We're learning YouTube in public.
So you don't have to start from scratch.
Most YouTube advice is either outdated, buried behind a course paywall, or only makes sense after you already have 100K subscribers. We got tired of that. So we started documenting everything we learn — in real time, while building a channel from zero — and putting it here for free.
This is not a finished guide. It's a living document. Every time we watch a creator break down the algorithm, test a new strategy, or discover something that shifts our understanding, we add it here. No fluff. No filler. Just what actually works right now.
If you're building on YouTube too — this is for you.
The YouTube
Playbook
Everything we've learned about the algorithm, thumbnails, titles, retention, and niche strategy — documented as we go. Quick reference format. No fluff.
YouTube Recommendations = your real distribution engine. Not subscribers. Not search. Recommendations.
Source: "I launched my YouTube Channel and nobody watched" — real analytics data
Key insight: Subscriptions = only 90 views. Search = only 18. Don't build for subscribers. Build for the algorithm to push you to strangers.
algorithm recommendations
triggers the push
at new channel stage
YouTube selects your seed audience — based on context signals in your title, thumbnail, description, tags, and hashtags. It finds the first small group whose watch history matches what your video appears to be about. If they engage well, YouTube expands to more like them.
Email lists, subreddits, Facebook groups — these can help get your first views but they are NOT your seed audience. The seed audience is YouTube's internal matching system. You don't choose it. You feed it with metadata.
Title + description first 40 chars + tags + hashtags tell YouTube exactly WHO this is for and WHAT it covers
Finds people whose watch history matches your signals — this is the seed
Good CTR + AVD signals: push this to more people like them
YouTube pushes to larger audiences. Your job is to make the first 12 hours count by having the right metadata from upload.
- Specific title — WHO + WHAT in first 40 characters
- Matching thumbnail — visually confirms the title's promise
- First description line — repeat WHO + WHAT, not a greeting
- Accurate tags + hashtags — niche-specific, not broad categories
- Vague title — "My Money Tips" tells YouTube nothing
- Broad category tags — "finance" or "money" = too wide
- Description starts with greeting — wasted signal space
- Thumbnail doesn't match title — mixed signals break matching
Aim for 30–50% of your video length as Average View Duration. 3:47 on an 11-min video = strong signal.
5%+ CTR = your target. Below 5% = you have a thumbnail problem. Fix the thumbnail before anything else. Half of all channels sit between 2–10% — the difference is whether your packaging matches your audience.
Your video content = 20% of your views. Your title + thumbnail = 80%. Spend 50% of your production time on packaging.
- 3 elements max — face, subject, text. Text counts as one element. No clutter.
- 3–6% text coverage — text fills a tiny fraction of the image. Less than you think.
- 4–9 words on the thumbnail maximum
- Face performs best on finance channels — expressive, not posed
- 2+ faces outperform a single face when it makes sense
- New channels: no-face thumbnails can outperform at the start
- Top colors: cyan, green, orange — highest CTR in testing
- No numbers — thumbnails without numbers outperform those with
Use as accent colors against dark or clean backgrounds. They pop against YouTube's white interface.
Split-screen emotional contrast — one thumbnail format that works
The "find 2 outlier videos and combine their thumbnails" method was effective in 2024–2025. It is now oversaturated. The current approach: find under-saturated topics with high views and low competition — similar to the Golden Keyword Ratio in SEO. Use outlier thinking as inspiration, not a copy formula.
If your CTR is below 5% — you have a thumbnail problem. Stop tweaking your content. Fix the thumbnail first.
Strong topic-audience fit. YouTube knows who you're for. Your packaging is working.
✓ You've found a niche and you're serving it well
Getting some attention but it's not an obvious click. Your packaging doesn't match your audience clearly enough.
→ Niche down. Your packaging is too broad or too vague.
Serious problem. Wrong audience or bad thumbnail. Stop posting, rethink the packaging.
Getting shown, not connecting. Niche down and test a stronger thumbnail.
Solid. Packaging is working. Focus on improving content quality and retention now.
Excellent. Strong niche fit. Your audience knows exactly what they're clicking. Scale this.
Source: YouTube Help. Half of all channels have a CTR between 2% and 10%. The difference between 2% and 10% is the difference between 1 person in 50 and 1 person in 10 clicking. That is a massive gap.
Make the background dark and the words light.
Light backgrounds with light text blend into YouTube's white interface. Dark thumbnails with bright text pop and stop the scroll. Forces visual contrast your eye goes to first.
Always put a WHO and a WHAT in the title. This gives YouTube the context clues it needs to find your exact seed audience fast.
Notice: the right-side titles have a WHAT (the thing) and a WHO (the person it's for). YouTube uses this to match you to the right seed audience immediately.
Find 3 videos on the exact same topic that are already performing — posted within the last year with 100K+ views. Then do this:
Look at the 3 videos side by side. What words do ALL of their titles share? Those repeated words are the keywords the algorithm is using to push them. Write them down. Your title must include those exact words.
Example: "Meal Prep" + "High Protein" + "Easy" + "Healthy" — these appeared across 3 top-performing meal prep videos. All 4 keywords in your title = algorithm signal match.
What does YOUR video cover that they didn't? That specific angle goes into the title too — after the shared keywords. Now you get algorithm ranking from the keyword match AND curiosity clicks from the unique angle.
Keywords they share = how YouTube finds you. Your unique twist = why someone picks your video over theirs.
Keep a running Google Doc. For every video you make, list the 3 reference videos, extract shared keywords, then draft your title to include those keywords + your angle.
- 5–7 words — sweet spot. Longer titles underperform.
- No numbers — "Why you're still broke" > "7 ways to save"
- First 40 characters — your WHO + WHAT must appear here
- Broad category words kill you — "finance" "money" "recipes" too vague
- Be specific, name the actual thing — don't dance around it
"The thing nobody tells you about money" — creates a gap they have to close
"Why you can't stop spending money" — they feel seen and have to click
A made-up name for something that already exists. Proprietary feel = unique curiosity. e.g. "The Automagic Goals Method"
- How-to questions — "How do I make a budget?"
- What is X — "What is a Roth IRA?"
- Generic lists — "5 ways to save money"
- Definitions — "What is debt snowball?"
AI is eating this content whole. You can't compete here as a new channel.
- Why questions — "Why do I keep failing at budgeting?"
- Experience-based — "I managed money alone for 10 years"
- Contrarian takes — "Stop budgeting and do this instead"
- Lived truth — "What divorce actually does to your finances"
AI cannot replicate lived experience. This is your permanent advantage.
Every successful YouTube video has at least 2 of these 3 forces. One strong enough can carry the whole video alone.
Your personality, vibe, presence. People watch you not just the topic.
Em on the Brain — normal girl, relatable energy. Her presence alone pulls people in.
The information or insight is so rare, people forgive everything else.
Daniel Barada — no edits, reads from a doc. 13K views in 1 day. Value alone carried it.
Pacing, cuts, text overlay, b-roll. Helps when the other two aren't carrying it.
Most beginners over-rely on this. Editing can't fix weak value or zero magnetism.
If you have exceptional value — teach things people can't get anywhere else. If you have strong magnetism — show up consistently and be real. Either one can carry you. Both together = momentum.
Thumbnail and title must complement — not repeat each other. The thumbnail raises the question. The title completes it. Together they open the click.
THE LOCK.
THE KEY.
by watching.
You usually don't have an editing problem. You have a pacing and structure problem. More cuts won't fix a slow story.
- Every 60–90 sec — something must change
- No filler sentences — cut anything that doesn't move the story
- Front-load value — give something useful in first 2 min
- Pattern interrupts — text on screen, question, pivot
- Mini-payoff — end each section before the next
Write your video: Problem → Proof → Solution → Action. Tight outline = natural pacing. Editing can't fix a weak structure.
Optimal video length differs by niche. Don't copy gaming or entertainment durations for a finance channel.
for short-form views
for long-form watch time
highest AVD overall
When someone watches your video AND then keeps watching more YouTube after — that's session time. YouTube rewards creators who extend sessions by pushing their videos harder.
your video
videos waiting
extends
you further
- Build a catalog — the more related videos you have, the longer sessions run
- End cards + playlists — point viewers to more of your content, not away from YouTube
- Consistent niche — if all your videos serve the same audience, sessions compound naturally
- Launch with 2 videos on day 1 — gives YouTube more to push and extends opening sessions
Pick one niche. Study the top creators. Score them on the 5 skills. Find their gaps — usually 3/10 or 5/10 on thumbnails, titles, or storytelling. Outrank those gaps. Repeat.
Not "personal finance" — "money system for financially solo women in their 30s"
Score each on thumbnails, titles, editing, storytelling, cinematography (1–10)
Like the Golden Keyword Ratio in SEO — fewer videos on topic + high views = opportunity. Not the same as the outdated outlier hack.
If the top creator in your niche has a 4/10 thumbnail — make a 9/10 thumbnail on the same topic
Keep It Sweet Kitchen: sourdough felt impossible → simplified it to flour, water, salt. Monetized in 11 days.
She went 0 → 96.5K in 10 months. 8K email subscribers off 2 videos. Same niche. Same audience. Consistent.
- 1Title drafted — 5–7 words, WHO + WHAT in first 40 chars
- 2Title type confirmed — curiosity gap, mind-read, or mechanism
- 3Thumbnail concept — 3 elements max, face + minimal text
- 4Hook written — pain named in first sentence
- 5Structure mapped — Problem → Proof → Solution → Action
- 6Real number or specific personal story confirmed
- 7Target length: 14–20 min for finance. 1–2 min for Shorts.
- 1Thumbnail — 3 elements only, 3–6% text, face expressive
- 2Thumbnail background — dark background, bright/light text
- 3CTR self-check — is this thumbnail a 5%+ click or a 2% click?
- 4Title — complements thumbnail, does NOT repeat it
- 5No numbers in title — emotional framing wins
- 6Description line 1 — WHO + WHAT, not a greeting
- 7Tags + hashtags — niche-specific, not broad categories
- 8First 30 sec reviewed — would YOU keep watching?
- 9End card set — point to more of your own content
- 10Pin a comment after publishing — tell viewers what to expect next, give them a reason to subscribe
Pain points that pull in strangers. Titles that work without knowing your channel at all.
Deeper dives, story updates, system walkthroughs. Rewards people who already trust you.