Most channels guess their next video. The channels that win treat YouTube as a data set and crawl it systematically. This is the depth-first research method Jake Trinder uses across the client channels ViewsToClients runs — written up so you can run it yourself.
YouTube is an information game: the channel that has seen the most of what's proven to work makes the best decisions on what to make next. So instead of brainstorming, you crawl: list niches whose viewers behave like yours, generate seed search terms for each, then explore every search result and every promising channel the way a web crawler explores the internet — exhaustively, one branch at a time, noting every outlier (a video that beat its own channel's baseline by a statistically meaningful margin). Patterns across those outliers become your video ideas — each one backed by evidence before you film a second.
The video idea — expressed through the title and thumbnail — decides how much of your market's attention you capture on a feed of twenty competing videos. Editing polish can't rescue an idea nobody wants to click. And you can't reliably judge what a good idea looks like from inside your own niche: you need visibility over what's actually working across the platform right now. That visibility comes from research, not intuition.
Adjacent niches aren't just similar topics. They're niches whose viewers share demographics, psychographics and viewing behaviour with yours — same intent behind the view, same things that make them click. A format proven there will usually transfer to your niche; if nobody in your niche has used it yet, you get first-mover advantage with evidence.
Ask an AI assistant to build the list: describe your niche and your viewer, and ask for ten-plus niches whose viewers behave the same way. Refine until additions stop being relevant. Order the list by relevance — you'll work through it top-down.
For each niche, build 10–20 search terms that between them cover the whole territory ("personal brand", "how to grow on LinkedIn", "Instagram growth strategy"…). These are the entry points of your crawl. Order them by how much demand you'd expect behind each.
Search the first term. Scan results for videos that performed outstandingly relative to their own channel — not raw view counts. A 120k-view video on a channel that normally does 25k is a signal; a 500k video on a 5M-subscriber channel is Tuesday. When you find one, open the channel in a new tab and explore it fully too — a creator with one outlier usually has more, because they understand the game. When a branch is exhausted, step back one level and continue. That's the whole algorithm: go deep, retrace, repeat, until the term — and eventually the niche — is fully explored.
Log every verified outlier in a spreadsheet: title, link, views, outlier multiple, niche. The database is the asset — patterns only become visible across dozens of entries.
After enough entries, formats start repeating across niches: the "every X explained in N minutes" format, the "give me N minutes and I'll…" format, the full-course format, the challenge format. When a format has worked for viewers who behave like yours but hasn't appeared in your niche, that's your shortlist. One example from the video: a speedrun-format video spotted in the dropshipping niche became the launch video of a brand-new client channel in a completely different market — and that channel's first videos each cleared six figures of views, on the back of an idea that was found, not invented.
Outlier multiples are the backbone of the whole process. Our free tools compute them against each channel's own like-for-like baseline (era-aware since YouTube's August 2026 view-count change): the Outlier Finder scans any niche for statistical outliers, and the Lens extension shows the multiplier under every thumbnail while you browse — so the crawl happens as you watch.
🔍 Open the free Outlier Finder 📊 Get VideoCore Lens (free)Want the whole system run for you? ViewsToClients runs YouTube channels for brands and businesses →
Based on "Use This Algorithm, It'll Blow Up Your YouTube Channel" by Jake Trinder, co-founder of ViewsToClients. Published 29 August 2026.