Two jobs in one place: a research vault of everything we've learned, and a working tool to pick niches, score opportunities, and manage launches by the numbers. Start with the nine truths, then use the Plan & Operate tools to make decisions.
The whole strategy in 9 truths
Everything in the Research Hub distills to these. Use them as the filter for every decision.
| 1 | It's the system, not the niche or "being faceless." Repeatable execution + testing taste at scale is the moat. → Builders, Pulse |
| 2 | AI is a tool, not the creator. Human edit/voice/variation survives policy; templated slop gets demonetized. → Rules, AI Stack |
| 3 | Money = high-RPM evergreen + Tier-1 geo, or low-comp non-English geo. Finance, history, crime, ASMR/sleep. → Niches, Creative Matrix |
| 4 | Package first. Decide title + thumbnail before you build; the algorithm ranks CTR × retention. → Packaging |
| 5 | Build the money-video flywheel from Day 1 (email + high-ticket). AdSense is the floor, not the ceiling. → Builders |
| 6 | Validate with data, not vibes. Mine outliers, ship ~15 test videos / 2–3 months, then pivot or double down. → Opportunity Scorecard |
| 7 | Speed = brutal cadence (1–3/day) in a high-retention niche. Rare but documented. → Speedrun |
| 8 | Protect structurally: separate Gmail per pod, entity-level AdSense, never circumvent. → Account Setup |
| 9 | Your edge: production craft + remixed delivery (voice/audience/language) in niches that reward better production. → Creative Matrix |
Use it to research
Open the Opportunity Scorecard while you're mining a niche. Plug in the numbers you find (RPM, outlier multiples, competition) and get a go/no-go score with target thresholds.
Use it to plan
The Channel Planner turns the creative matrix into a concept builder; the 90-Day Roadmap sequences the launch.
Use it to operate
Launch KPIs tells you, week by week, whether a channel is working — with the exact thresholds and the decision rule (keep / iterate / kill).
The KPIs that tell you, during research, whether a niche/format is worth launching. Read each signal against its target, then use the live scorer to get a composite go/no-go.
The 7 opportunity signals (what "good" looks like)
| Signal | What it measures | Target = opportunity |
|---|---|---|
| Niche RPM | $ kept per 1,000 views (after YouTube cut) | $8+ good · $15+ excellent |
| Outlier multiple | Top video views ÷ that channel's average | 10×+ (breakout format exists) |
| View : sub ratio | A breakout video's views ÷ the channel's subs | 50×+ on a small/new channel = whitespace |
| Competition depth | Active strong channels (>10–50k subs) on the core keyword | Few (<~30) and beatable |
| Demand | Search volume + Google Trends slope (12–24mo) | Medium-to-high & flat/rising |
| Retention potential | Does the format hold attention (story/curiosity vs facts-list)? | High (binge / loop / curiosity) |
| Your-edge fit | Does it reward better production / a delivery remix you can own? | High (craft compounds here) |
Live opportunity scorer
Plug in what you found while researching. The score weights RPM, the outlier proof, and competition most heavily.
Guide: 75+ strong go · 55–74 promising, test small · <55 weak field, keep looking. Weighting: RPM & competition 20% each, outlier proof 18%, demand 14%, retention 12%, view:sub 8%, edge-fit 8%.
Once a channel is live, these are the numbers that tell you to keep going, fix something, or move on — and the decision rules that turn the numbers into action.
get_my_* tools — get_my_channel_overview (views, subs, watch time), get_my_audience_retention (APV + first-60s), get_my_revenue_report (revenue + RPM), get_my_traffic_sources (browse vs suggested vs search), and get_my_top_videos (which formats to double down on). No manual entry needed.The operating metrics & their thresholds
| KPI | Why it matters | Green | Amber | Red |
|---|---|---|---|---|
| CTR | Demand for the package; feeds browse/suggested | ≥5% | 3–5% | <3% |
| Avg % viewed (long) | Satisfaction; drives the algorithm | ≥40% | 30–40% | <30% |
| First-60s retention | Whether the hook works | ≥60% | 45–60% | <45% |
| Watch-hours pace | Track toward 4,000/yr for YPP | on/above pace | slightly behind | stalled |
| Sub velocity | Audience compounding | accelerating | steady | flat/declining |
| RPM (once monetized) | Earnings efficiency by niche/geo | ≥ niche benchmark | near | well below |
| Traffic mix | Browse/Suggested = algorithm love; all-search = capped | growing browse/sugg. | mostly search | no external reach |
Channel health checker
Enter your current numbers; each lights up vs the threshold and rolls into a stage-gate verdict.
Decision rule appears here.
Decision rules (turn numbers into action)
- Before 15 videos: don't judge. You're building the data set. Keep cadence, keep testing packaging.
- Low CTR, healthy retention: packaging problem → iterate thumbnails/titles (A/B), keep the content engine.
- High CTR, low retention: the promise isn't delivered → fix the hook (first 30–60s) and pacing.
- 15+ videos, still red across the board: the niche/format is wrong → pivot the angle or kill it (operators delete channels that don't clear profit after a 2–3 month test).
- Anything green & rising: double down — make a series, interconnect videos, clone the winning template, then replicate into a 2nd channel (separate Gmail/pod).
Your live research worktable, wired to NexLev. Columns map 1:1 to what NexLev returns — just say "research these channels and add them to the tracker" and I drop each one's stats straight in. Edit any cell and Score, Verdict, Status, and your projected earnings recalculate instantly. Autosaves in this browser.
My-channel projection assumptions
| Channel | Niche | Subs | Mo. views | Avg views/vid | Outlier | RPM $ | NexLev $/mo | Quality | Status | Model's style | My remix angle | Edge | Score | Verdict | My $/mo |
|---|
Score: RPM 25% · Outlier 25% · Format power (avg views/vid) 25% · Quality 15% · Edge-fit 10%. Verdict: ≥75 strong · 55–74 test · <55 pass. Status flags momentum, not quality — a Dormant channel with high avg views & outlier is a proven format nobody's currently running = opportunity to revive.
How NexLev fills this table
| NexLev field | → Tracker column |
|---|---|
title / category.name | Channel / Niche |
stats.subscribers · stats.monthlyViews | Subs · Mo. views |
stats.avgViewsPerVideo | Avg views/vid (the format-power metric) |
outlierScore · stats.rpm.total · stats.monthlyRevenue | Outlier · RPM $ · NexLev $/mo |
quality · lastVideoDate | Quality · Status (Active/Slowing/Dormant) |
format.name + my read | Model's style → I propose your remix angle |
Key interpretation baked in: I judge a format's power by avg views/vid + outlier score, and its current momentum by monthly views + last-upload date — so a low "NexLev $/mo" never hides an elite-but-dormant template (e.g. a channel with 220k avg views but $1/mo because it stopped posting).
Turn a proven topic into a differentiated channel concept by choosing one option per axis. Most operators only vary topic/approach/style in English — the wins live in voice, audience, and language.
Concept builder — pick one per axis
| Axis | Options |
|---|---|
| Topic | finance · history · true crime · psychology · health/longevity · tech/AI · motivation · sleep/ambient |
| Approach | explainer · story arc · listicle · documentary · contrarian · case study · "you-are-there" · musical edutainment |
| Voice | authoritative · energetic · humorous-cartoon · calm/bedtime · relatable peer · stoic · Gen-Z slang · warm "grandparent" |
| Audience | Gen Z · millennials · seniors · low-income/aspirational · professionals · kids · ESL learners |
| Production style | whiteboard/cartoon · charts+B-roll · cinematic stock · anime illustration · ambient loops · archival/maps · 3D · screen-recording |
| Language | English · Spanish · Portuguese (BR) · Hindi · Arabic · Indonesian … |
Gap recipes worth testing first
| Concept | Why it's open | Edge fit |
|---|---|---|
| Finance/health for seniors — warm voice, large-text simple animation (EN/ES/PT-BR) | $15–40 CPM topic, aging audience nobody serves faceless; "~19× growth, low comp" | High |
| Finance in PT-BR, hip-hop/musical for young low-income viewers | Nath proves demand; rap/song-style format wide open | High |
| Tech/AI or motivation as ambient/sleep loops in PT-BR | High-demand topic in a passive format; minimal comp, huge watch-time | Med-High |
| True crime/history as cozy bedtime + light music (ES/PT-BR) | Proven in English; localized variants rare | High |
Full creative formulas per niche & more recipes live in the Creative Matrix tab.
How to run the matrix as an experiment
This is the disciplined version of "I'd be great at many niches" — you're not spreading thin, you're A/B-testing delivery on a winning topic.
Sequenced launch for a producer using AI to scale original work.
Weeks 1–2 · Pick & validate
- Choose a high-CPM topic + a delivery remix you can own (use the Channel Planner)
- Run the Opportunity Scorecard; aim for 75+ before committing
- Build a swipe file of 20–50 model channels, narrow to 5–10
Weeks 3–6 · Build the system
- Lock the AI stack (script → VO → visuals → edit → thumbnail)
- Write the SOP: package-first, script template, editing rhythm, QC checklist
- Batch 15–30 videos so cadence is survivable; ship + A/B every thumbnail
Weeks 7–12 · Scale what works
- Use Launch KPIs weekly; double down on rising-retention formats
- Build the money-video flywheel (email + high-ticket) from the start
- Hit Tier-1 thresholds; prep a 2nd channel on a separate Gmail/pod once proven
Do
- Keep visible human authorship in every video
- Vary format/angle; test relentlessly
- Disclose realistic synthetic media; license music/footage
- Optimize for Tier-1 geo; chase CTR × retention
Don't
- Mass-produce near-identical templated videos
- Lean on stock TTS + looping AI images with no POV
- Run pure 2-hr ambient sleep for revenue
- Clone real voices without permission; misrepresent thumbnails
Repeatable, daily-able video templates that independently hit the winner-curve ($0→$1–10k/mo in 2–5 months). The "make many" unlock: build one production system per engine, then re-skin it across domains. All channels below are verified live in NexLev (new, faceless, monetized).
The engine library (verified)
| Engine | Proof channel ($/mo · RPM) | Clone-to domains |
|---|---|---|
| "Every Level / Rank of X" (ladder) ✅#1 | Master POV $11.8k · Ladder Theory $7.0k · Hypothetically $7.0k | wealth, careers, status, military, professions |
| "Every [X] Explained in N min" | Bluntly Explained $4.2k (high) · Animated History $3.9k (outlier 30) · Explainer Chris $2.5k · ExplainTory $2.1k | finance, tech, space, weapons, history, anything |
| "Why/When did humans X" (human-origins) | Ink Explainer $5.8k (high) · Mack $3.1k (1mo old) | psychology, science, history, biology |
| "Why your brain does X" (behavioral psych) | TheObservationDeck $0.9k (outlier 28) | psychology, productivity, self-help |
| First-person affirmation monologue | Willspire $2.3k (outlier 25.7, RPM $6.1) | motivation, discipline, identity |
| Calm "quiet wealth" finance | Wealth Logic $6.0k · Money Logic $1.7k (2mo) · Stoic Economist $1.0k | finance, psychology-of-money |
| Philosophy + sleep (Zen, 3hr) | Midnight Monk $4.2k (RPM $9.8) | philosophy, spirituality, sleep |
| "Why [Company] Collapsed / Abandoned & Left to Rot" | Bankruptcy Breakdown $2.3k (RPM $7.2, 2.8M-view hit) | business, brands, theme parks |
| Cinematic history battles / docs | History Reforged $3.8k (high) · Animated History $3.9k | history, military, empires |
| "What if [modern] vs [historical]" | Plot Twist Files $3.6k | history, military, hypotheticals |
| "Living in [country]" 4K docs | Nationscape $3.7k | travel, geography, culture (~195 countries) |
| Original true-crime docs | Dead End Files $2.8k (high) | crime, mystery |
| Animated science curiosity ("Do animals X") | Think Science $7.8k · Franz $5.6k (484k avg views) | science, biology, health |
| "I [task], You Sleep" (gaming+sleep) | DangerouslySleepy $8.4k | gaming, ASMR, ambient |
| Brand value audits ("ranked worst→best") | Her Highstreet Audit $1.2k (high) | any product/retail category |
| Cryptid/mystery science ("X keeps being spotted") | Primal Lens $1.8k | wildlife, mystery, science |
Saved models live in your NexLev swipefile "Launch Candidates — Faceless Empire." Earnings are NexLev estimates. RPM bands: explainer/finance $5–9; philosophy-sleep ~$10; science/animation $2–5 (volume play).
How to pick & clone an engine
- Pick by RPM × your edge: for highest $, lean engines with $5–10 RPM (explainer, quiet-wealth finance, corporate-collapse, philosophy-sleep). For reach, the high-outlier ones (every-level, animated science).
- Build one production template — VO voice, illustration system, title formula, thumbnail style, intro/pacing — so each new video is assembly, not invention.
- Re-skin across domains — the same "Every X Explained" pipeline becomes a finance channel, a tech channel, and a PT-BR channel. That's how one operator runs many.
- Score every candidate in the Opportunity Scorecard, log to the Channel Tracker, and judge after ~15 videos via Launch KPIs.
- Keep it fed — the weekly winner-scan surfaces new engines automatically.
CPM = advertiser bid / 1,000 impressions. RPM = what you keep / 1,000 views (~30–55% of CPM). Geo mix swings these hard.
| Niche | CPM | RPM | Notes |
|---|---|---|---|
| Personal finance / investing | $15–65 | $10–35 | Evergreen, long watch time, buyer intent |
| Insurance | $20–60 | $7–32 | Highest-value leads |
| Make-money / business / B2B SaaS | $15–45 | $8–22 | Strong B2B budgets |
| Legal / true-crime storytelling | $15–45 | $12–25 | Among highest reported RPM |
| Real estate / mortgages | $15–40 | $7–18 | High intent, cyclical |
| AI news & tools | $8–25 | $5–14 | Exploding, fast-moving |
| Gaming / entertainment / lifestyle | $1–8 | $0.5–5 | Volume play only |
| Shorts (any niche) | — | $0.01–0.06 | Growth engine, not revenue |
Validation workflow
- Seed 5–10 ideas from high-RPM categories
- vidIQ/TubeBuddy: medium volume, low/med competition
- Analyze top 10–20 videos (views, subs, cadence, length)
- Flag outliers (small channel, big views)
- Google Trends 12–24mo + by region
- Count strong competitors >10–50k subs
- Test 3–5 long-tail variations
- Ship 3 pilots, watch 30–90 days, iterate
Avoid / saturated
General gamingBeauty/makeupLifestyle vlogsGeneric motivationPlain meditation musicUndifferentiated ASMR2-hr stock sleep loopsDrill into micro-niches instead: "dividend investing for retirees," "WWII untold stories," "cold cases solved by AI forensics."
The arbitrage is delivery, not topic. The facts can be near-identical; voice, tone, style, audience, and language unlock a fresh, lower-competition audience.
Winning creative formulas by niche
| Niche | Formula that wins | Examples |
|---|---|---|
| Finance | Whiteboard/cartoon characters or charts + rise/fall "movie" arcs; authoritative/energetic; hook-heavy | Nick/Bob Invests, Economics Explained, MagnatesMedia |
| History | Humorous cartoon, epic animated maps/battles, or soothing documentary; 15–40 min | OverSimplified, Kings & Generals, Voices of the Past |
| True crime | Doc/podcast narration over timelines; dramatic or calm-bedtime; 20–60 min binge | Bedtime Stories, Lazy Masquerade |
| Psychology | Cute animated explainers or cinematic B-roll essays; insightful/calm | Psych2Go (13M), Aperture |
| Tech/AI | Screen demos + comparisons or cinematic B-roll deep dives | ColdFusion, TheAIGRID |
| Motivation | Speech/stock compilations or animated transformation stories | Motiversity (11M), Improvement Pill |
| Sleep/ambient | Looping visuals + pure audio or soft guided narration; 8hr+ streams | Lofi Girl (15M, ~$20–45k/mo) |
Proven remix wins (same topic → new delivery)
- Finance → cartoon characters (Nick/Bob Invests): dry charts as Family-Guy-style whiteboard → 100k+ subs in months.
- Finance → young Brazilian-Portuguese (Nath Finanças): elite jargon → favela-relatable peer tone → explosive growth.
- True crime/history → soothing bedtime (Bedtime Stories, 1M+): dramatic docs → calm "fall asleep to."
- History → first-person "you are there" (Voices of the Past, 1M+).
- Any EN winner → AI-dub localization (ES/PT/Hindi): documented thousands/mo added; the #1 thing courses quietly teach.
The algorithm ranks CTR × average view duration + satisfaction. Validate the package before you build.
Thumbnail principles
- One focal point, 2–3 elements, high contrast, mobile-readable at 320–400px
- Faceless emotion proxy: reaction graphic, before/after, expressive object
- 0–4 words ultra-bold; never duplicate the title; curiosity gap
- Custom beats auto-generated by 60–70%
CTR: <3% fix · 3–4% avg · 4–5% solid · 5–7%+ good · top creators 5–10%+.
Title formulas (50–70 chars)
- Number + Outcome: "7 Things That [Result] (Most Miss #4)"
- How-to compression: "How to [Big Result] in [Small Time]"
- Contrarian: "Why [Belief] Is Wrong (And What to Do)"
- Authority: "[Expert] Explains: [Claim]"
Clarity first, curiosity second half; match thumbnail intent; test 5 variants.
First 30 seconds (target 60%+)
0–5s: pattern interrupt / stakes · 5–15s: payoff promise + proof · 15–30s: roadmap + open loop. Rebuild if >40% drop in 60s.
Long-form scripting
- Nested open loops, story spine, re-hook every 2–4 min
- Pattern interrupts each segment; chapters; dense value
- APV 35–42%+ healthy; curve shape > raw number
Treat winning videos as data, not templates. Extract the blueprint; re-author with your own angle, script, visuals, voice.
Find & swipe
- Outlier finders: 1of10 (10×–100×), Spotter Studio, TubeLab, vidIQ trends
- Competitor trackers: vidIQ Competitors, TubeBuddy comparison, TubeAnalytics
- Swipe file (Notion/Airtable): thumbnail, title, hook transcript, structure timestamps, cadence, monetization signals
- Start 20–50 channels → narrow to 5–10 models; refresh weekly
Reverse-engineer packaging
Thumbnails (70–80% of the click): bold 3–5 words, contrast, emotion proxy, arrows, curiosity gap. Titles: number+benefit, question, curiosity+keyword. Hooks: transcribe winners to study rhythm.
Differentiation rule
If a viewer could mistake your video for the original (same script/footage/packaging w/ minor changes), it's too close — a reused/inauthentic risk. Vary visuals 20–50%+, write original hooks/analysis, change sub-angle/data.
Enforcement is pattern-based and channel-holistic, leaning on engagement signals + viewer reports.
| Area | Hard rule (will hurt you) | Bendable / gray |
|---|---|---|
| Inauthentic content (Jul 2025) | Templated, little-variation, mass-produced, replicable-at-scale with no human input | High-effort AI-assisted; loose templates w/ unique value & variation |
| Reused content | Reposts, minimal edits, scraped/duplicative | "Significant" added value — commentary/reaction lives here |
| AI disclosure | Realistic synthetic people/voice/scenes — must toggle "altered" | Clearly unreal/animated, assist-only = no disclosure |
| AI voice / TTS | Repetitive TTS no value; cloning real voice w/o permission | AI narration + strong scripting monetizes fine |
| Thumbnails / metadata | Materially misrepresenting; systematic fake-outrage | Curiosity-gap tolerated if video delivers |
| Copyright / Content ID | Unlicensed music/clips, untransformed reuse, repeat strikes | Fair use for commentary/criticism/education — case-by-case |
YouTube penalizes the operator/identity, not just a channel — via ~7 signals (Google, AdSense, IP/device, recovery info, managers, content similarity). Isolate to contain damage.
Separate Gmail / pods vs umbrella
One Google = 1 personal + up to 100 Brand channels. Umbrella is convenient but shared login/IP links everything. Best practice: separate Googles in pods of 2–5; lose a pod, not the network.
Contagion
- Single video flag: usually contained
- Channel termination: cascades — "applies to all your existing channels, new channels, and channels you appear on regularly"
- Same Google = highest · separate = lower · same AdSense = very high
AdSense = the strongest link
One per person (or per documented entity). A severe violation can disable the whole AdSense → all linked channels die at once. Split liability via separate LLCs each with own AdSense.
Limits
- 100 channels / Google account
- No hard cap on accounts a person owns if compliant
- Real limit = quality + management
Setup checklist
- Foundation: dedicated Gmail per pod, own recovery info, a private records sheet
- Monetization: one clean AdSense (simple, shared risk) OR LLC + AdSense per pod (max protection)
- Content separation: distinct niche/voice/visuals per channel (similarity is a linking signal)
- Operate: monitor Studio, appeal fast; if terminated, never create a replacement on a linked identity
Hybrid by design: AI does grunt work (research, draft, VO, B-roll, rough cut); you keep creative ownership (angle, edit, story, polish).
| Stage | Tools | ~Cost |
|---|---|---|
| Script / research | Claude, ChatGPT, Gemini, Perplexity; OverseerOS | $20/mo |
| Voiceover | ElevenLabs (multilingual, cloning) · Murf, Play.ht | $5–22/mo |
| Video / image | Veo 3.1, Kling 3, Sora 2, Runway Gen-4, Higgsfield; Midjourney | $20–50/mo |
| Assembly / edit | InVideo/Pictory, Descript/VEED, CapCut/DaVinci, OpusClip | $19–99/mo |
| Thumbnails | Canva Pro + AI, Midjourney, Photoshop AI | $15/mo |
QC checklist (human watch-through)
- Unique angle? Hook every 30–60s?
- Script: original insight, no generic AI tropes
- VO: emotion matches, no artifacts, consistent
- Visuals: coherent, no glitches
- Edit: tight pacing, readable text, music enhances
- Thumbnail/title: high-CTR & accurate; policy check
From every substantial operator interview + a decode of the paid courses (2,139 sources).
| Operator | Scale | Signature tactic |
|---|---|---|
| Wanner | 75 channels, $7–8M lifetime | Evergreen ASMR/sleep loops; 90% cost cut via repurposing; low-comp geo monetized in 9–10 days; Content ID licensing stack |
| Matt Par | 9–12+ channels, students $3–8k/mo | Tier-1 CPM; hook→pattern-interrupt every 30–60s→CTA; A/B thumbnails (+20–40% CTR) |
| Noah Morris | 18–20+ channels, $1M+/yr | "Find niches where even the worst content goes viral"; thumbnail+title one-two combo |
| Caleb Boxx | $3M+ personal | History/animation; prefers Gemini for scripts; judge after 15 videos |
The money-video flywheel (the underused lever)
Build videos that funnel into one long "money video," then convert. Matt Par's math: "$500 product at 4% conversion on 100k views = $20k/mo." Every view becomes a view + an email + a customer. Almost nobody builds the email/high-ticket layer from Day 1 — do it.
Course substance (decoded)
High: Matt Par (Tube Mastery), Nexlev — real templates/tools/communities. Medium: Caleb Boxx (great free videos, service-y paid). Low/hype: Frelan/AutoTube/2025–26 wave. The free promo substance = the whole playbook; paid mostly sells acceleration.
Raw operator signal — earnings claims unverified, directional only. This is what the daily 6 AM run keeps freshest.
Working / claimed (Jun 2026)
| Handle | Claim |
|---|---|
| @ChannelProfiler / @MuteeAutomation | Hyper-specific explainers breaking out — "Trucks Explained" 199K-view video ≈ 150× view:sub; Lulusia 30-min deep-dive 117K views on 2 videos |
| "Boring" nostalgia / history-of-simple-things | One channel 9.4M views ≈ ~$90K/mo (unverified) |
| Crude finance explainers (Nick/Alicia Invests style) | Cloned → ~1M views/mo ≈ $20–29k AdSense at ~$20 CPM |
| @erasay_new | Spanish "sleeping market" — 740K views in 28 days from scratch |
Getting hit
2-hr ambient sleep loops (low RPM, asleep during ads), generic top-10 listicles, zero-IP AI stories — fastest-flagged. A 588K-sub channel ($30k/mo) fully demonetized while still ~1M views/mo. "AI as a tool = OK; AI as creator = demonetized."
Outlier cases of scratch → monetized in under ~30 days. Rare and real — but "monetized" means hitting the YPP threshold fast, not $1k landing in week one.
| Handle | Numbers | Cadence |
|---|---|---|
| @GoodnessDeFi | 2nd channel monetized in 1 month; 28d = 1,040 subs, 168.3k views, 12k watch hrs | 1 video/day, self-built |
| @sniffyy_ | 3rd channel approved <24h that month; 97.8k views, 7.1k watch hrs, +1k subs | 2→3 uploads/day |
| Beyond the Veil | $7,039 in 28 days, 853k views (ceiling example) | Spirituality storytelling, 83 vids |