YouTube is the second-largest search engine in the world, and for a new channel search is the only traffic source you can meaningfully influence. Suggested feeds reward existing performance, browse features reward existing subscribers, but search rewards relevance — and relevance is something a brand new channel can win with nothing more than good research.
This masterclass covers how YouTube search actually works, how to find keywords a small channel can realistically rank for, and how to build a content calendar where every video reinforces the last.
How YouTube decides what to rank
Three broad factors govern search placement. Relevance is the match between the query and your title, description, tags, transcript and on-screen content. Engagement covers click-through rate, average view duration and how satisfied viewers appear to be. Authority is YouTube's accumulated confidence that your channel produces good content on this specific topic.
A new channel has no authority and no engagement history. What it does have is complete control over relevance. That is why keyword selection is disproportionately important early on: you are competing on the one dimension where you are not automatically behind.
Step one: build a seed list
Write down ten to fifteen broad phrases that describe your niche. If you cover budget filmmaking that might include "cheap camera", "lighting setup", "gimbal", "colour grading" and so on. These are deliberately generic — they are inputs, not targets.
Feed each into the SmartLabsAI keyword research tool. Each seed expands into dozens of real queries pulled from YouTube's live autocomplete, meaning every suggestion reflects language actual viewers are typing right now rather than a synthetic permutation of words.
Step two: filter for winnable keywords
Now the crucial part. Sort your expanded list by competition, not by volume. For a channel under ten thousand subscribers, the target zone is medium interest paired with low competition. A phrase with two hundred monthly searches you can rank first for will out-earn a phrase with twenty thousand searches where you land on page four and are never seen.
Sanity-check every shortlisted keyword by searching it on YouTube directly. Look at the top five results and ask three questions. How large are those channels? How old are the videos? Do they actually answer the query well? If the top results are two years old, from mid-sized channels, and only partially address the question, you have found an opening.
Step three: mine the question queries
Queries beginning with how, what, why, can and should are the richest vein in any keyword list. They carry explicit intent, they tend to be longer and therefore less contested, and each one is effectively a finished video title. A viewer searching "how to fix grainy footage in low light" knows exactly what they want and will watch until they get it — which is precisely the retention signal YouTube rewards.
Long-tail question keywords are the single best starting point for a channel with no authority. They are also easiest to satisfy completely, which produces the high retention that eventually buys you the authority to compete for harder terms.
Step four: cluster into topic groups
Do not treat keywords as isolated video ideas. Group them into clusters of five to eight closely related terms and cover the cluster systematically. Publishing several strong videos on adjacent aspects of one topic tells YouTube your channel is an authority on that subject, and the videos begin to appear in each other's suggested feeds — which is free, compounding distribution.
Concretely: rather than one video on lighting followed by one on audio and one on editing, publish four consecutive videos on lighting for small rooms, lighting on a budget, fixing bad lighting in post, and lighting mistakes beginners make. That cluster will outperform four unrelated videos every time.
Step five: place the keyword correctly
Your primary keyword belongs in five places. The title, ideally near the front. The first two sentences of the description. Somewhere in the first thirty seconds of spoken dialogue, because YouTube reads your transcript. Two or three of your tags. And the video file name before you upload it, which is a small signal but a free one.
Everything else should read naturally. Keyword stuffing was defeated as a tactic years ago and now actively hurts, because a description crammed with repeated phrases is a poor description for the humans deciding whether to click.
Step six: measure and iterate
Record every target keyword in a spreadsheet with the video that targets it. Once a week, check positions with the rank tracker from a neutral, logged-out session — the results you see while signed into your own account are personalised and misleadingly good.
After four weeks the picture becomes readable. Videos climbing steadily should get promotional support: pin a comment, link them from newer uploads, cut a short from the best segment. Videos stuck outside the top twenty are usually targeting keywords beyond your current authority — move down to a longer-tail variant rather than fighting a losing battle.
Common mistakes
- Chasing high-volume head terms with a small channel and ranking for nothing.
- Targeting the same keyword with several videos, so they compete against each other.
- Doing research once and never revisiting it as search language shifts.
- Ignoring the transcript, which YouTube reads and most creators never optimise.
- Judging results after five days instead of the four weeks rankings actually take to settle.
A realistic first ninety days
Month one: research fifteen seeds, build a list of forty candidate keywords, and publish four videos targeting the lowest-competition question queries you found. Month two: track weekly, identify which cluster performed best, and publish four more inside that cluster. Month three: refresh the titles and descriptions of month one's videos using what you learned, and begin targeting slightly more competitive terms in your strongest topic.
This is unglamorous work and it is exactly what separates channels that compound from channels that plateau. Ten minutes of research before each video, thirty seconds of rank checking each week, and a willingness to go after small keywords first. Every tool you need for it is free on SmartLabsAI.