video. That one change saved weeks of wasted production.