The Growth Playbook 🌱

Every social-media platform has one goal: keep people scrolling. The algorithm is a matchmaker between content and viewers. When you post a video, the platform builds a "digital fingerprint" of it using:

• Computer vision — what's happening visually
• Audio fingerprinting — what's being said
• Metadata — captions, hashtags, creator info

Then it picks ~200 people to test your video on first. Most of those 200 are non-followers — the algorithm wants to see if strangers like it. If they engage, the video gets pushed to 2,000 → 20,000 → 200,000 in waves. If they don't, the push slows or stops immediately.

2026 note for YouTube: that "metadata" bullet got demoted. YouTube now builds the fingerprint from the video itself (frames + audio) and uses your title, description and tags to confirm it, not to define it. Metadata that disagrees with the footage is a contradiction, and contradictions are what the clickbait filter hunts. Full lesson: The 2026 YouTube shift.

• Average watch time + completion rate — how long viewers stay, and what % finish
• Engagement rate — likes + comments + shares ÷ views
• Watchtime session share — how much of a user's TOTAL session was YOUR content. You can't see this anywhere in your analytics dashboard, but the algorithm weighs it heavily.

Your video needs to hit all four of these to get pushed:

• Relevant — solves a real problem your target viewer has
• Non-obvious — says something they haven't heard before
• Absorbable — communicated clearly enough to actually understand
• Actionable — they can implement it and get a result

Platform	What works
Twitter/X	Keyword-rich text, NO hashtags, post 1–3× daily, short looping clips, upload video directly (don't link to YouTube). Replies matter MORE than your own tweets.
Instagram	5–7 hashtags sweet spot, great for discoverability, use Stories, dots hide hashtags from preview
TikTok	Stay in your lane, keywords in caption, minimal hashtags (1–3 max), trending sound > perfect tags
YouTube Shorts	Include 2–3 brand-relevant hashtags, treat it like its own TV channel, top 2 in title for indexing
Fansly (18+)	FYP works extremely well. Hashtags MATTER here — tag specifically + always #vtuber. Shorts 6–20s.

What changed. YouTube used to work out what a video was about from what you typed. In 2026 it switched to reading the video itself: Gemini watches the frames, listens to the audio, and builds its own profile of what you made and who would want it. Your title, description and tags now confirm that profile instead of creating it. When they disagree with the footage, that is a mismatch, and mismatch is exactly what the clickbait filter looks for.

Two new discovery surfaces came with it:

• Ask YouTube: viewers ask a full question or describe a vibe; YouTube answers with videos and starts playback at the timestamp that answers it, not at 0:00.
• Custom Feeds: a viewer describes what they want to watch and YouTube builds a home-page feed from that prompt.

The four free wins (do these this week):

1. Plan around questions, not one keyword. Pull the question-shaped searches for your game before you plan (vidIQ keyword tool ➜ questions). Real numbers, Sept 2026: guild wars 2 = 314k searches at competition 44.5; how to get mesmer shield warframe = 9.6k searches at competition 7.4 and a higher opportunity score. Sort by opportunity, prefer competition under 20. Your hook rules do not change; this only narrows which words you pick.
2. Build in segments. One clear subject per block. Ask YouTube serves timestamps, so a cleanly divided video beats a ramble on structure alone. Make a few segments strong enough to stand alone as Shorts: one shoot, two surfaces.
3. Descriptive chapters. This reverses old advice. "Method 1 / Method 2" now hides the information YouTube uses to place you. ❌ 0:00 Intro · 2:10 Method 1 ✅ 0:00 Why I came back to Guild Wars 2 · 2:10 Gold farming with the Drizzlewood meta. Still needs the first stamp at exactly 0:00 and at least three chapters.
4. Say the subject out loud. The exact game, boss, product, place, early and whenever it changes. ✅ "Where Winds Meet just changed how the grapple works" ❌ "so this game just changed how this works". Do not overcorrect into robot narration; retention still decides whether you rank at all.

Stop: treating the tag field as a ranking lever (fill it in 30 seconds for ambiguous names, then move on) · keyword-stuffing descriptions · vague chapter names · planning only around the head keyword.
Keep (none of this changed): CTR, average view duration, the first 30 seconds, thumbnail contrast, a title a human wants to click. Gemini changed how you get matched; it did not cancel having to hold the viewer.

New viewers only see your title before clicking. A good title creates curiosity, shock, or confusion — anything that makes them stop scrolling.

Format that works on Twitch:

[Aesthetic emoji/decoration] + [Hook/Meme] + [Game] + [VTuber] + [Commands]

• "Chill vibes" tells nobody anything — give them a reason to click
• Challenges work great: "If I die, I restart" beats "cozy stream" every time
• "24 hour stream" creates instant curiosity
• Questions work extremely well — "HE DID WHAT WITH THAT SPATULA?!"
• Goal: create confusion, shock, or curiosity → drive clicks

YouTube long-form, 2026: keep the hook, but pick your words from the questions people actually ask. "What class to pick in Guild Wars 2" is searched 5k times a month at a sixth of the competition of "guild wars 2". Write the title as the human-readable answer to a real question, then make it clickable. Details in the 2026 shift.

Always use a mix of small / medium / large tags so your content can perform at ANY stage of the algorithm:

• Small tags (under 500k posts) — low competition, easier to rank in. Your easy-win slot.
• Medium tags (500k – 1M) — balanced reach. Most of your discoverability lives here.
• Large tags (1M+) — high exposure if you break out. Don't lean on these alone or you'll get buried.

Standard mix: 1 small / 2 medium / 2 large. If you already have audience traction, lean more on medium + large.

One thing this formula is NOT (2026): the YouTube tag field on the upload screen. That box is close to dead now that YouTube reads the video itself. Fill it in 30 seconds for genuinely ambiguous names, then move on. Hashtags in captions and descriptions on Shorts, TikTok and Reels still work the way this chapter describes.

• Shorts are fed through a queue — no impressions, just views vs. swipes
• Aim for 45–55 seconds with 85–90% retention for best results
• Don't go under 25 seconds unless your short is extremely rewatchable
• The "viewed vs. swiped" stat is your most important early metric — aim for 75%+
• Edit your shorts. Always. Auto-converting landscape clips does not work — vertical-first or it dies
• Watch shorts in your niche for 1–2 hours a week — study what's working
• Say the game or subject out loud in the first line (2026). YouTube transcribes Shorts and matches them by what is said and shown. "This game" tells it nothing; "Warframe" gets you into every Warframe prompt

• Take a hard stance — people comment when they strongly agree OR disagree
• Pick the contrarian side when you can
• Ratchet up your framing — be bold, not wishy-washy
• Talk about cult-loved brands, games, or communities people already have opinions on
• Make people FEEL something — emotion drives comments more than logic

If a new viewer types in chat they're 50% more likely to return. Welcome EVERY new chatter — sounds obvious but most streamers don't do it.

• Use open-ended questions, not yes/no — "What's your favorite game from childhood that still holds up?" not "do you like this game?"
• Put a question or hot take IN your stream title to prime engagement before they click
• Stream Elements timers — remind viewers of commands and follows on a loop
• Loyalty points + commands like !gamble or !heist keep chat participating
• Fun commands like !iq or !based give chatters something to spam at each other
• Words on Stream, Gartic, or chat-controlled game mods spike engagement fast — these also trick the algorithm into thinking your chat is more active than it is

Stream structure tips:

• Have a routine 20 minutes before going live — helps you show up consistently
• Introduce yourself every time someone new follows or raids — "Hi I'm [name], I play [X], love [Y], welcome to the cult"
• Always be plugging what you're working toward — fundraisers, goals, milestones

Core mindset shift: if you only post to drive people elsewhere, you'll struggle on ALL platforms. Learn each platform for what it IS — growth on those will naturally convert. People who bond with you on short-form will seek out your stream on their own. You don't need constant CTAs — if they like you, they'll find you.

Why short-form converts at all:

• Short-form gives viewers a concentrated taste of your personality
• They see the edited, best-of version of you first
• If they like THAT, they'll want the unfiltered live version
• Make sure your short-form and stream personality are consistent — high-energy shorts + dead-silent stream = they leave

What actually converts:

• Stream schedule clearly visible in your bio everywhere
• Content backlog — new viewers want old videos to binge
• Discord ready BEFORE you go viral — people need somewhere to land
• Consistent upload schedule signals you're worth following long-term

Funneling YouTube → Twitch:

• Don't start your YouTube stream until you begin gaming — YouTube hates category switches mid-stream
• Do your Just Chatting / reaction segment first, THEN start the YouTube VOD
• Keeps your YouTube content clean and focused for the algorithm
• Clip your stream highlights and post them as shorts — double duty

Content that travels: your model is a costume — the CONTENT is what gets viewership. Always ask: "Would someone who has never heard of VTubing still find this funny, useful, or interesting?" If no, go back to the drawing board. Reaction videos, gaming fails, hot takes, and unhinged opinions travel furthest. Inside jokes and VTuber-only memes keep your growth flatlined.

Everything Twitch — start at the top, or jump to a sub-section. The 2026 deep-dive (#twitchDeep) and OBS setup (#obsSetup) are folded in below as anchors so old links still work.

Twitch's discovery system is broken for small streamers. Categories are sorted highest viewers first — you get buried. Streaming alone and hoping people find you will not work. You need to build your audience BEFORE relying on Twitch to push you.

The actual strategy:

• Build your presence on a platform with a working algorithm first (TikTok, YouTube Shorts, Instagram Reels)
• Funnel viewers from short-form into your stream
• Even 1% click-through compounds over time
• More viewers → higher in category → Twitch starts pushing you organically

Niching down is non-negotiable:

• "Variety streamer" puts you in a box with 300,000 other people
• You need to be describable in one sentence
• Example: "I play Souls games, love Warhammer, and give VTuber tips"
• Pick games with viewers but LOW streamer competition — use TwitchTracker to find peak ratio time slots

Mindset shift:

• Consistency ≠ streaming every day — it means reliable patterns
• Don't chase trends blindly — find where trends OVERLAP with your content
• Passion alone grabs attention. Passion + strategy keeps it.
• You don't need to go viral. You need to compound.

The truth about Twitch discoverability. Twitch's internal discovery is still limited. Real growth comes from external platforms (X / YouTube / TikTok), networking and collaborations, and community loyalty. Twitch rewards watch time, returning viewers, and chat activity — not follower count.

Basics that have to be solid before anything else: clear microphone audio, stable internet, functional VTuber model (or PNG), readable alerts and overlays. Perfection is not required — clarity is.

Define your stream identity. Ask: what vibe do I offer (cozy, chaotic, chill, competitive)? What emotions do viewers feel? Why would someone come back tomorrow? Pick one vibe and stick to it. Don't let your streams feel "raw" — get comfortable, and viewers will too.

Content selection — the game trap. Don't camp in oversaturated games with thousands of streamers. Better growth categories:

• Mid-size games with active communities
• Niche genres you genuinely enjoy
• Rotating "main game + variety"

Sample rotation: 60% main game · 30% secondary growth games · 10% experimental.

Consistency over frequency. Streaming 3–4 days a week consistently beats daily random streams. Same days, same times, 2–4 hour streams. End while your energy is still high. Reliability builds trust.

How to make your streams entertaining.

• Talk even when chat is quiet. Explain your thoughts out loud. React visibly with your model. Treat lurkers like viewers — most viewers never chat. Silence kills retention faster than bad gameplay.
• Reframe: think of quiet streams as recording sessions. Every clip is future short-form content. If you're silent there's nothing to clip.

Building chat culture + community identity. Set the tone early — your community is a reflection of YOU.

• Establish clear rules + boundaries
• Model the behavior you want to see
• Shut down negativity quickly · don't be scared to ban
• Create a community name + inside jokes

High-impact viewer interaction tools: channel point redeems · polls and predictions · viewer-named challenges ("for 5 subs do X", "I bet you won't") · chat-chooses (game, champion, language) · spin-the-wheel punishments (no empty / punishing options for viewers — that becomes gambling).

How to grow a deeper connection with chatters:

• Always say their name out loud (a piece of it is fine if it's long, just make sure they know it's them)
• When responding to a chat message, read it out loud fully — chatters get excited to write knowing you might read it, and clips with read messages look much more professional
• Greet every new follower, raid, and first-time chatter — you have ~3 seconds to hook them before they leave
• Remember small details viewers share (timezone, job, name, inside jokes) — builds deep connection. But watch the parasocial line carefully and cut it short if someone crosses it.
• Include your chatters in your content — sub names on your model, screenshots with their name in clips, "chat" → their actual handle
• Announce every ad break 60 seconds before it hits. Use the break for a restroom / drink. Viewers stay because you didn't continue without them
• Be a little playfully mean to regulars sometimes — sarcastic clips land really well, but ONLY with chatters you know won't get offended. Never with newcomers (unless they were mean first)
• Pin your Discord invite at the top of chat and interact with viewers off-stream too

Networking + collaboration — the right way. Networking is building genuine long-term creator relationships. It is NOT cold DMing strangers, dropping links in unrelated Discords, or treating other VTubers as stepping stones. It IS supporting creators you genuinely enjoy, showing up consistently in shared spaces, and creating value before asking for anything.

Channel readiness checklist before serious networking: solid audio · functional model + a PNG of your model your collab partner can put on screen · semi-consistent schedule · public VODs so future partners can vibe-check you.

Chat etiquette in other people's streams:

• NEVER mention your own stream
• Match the streamer's energy
• Don't dominate conversation (no spam)
• Remember inside jokes, ask thoughtful questions, show up consistently

Turning acquaintances into creator friends: move from public chat to DMs · share struggles (but don't complain 24/7) · support their off-stream content (covers, merch drops) · give without expectation (retweet announcements, attend debut streams).

How to ask for a collab. Ask after multiple friendly interactions, shared interests, and ideally after they follow you back. Template:

Do NOT copy-paste the same message to many people — the platform will flag you as spam and your messages won't go through.

Planning collabs: define stream length, game, POVs + content boundaries, time zones, backup plans. Things go wrong — reschedule and run your own stream as planned.

On-stream collab etiquette: share the spotlight · don't talk over others · include everyone · don't laugh someone off if they make a mistake · don't assign blame live · stay professional.

After the collab: thank collaborators publicly, share clips from all POVs in one edited clip, follow up privately, stay in touch casually (avoid only messaging for collabs).

Warning signs someone is using you: only talks about numbers · pushes collabs immediately · disrespects you publicly · only ever texts you to collab · doesn't want to spend time with you off-stream · mentions you a lot in a way that feels performative.

Collab activity menu — low energy: Gartic Phone · GeoGuessr · Jackbox · tier list debates · Guess the VTuber · rating VTuber designs · smash or pass (Pokémon edition etc.) · reacting to videos · Among Us · viewer VC interactions (therapy sessions, confession stories, embarrassing storytime).

High energy collab activities: Minecraft · competitive duos (LoL, OW2, Apex, Valorant) · Mario Kart for big groups · Lethal Company roleplay · Phasmophobia · Peak · Repo · Roblox games.

Platform roles for cross-promo:

• Twitter/X: personality, brand building, networking
• TikTok / Instagram: discovery + virality
• YouTube: discovery, virality, long-term growth
• Discord: community hub — most loyal fanbase lives there

One strong platform is enough to start.

Clipping strategy. Clip daily if possible (1 video a day is enough — ignore the "post 2–5/day" advice, it's not sustainable). Add captions ALWAYS. Post consistently — if you can't do daily, pick 3 days a week and stick to them. Treat every stream as clip farming time.

OBS — video settings (Settings → Video):

• Base (canvas): 1920×1080
• Output (scaled): 1920×1080 for strong PCs, 1280×720 for weaker
• FPS: 60 (drop to 30 only if your PC can't handle it — viewers can tell)

OBS — output / encoder (Settings → Output → Streaming):

• Do NOT use x264 — that's your CPU and it will cook your PC
• Use the encoder with your GPU's name (e.g. NVIDIA NVENC H.264)
• Bitrate: 4500–6000 kbps (Twitch's max is 6000)
• Keyframe interval: 0s
• Preset: P6: Slower (Better Quality)

OBS — audio. Bad audio makes new viewers click off instantly. Bad video quality matters way less than bad audio.

• Sample rate: 48 kHz (match Windows)

Microphone filters (right-click mic source → Filters):

• Noise Suppression: RNNoise (best quality)
• Compressor: Ratio 4:1 · Threshold −18 dB · Attack 6 ms · Release 50 ms · Output gain 0 · Sidechain None
• Noise Gate: Close −55 dB · Open −60 dB — tune so it doesn't cut your sentences at the start/end
• Expander: Ratio 2.0 · Threshold −40 dB · Attack 10 ms · Release 50 ms · Output gain 0 · Detection RMS
• Gain: ~9 dB (tune to your voice)

Separate music from your Twitch VODs. Use OBS's Audio Application Capture source to isolate music on its own track. Why:

• Protects you from copyright strikes / bans
• You can listen to whatever you want, not just royalty-free
• Clean canvas for editing clips later (no muting needed)
• You can upload VODs to YouTube and actually monetize them

Add an Audio folder per scene with sources for music (e.g. iTunes) and game audio. Mute the global desktop audio source so it doesn't double. Set Twitch VOD to Track 2 in Settings → Output. In Advanced Audio Properties, tick tracks 1 + 2 for everything you want in both stream and VOD — then untick Track 2 on your music source. Music plays live but is silent in the VOD. Walkthrough: YouTube tutorial. (This feature is Twitch-only currently.)

VTube Studio capture — use Spout2. Default game capture is heavier on CPU than it needs to be. Install the obs-spout2-plugin, then:

• In VTube Studio: Change background → color picker → BLACK (not green)
• In OBS: add a Spout2 source — VTube Studio should auto-detect (otherwise pick VtubeStudioSpout)
• Tick Allow transparency

Don't open VTube Studio via Steam — performance hit. Right-click VTube Studio in Steam → Properties → Installed Files → Browse → Start_without_steam. Make a desktop shortcut. (VBridger is fine to open normally.)

Model framing. Don't cut the top of your model in VTube Studio — keep it smaller in VTS and zoom in OBS. Frame from waist up or bust up for the best model-to-game ratio.

Scene organization (the 4-scene template):

• Starting Soon: animated/still character illustration · chatbox · alertbox · timer · sub or donation goal · audio capture folder
• Just Chatting: chatting overlay · chatbox · alertbox · goals · audio folder · Spout2 model
• Game Scene (one per game): copy Just Chatting, swap chatting overlay for the game · add an in-game overlay if you have one
• Screen / Browser capture: copy Just Chatting, replace overlay with Window Capture (NOT Desktop Capture — risks doxxing yourself)

Always reuse existing sources rather than making new ones — huge performance win. Always name your sources.

Twitch profile basics.

• Name: avoid numbers (looks less professional). If your handle is taken, add cute letters or doubled vowels — looks cleaner than tacking on a number
• PFP: face of your model or an emote. No AI images.
• Banner: simple — single color is usually enough
• Offline screen: pretty illustration of your model on a plain colored bg, "Currently offline" text, social handles with icons under it

Essential Twitch panels: About Me · Rules · Socials · Subscribe/Donate with perks listed · any sponsor panels. Optional: PC specs, schedule. Always add text under panel images — don't rely on image-only panels.

Twitch panel markdown: > at the start of a line makes a purple bar · [name](link) for inline links · **bold** for bold.

Stream widgets.

• Chat overlay: Streamelements/Streamlabs · big readable font · enable BetterTTV emotes so they render on screen too
• Alerts: 3–5s · distinct image + sound per type (sub vs follow) — use your emotes if you don't want custom art · for custom sounds, record audio with OBS then convert via cloudconvert
• Sub/follower goals: set small achievable goals ("0/5 daily subs") · keep them small on screen · add a text goal underneath ("cooking stream 0/15 subs") so viewers know what they're funding · don't use the default Streamlabs dono goal every stream — looks cheap
• Stream labels: latest follower / sub / donation — supporters love seeing their name on screen
• Sound alerts: set up via soundalerts.com · sounds priced in bits or channel points · test the volume in OBS

Channel point redeems. Mix cheap goofy ones ("Lurking", "Hi" — 100–300 points) with rare expensive ones ("VIP in chat", "chibi doodle by me"). Connect Mix It Up to trigger VTube Studio model toggles via redeems (hairstyle swap, outfit change, etc.). Add images + colors to redeems — your emotes work great.

Chat interactions: built-in polls + predictions · stream avatars/pets (streamavatars.com) that wander around your screen representing active chatters.

Performance + stability tips: close unused Chrome tabs · close Discord when not using it · monitor CPU temps · close Steam if you're not using VBridger · drop to PNG model if a game is choking your PC (troubleshoot after) · ask chat occasionally if they're seeing dropped frames.

Common mistakes: streaming only oversaturated games · ignoring chat · no off-platform content · obsessing over viewer count (hide the counter — focus on entertainment).

Sustainability: schedule days off · don't force a strict weekly schedule · rotate content · take breaks without apology. Growth is uneven, plateaus are normal.

Twitch tags are not hashtags. They're the searchable labels under your stream that tell Twitch — and viewers browsing a category — what your stream actually is. They power three things: directory filtering (viewers narrowing a category), search, and recommendations (Twitch's own help confirms it shows people more streams with tags they already watch). They do not replace your category — they refine it.

The rules:

• Up to 10 tags per stream, 25 characters each
• Letters, numbers, accent marks only — no spaces, so multi-word tags are CamelCase (PlayingWithViewers)
• Set them in Creator Dashboard → Stream Manager → Edit Stream Info → Tags; your language tag is added automatically
• They don't reliably persist between streams — re-check them each time so they match what you're actually doing

• Identity stack goes on every stream: VTuber always; add ENVTuber if you stream in English. These are your highest-value tags — the VTuber audience actively browses them.
• 5–7 genuinely relevant tags beats maxing all 10 with filler. Relevance is what compounds; irrelevant tags pull in viewers who bounce, which hurts you.
• Lead with mid/low-competition tags, add 1–2 broad ones. On a saturated tag you're buried on page 200; on a niche tag you're actually visible.
• Match tags to your title + category. Mismatch tanks click-through.
• Measure what works: Dashboard → Analytics → Discovery → Tags shows impressions + click-through per tag (last 30 days). Drop whatever has the worst click-to-view ratio.

The shape per stream: VTuber + ENVTuber + 1 category-true tag + 2–3 content/mood tags + (optional) 1 community tag.

Category = the specific game. Custom tags refine genre, play-style, and experience. Pick one from each group:

Group	Tags
Genre	FPS RPG JRPG MMORPG Soulslike Roguelike Gacha CozyGaming Survival Metroidvania VisualNovel Sandbox Strategy
Play-style	FirstPlaythrough NoSpoilers NoBackseating Casual Competitive Ranked CoOp Completionist Hardcore Modded Nuzlocke Speedrun BeginnerFriendly
Community	PlayingWithViewers ViewerGames Charity Subathon
Mood	Chill Cozy Chatty HighEnergy Funny

Horror is a gaming sub-genre — same category (the game), add Horror SurvivalHorror IndieHorror PsychologicalHorror Jumpscares Reactions, plus FirstPlaythrough + NoSpoilers for blind runs (the old BlindPlaythrough tag was retired in 2020). Spooky/Halloween spike every October.

Avoid: Gaming, Live, Twitch, ProGamer — too broad / no intent. Don't use Speedrun, Competitive, or Nuzlocke unless it's literally true. DropsEnabled is auto-applied — don't type it.

Examples: Cozy sim → VTuber ENVTuber CozyGaming Cozy FirstPlaythrough NoSpoilers · Gacha daily → VTuber ENVTuber Gacha RPG Chatty PlayingWithViewers · Ranked shooter → VTuber ENVTuber FPS Competitive Ranked Chatty · Co-op horror → VTuber ENVTuber Horror CoOp Jumpscares Reactions

The most saturated category — tags + title + personality carry it. Format: AMA Storytime Reaction WatchParty Podcast Yapping. Engagement: Chatty Interactive CommunityFocused. Community: LGBTQIAPlus SafeSpace Wholesome. Mood: Chill Cozy GoodVibes HighEnergy. Don't tag JustChatting (it's your category) or Variety/Fun (no intent).

Examples: Q&A → VTuber ENVTuber AMA Chatty GoodVibes Interactive · Cozy yap → VTuber ENVTuber Yapping Chill Cozy CommunityFocused

Category = Art. Medium: DigitalArt Drawing Illustration Painting TraditionalArt Sketching PixelArt CharacterDesign Animation Live2D VTuberArt. Activity: ArtCommissions TakingCommissions ArtRequests ArtTutorial EmoteArt Rigging. Tools: Procreate ClipStudioPaint Krita Blender. Mood: Cozy Chatty LGBTQIAPlus.

Examples: Commissions → VTuber ENVTuber DigitalArt ArtCommissions CharacterDesign Procreate · Model art → VTuber ENVTuber Live2D VTuberArt Rigging Sketching

Category = Music (DJ Program has separate licensed-music VOD rules). Format: Singing Karaoke Cover OriginalMusic Songwriting LiveMusic MusicProduction DJ Utaite. Genre/lang: AnimeSongs JPop KPop Vocaloid LoFi Jazz Rock EDM. Instruments: Guitar Piano Acoustic. Mood: Chill Cozy Chatty.

Examples: Karaoke → VTuber ENVTuber Singing Karaoke AnimeSongs Chatty · Original showcase → VTuber ENVTuber OriginalMusic Songwriting Vocals Chill

Variety is a brand, not a tag — plain Variety has almost no search intent. The move: tag for whatever you're actually doing this session (swap them as you switch games/segments) and let your identity + community tags hold steady. Keep VTuber ENVTuber + VarietyStreamer as a soft anchor, then layer the live activity's real tags. Mood tags that travel across anything: Chill Chatty HighEnergy Funny CommunityFocused.

Example: game half → VTuber ENVTuber VarietyStreamer + that game's genre/play-style tags; chatting half → swap to AMA Chatty GoodVibes.

Category = Science & Technology or Software and Game Development. Subject: Coding Programming SoftwareDevelopment GameDev WebDev Science TechTalk Educational LanguageLearning. Format: Tutorial Teaching LiveCoding StudyWithMe Coworking Pomodoro. Tools: Python JavaScript Unity Godot Blender. Mood: Chill BeginnerFriendly Chatty.

Examples: Dev stream → VTuber ENVTuber Coding Python LiveCoding BeginnerFriendly · Study-with-me → VTuber ENVTuber StudyWithMe Coworking Pomodoro Focus

Maps to Creative subcats (Makers & Crafting, Beauty & Body Art, Fitness & Health, Writing & Reading). Activity: Cooking Baking Fitness Workout Crafting Crochet Cosplay Beauty Makeup Reading Pets. Format: Vlog DayInTheLife GetReadyWithMe StudyWithMe BehindTheScenes HowTo. Mood: Cozy Chatty GoodVibes.

Examples: Cooking → VTuber ENVTuber Cooking Food Chatty GoodVibes · Crafting → VTuber ENVTuber Crafting Crochet Cozy Chill

Category = IRL / Travel & Outdoors. Activity: IRL Travel Outdoors WalkAndTalk EventStream FoodTour. Format: Vlog DayInTheLife JustChatting-style hangout. Mood: Chatty GoodVibes HighEnergy CommunityFocused. (IRL with no avatar? You can still tag VTuber if that's your brand, or drop it for an IRL-only session.)

Example: Out-and-about → VTuber ENVTuber IRL Travel WalkAndTalk Chatty

Category = ASMR (audio-quality dependent, moderation-heavy — mind Content Classification Labels, which carry over between streams). Type: ASMR Relaxing Tingles WhisperASMR SoftSpoken SoundASMR Meditation SleepAid. Trigger/activity: Tapping RolePlay Reading Whispering. Cozy crossover: Cozy Comfy Wholesome LowStakes Calm.

Examples: ASMR session → VTuber ENVTuber ASMR WhisperASMR Tingles Relaxing · Cozy night → VTuber ENVTuber CozyGaming Cozy Wholesome LowStakes

Category = VRChat. Activity: VRChat VR SocialVR VRChatWorlds Avatars FullBodyTracking Exploration RolePlay. Mood: Funny HighEnergy Chill Chatty CommunityFocused. VRChat skews comedy/social — lean into Funny and PlayingWithViewers when you're meeting people in-world.

Example: World-hopping → VTuber ENVTuber VRChat SocialVR Funny Chatty

How X works in 2026. The platform now prioritizes replies and conversations, time spent on a post, profile visits after engagement, and repeated interaction between the same users. Follower count matters less than it used to. Returning people bring the most new traction to your account — so going viral is optional, but building a recognizable presence is mandatory. Growth is cumulative, not explosive.

Profile checklist — your profile should instantly answer "who are you, what do you stream, why should I follow."

• Clear name with an emoji — people will recognize you by the emoji alone
• Big creators have stopped putting "vtuber" in their display name; don't lock yourself into it
• Social media links must be findable — put a Carrd or Linktree in the bio with socials, merch, Discord, newest YouTube video, donation link, affiliate codes
• List your partnerships: Twitch partner, YouTube partner, GamerSupps code, etc.
• Contact email visible — most credible brands try email first. Envelope emoji in front keeps it cute
• Include your art tag
• Model / PFP credits if you didn't make them yourself
• Use the new "extended bio" option for the rest

Pinned post. Should include: short intro, model references (reference sheet + render PNG), social media links (Twitch / YouTube / TikTok most important), contact email at the bottom. No hashtags in the post itself — hashtags limit engagement on X. Use the pinned post as a link drop for art raffles too, so your media tab doesn't get cluttered with model PNGs. Bonus: views on your pinned roughly tell you how many people visited your profile.

The 4 core VTuber content pillars:

• Personality posts — jokes, opinions, relatable thoughts
• Stream content — schedules, clips, screenshots
• Community interaction — replies, questions, polls
• Creator networking — supporting other VTubers

Relatability = shareability. Relatable posts perform because people think "this is literally me." Examples: "Buying new emotes instead of fixing audio because priorities" · "Being nervous before stream even after doing this for years is crazy actually."

Stream promotion without killing engagement. Never post links on their own — links cut your reach a lot. Counter that by uploading something with the link: a meme of your character or a vertical clip from your TikTok/YouTube. Or, post the announcement and drop the link in a comment instead. Even text-only stream announcements should have an image.

Media tab best practices. 90% of new people check your media tab before deciding whether to follow. Keep it clean.

• 1 image or clip per tweet (avoid cropping)
• Subtitles on every clip
• Casual screenshots outperform polished graphics — overly professional VTubers don't land well
• High performing media: fanart reposts with credit, commissioned art, reposted TikToks/Shorts, short reaction clips from streams, VTuber trends using your model, meme edits with your model, achievement screenshots
• Avoid: replying with GIFs (clutters media tab), posting your model PNG under art raffles (link the pinned post instead), reposting the same thing too often, posting NSFW on main (can kill your reach for months — make a dedicated alt)

Premium subscription — necessary? Field-tested on both a 51K and 16K account — no meaningful increase in views or engagement. Premium boosts comments' visibility (your replies show first in others' threads) and the view count those rack up makes your own analytics look inflated. The features that actually justify it: ability to earn from ads in your comments (if your country supports it), longer posts, post editing, streaming on X. If you can monetize: questions, polls (asking for comment-votes, not poll-clicks), and "your last saved image is a reaction to this post"-style posts farm comments and ad revenue.

The power of replies. Replies matter MORE than your own tweets. Being active on others' posts signals to X that your account is active, which pushes your own posts further too.

• Daily goal: 5–10 genuine replies under posts of creators with similar content
• Reply with small harmless jokes, compliments, shared experiences, or relevant questions
• Keep replies relevant — don't look like a bot

Networking with other VTubers. Real networking is commenting, liking, retweeting, and showing up in their Twitch chats. Never ask for a collab cold. Follow first, comment a few times (not every single post), say hi in chat, and once they interact back — then you can DM. Quality over quantity. Not everyone will like you, focus on those who do.

Posting frequency. 1–3 tweets per day, spread throughout the day. Post when your audience is awake, be consistent, test times to find when your followers are most active. Don't overwork yourself — burnout is not worth it.

Mistakes that kill VTuber growth:

• Only tweeting stream links
• Complaining
• Chasing virality at all cost
• Comparing numbers
• Using AI
• Spamming · not posting enough · not being consistent
• Only making content in a language that isn't English — biggest underrated growth-killer. Limits reach to one region. English opens you to the whole globe.
• Constantly rebranding — pick who you want to be and stick to it

Algorithm myths:

• Shadowbanned? Unless you recently posted NSFW, probably not. Your post just hit at a bad time or didn't resonate. Stay consistent.
• You need viral tweets to grow? No. Tweets that consistently get 200–1000 likes are really good. Consistent engagement beats one banger every now and then.

Sustainable growth + mental health. Mute people who make you anxious (muting is private — blocking is visible from their profile). Mute big news accounts. Avoid doom-scrolling and comparison. Set work hours and stop checking after them. Stop reading comments under huge posts if it stresses you out. Growth isn't linear — 1000 followers in a few hours one day and zero for a month after is normal.

Everything money — start at the top, or jump to a sub-section. The deep-dive (#monetizeDeep) and Fansly (#fansly) are folded in below as anchors so old links still work.

Mindset first: you are not begging your audience — you are a virtual busker. Buskers perform, and if you like them, you throw coins in the case. Entertainment has value. You are allowed to ask for support. You earned it by showing up. Shift from "I hope someone donates" to "here's what we're working towards."

Diversify your income — don't rely on one source:

• Twitch subs + bits
• YouTube ad revenue
• Sponsorships + affiliate codes
• Merchandise
• Tip jar / StreamElements
• Throne wishlist
• Paid collaborations

Don't quit your day job until streaming has brought in more than your living expenses for at least 6 months. Streaming income can drop as fast as it rises.

YouTube monetization.

• YouTube Partner Program (ad revenue) — requires 1,000 subs + 4,000 public watch hours in 12 months OR 10M Shorts views in 90 days
• Ad revenue is very inconsistent month to month
• To actually earn from YouTube, focus on long-form > 8 min (lets you put mid-roll ads — most of the revenue). Clip your streams into highlights. Use Shorts to bring traction to the channel, not to earn.
• You don't need to hire editors. CapCut for PC is free. Sacrifice 1–2 streaming days a week for editing — you can do a week of shorts in 2 days.
• Super Chats: require Partner Program. You can use your own donation link in description (Streamlabs / SE / Ko-fi) but use it sparingly — YouTube doesn't love it long-term.
• Channel memberships: require Partner Program. Monthly recurring income.

Twitch monetization.

• Affiliate (30-day): 50 followers · 8 hours streamed · 7 unique stream days · avg 3 concurrent viewers → unlocks subs, donations, bits, ads
• Partner: avg 75+ viewers, 6 streams on 6 unique days for 2 months running → higher ad splits, better discoverability
• Twitch's discoverability is awful — bring people from other platforms via clips
• Smart monetization: channel point → paid upgrades (subs earn faster), sound alerts, model upgrade goals (toggles + outfits), goal-based support (new emotes, assets, overlays, wallpapers), subathons / donothons with lots of goals + punishments. Avoid: aggressive donation goals, guilt-based messaging, over-monetizing early.

Twitter/X monetization. Twitter Premium + 500 followers + 5M impressions in 3 months + verified email/phone + Stripe + supported country. Don't buy Premium under 500 followers — Premium does NOT help you grow. Ads pay you via comments on your posts — posts that drive comments (questions, polls, "your 3rd saved pic is a reaction to this") earn the most.

TikTok Live. 1,000 followers · 18+. Great way to earn from gifts AND funnel viewers to your Twitch. Streamlabs' multistream plugin for OBS lets you push to Twitch + TikTok simultaneously — set it once and forget it. Setup walkthrough: YouTube.

Getting sponsors — where to start:

• Audit your socials first — sponsors will look before they respond
• Remove or avoid: sad posting, drama, heavy personal venting on your main
• Keep branding consistent across ALL platforms — everything should look like it came from the same character
• Make sure your email and a link to your media kit are easy to find
• A Carrd or Linktree is the minimum — a full media kit is better

Your media kit should include:

• Your creator name + region
• Follower counts across all platforms
• Average viewership + total views
• Engagement rate
• Audience age range
• Content genres + popular videos
• Past collaborations and partnerships
• Current and past sponsors

Writing your sponsor letter:

• Treat it like a cover letter — sell yourself in the best light
• Include: who you are, what platforms you're on, your content type, your best stats
• Add a personal story related to the brand — makes you memorable
• Keep a generic template and customize per sponsor
• Worst they can say is no — doors reopen

Where to find sponsors:

• Twitter — keep an eye on new game releases and partner programs
• Publisher partnerships — Bandai Namco, Capcom, etc. all have creator programs
• Lurkit — sign up for game keys + paid sponsorship opportunities (lurkit.gg)
• Affiliate portals — most brands have them at the bottom of their website
• Streamelements + DM brands you like directly
• This community! (Eggie's Discord)
• Don't wait to be approached — pitch first, pitch often

Usually around 1K–10K followers needed — brands care about engagement more than follower count.

Spotting sponsor scams in your inbox (most cold emails are scams):

• Numbers or random symbols in the sender's email username → bot
• No business card at the bottom of the email (no website + contact number)
• Real brands won't contact you more than twice — scammers email you over and over
• Demands that sound impossible ("convince 100 viewers to register and get $50") → scam
• Wait a few days before responding — real brands won't desperately ping you

Small channel? You still qualify. CCV is not the only metric that matters. A small but engaged audience is valuable to the right brand. You can guest at conventions, land collabs, and get sponsors at any size. The difference between those who get opportunities and those who don't? They applied. That's it. Apply.

Direct support platforms.

• Ko-fi.com — no requirements · one-time donations, monthly memberships, commissions, shop sales · set donation goals ("For new microphone")
• Patreon — no requirements · monthly subscriptions · great for exclusive art (incl. NSFW if that's your thing), behind-the-scenes, art WIPs, polls, ASMR audios. Be active there so members feel the sub is worth it.

Merchandise.

• Physical merch platforms: Fourthwall · UwUmarket · MosoBox · others. You can DM them — most will say yes if you already have art
• Commercial rights matter: never use art you commissioned for merch unless you bought commercial rights from the artist (or buy them retroactively — most artists offer this)
• Fresh, unseen art works best in drops — adds surprise to the announcement
• Plan drops around a debut, subathon, or other big event — never drop without a teaser
• For first drops, partner with UwUmarket or MosoBox (they offer lots of help). Lean toward your own shop later.

Digital products: Gumroad · Ko-fi · Etsy · Vgen. Guides, stream assets, wallpapers, voice packs, chat widgets — almost anything can be a digital product.

• Donation goals work because humans love seeing a progress bar fill. Tie the goal to something concrete ("for new emotes").
• Subathons: viewers sub/donate, the stream timer increases, when it runs out — stream's over. Don't do them more than once a month or you'll look greedy. Cute timer widgets on Ko-fi / Etsy / Vgen, both paid and free.

Common monetization mistakes:

• Monetizing too early — you need viewers first. Build the fanbase, then add income streams slowly.
• Overloading viewers with links — don't spam. One clean panel/socials section is enough; people who want to support will find it.
• Ignoring taxes — VTubing is a job. If you make profit, you likely owe tax. Talk to an accountant. Don't commit tax fraud
• Not using vtuber assets as tax write-offs — if streaming is your job you can write off your equipment, software, and even convention trips (TwitchCon, OffKai). Keep your receipts.

Legal hygiene:

• Track your income in a spreadsheet — income, expenses, taxes. End-of-year you. will. thank. you.
• Don't spend everything immediately — save for taxes + sub refunds + slow months
• Follow each platform's TOS — small streamers get ignored by support if you get banned
• Disclose sponsored content — in some countries (Poland, US) it's illegal not to
• Set up a separate email + PayPal for VTuber business — avoid accidental doxxing
• This is general info — talk to an accountant for your specific country

Why Fansly works well for VTubers. Multiple tiers, working FYP/discovery, ability to stream natively, Lovense integration for interactive streams, and very real income — top creators clear $2K–$10K/month. Variance is huge — most creators don't crack $150/month — but the ones who follow the playbook scale fast.

Branding. Stand out with design AND personality. Brat, submissive, dommy mommy — pick a lane you're actually comfortable with. Some creators run a totally separate persona/alt for NSFW.

• PFP: clean, doesn't need to be lewd
• Banner: spicier — full-body illustration or IRL pic. Avoid plain logo banners (looks cheap)
• Bio: spell out exactly what subs get

Subscriber tiers. 1–10+ tiers possible. Example setup:

• Tier 1 ($10) — VODs + DMs
• Tier 2 ($15) — adds NSFW audios, vtuber lewd drawings/pictures/videos (VRChat), DM priority
• Tier 3 ($25) — adds IRL lewd pictures/videos
• Tier 4 / 5 / 6 ($50 / $99 / $199) — extra support tiers, optional, mostly for fun (top tier may include findom in DMs)

2–3 tiers is plenty to start. Don't price below $10 — cheaper tiers attract worse people. Never leave DMs fully open — gate them behind even a $5 tier so you only hear from people who paid.

Promotions exist but save them for big moments (Christmas, Black Friday, summer, birthday) so people don't expect them. Set everything up under Creator Dashboard → Plans/Promotions. Every post needs access permissions set ("requires subscription" → minimum tier). Public videos can hit the FYP, gated ones can't.

Automated messages. Set up DIFFERENT auto-messages for new followers vs new subs (under Creator Dashboard → Messages). Intro yourself + link socials in the follower message. Thank + something cute in the sub message. Follow some other VTubers to see what their messages look like and adapt yours.

Pinned messages. Up to 3 pins — use all 3. Most VTubers pin a tip menu, intro, and a teaser clip. Tip menu = fan service: custom IRL pictures, dpic ratings in DMs, custom audios, etc. Make a graphic with prices. Minimum $10 per item, ideally $20+. Also pin your rules — if you don't show nudity (totally valid), make that visible to everyone.

Streaming on Fansly. The biggest discovery driver. Browse displays streams on the main page; viewer count doesn't decide ranking (you'll see 2-viewer streamers next to 700-viewer ones).

• Audio is the draw — moaning, breathing, ASMR. Only do what you're comfortable with
• Lewd content: Fansly is partnered with Lovense — viewers can tip to control your toys (speed, pattern). Lush 4 / Lush Mini, Domi 2, Nora, sex machines are common
• Non-sexual lewd: nude VTuber model + games in background — basically your Twitch stream but spicier. Works best if you already have an audience to bring over
• VRChat content: VR x IRL where you move your 3D VRChat model live. For interactive content you'll want a model with SPS (socket penetration software) — commission someone for the first build. Cheap base models on Booth
• Interactive redeems if you don't have a toy: "$10 — choose my toy" done manually, plus a $5–$25 spinning wheel with prizes like "control my toy for 5 min" or "lewd pic in DM after stream". No empty or losing options on the wheel — that becomes legally classified as gambling. You can include punishments for YOURSELF, just not for viewers.
• VODs: Fansly doesn't store stream VODs — record them yourself and upload as videos later. Upload same day to free up your drive. 2–3 hour VODs take 30–60 min to upload.

The FYP. Fansly's FYP works extremely well — this is the ONE place hashtags actually help. Tag everything specifically (#blowjob, etc.) AND #vtuber. To hit the FYP your video must be public (visible free to everyone on your wall too). Locked videos with a free preview can also be posted but reach is weaker due to lower watch time.

• Shorts: 6–20 seconds
• Can repurpose stream clips — pick parts with good audio (moans, breathing, spicy lines)
• Edit VODs into 15–30 min digest videos and sell for $1–5 — subs can either watch the full VOD (free with sub) or grab the TLDR for a small tip. Adds up fast.
• Tease, don't give it all away — cut 80% of the content, leave 20% as the payoff. Keeps subs craving more

DMs. Respond to every DM from subs / paid-per-message senders — except the harassing or boundary-crossing ones (call out + block). Most subbed to build a deeper connection, so be nice, chat casually, and cut sexual topics short if you don't want them (and lean in if you do!). Girlfriend-experience roleplay is the top-performing niche among VTubers on Fansly if it's within your comfort zone.

Growing Fansly using outside platforms. Outside reach speeds things up massively but isn't required (creators do grow from zero on Fansly itself).

• Keep Fansly content OFF your main accounts. No NSFW on main — risks permanent shadowban
• Make a dedicated NSFW Twitter/X alt, occasionally redirect from main
• On the alt: post spicy IRL pics (censored = Fansly exclusive), NSFW drawings of your model, audios, and ALL your Fansly FYP videos — doubles reach + brings fast subs
• Stay consistent — at least every 2 days, ideally 2x/day. Don't binge-post then ghost for a week
• Always announce when you're going live on Fansly

Content ideas for subs: IRL nudes (pics/videos) · lewd audios (moaning, blowjob sounds) · ASMR roleplay (SFW + NSFW) · voice packs · NSFW art (made or commissioned) · VODs · text posts, polls, tip goals · IRL tease pics/videos (no nudity) · pics of new toys with polls.

Realistic numbers. Most Fansly creators don't break $150/month — that's the cutoff between "trying" and "actually executing the playbook." Once past it, growth snowballs. Realistic example: ~$1,688 in 30 days · ~$2,011 in a strong month · ~$520 in a slow one with 90 subs and 7 streams. One stream might make $350, the next $10 — variance is real. Top creators clear $2K–$10K/month streaming + posting daily. Stream at night for max browse traffic.

This wing of the playbook is the VTuber Bible — a free, living guide to how growth actually works, platform by platform. No hacks, no "post at 7:03pm" myths: each chapter explains how the platform's algorithm thinks, what to post, how often, and the one metric that matters — then the later chapters cover clips, branding, community, money, sponsorships and the mindset to survive it all.

• Take it as a class: the launcher at the top of this page walks all 19 chapters in order, one lesson at a time, with your progress saved in this browser.
• Or raid it as a reference: every chapter stands alone — jump straight to the platform you're fighting with today.
• Tools that pair with it: the Optimizer (titles, tags & captions), the Thumbnail Checker, the Rate Calculator + Pitch Builder (the Sponsorships chapter, done for you), and the Debut Checklist for the big day.

1 · Every algorithm optimizes the same thing

YouTube, TikTok, Instagram and X are all trying to answer one question: "if I show this to more people, will they stay on my app longer?" Every metric — retention, completion, shares, session time — is a proxy for that. The master skill is making things people finish and share; every platform section below is just the local dialect of that rule.

2 · The hook is 80% of the outcome

People decide in 1–2 seconds (shorts) or on the title+thumbnail alone (long-form) whether to give you a chance. A mediocre video with a great hook outperforms a great video with a weak one — every time. Write the hook first, then make the content that pays it off.

Hook shapes that keep working: open a curiosity gap, start mid-action, make a claim that begs pushback, or promise a payoff with a visible timer on it.

3 · Retention beats reach

The algorithm shows your video to a small test batch. What that batch does decides everything after. You earn the second wave with watch-through — cut every second that doesn't earn its place; the delete key is your best growth tool.

4 · Consistency beats intensity

Ten videos a week for two weeks then burnout loses to a gentler pace held for a year, permanently. Algorithms reward accounts that reliably feed them; audiences bond to rhythms. Pick a cadence you can hold on your worst week, not your best — sustainable beats ideal, always.

5 · Twitch is home, everything else is a flyer

Live streams barely get discovered — Twitch's directory shows big streams first. Growth flows the other way: shorts/clips/TikToks find strangers → X/Discord turn them into friends → Twitch is where friends hang out. Treat every clip as a flyer for the stream, and the stream as the factory that makes clips.

6 · 80/20 value & personality

Around 80% of posts should give a stranger something (a laugh, a feeling, a fact); about 20% can be personal/promo. Streamers who only post "going live!" grow at exactly the speed of their existing audience: zero.

7 · Volume is how you learn

Your first 100 shorts are tuition. Nobody can predict virality — winners are found by shipping many honest attempts and reading the data, not by polishing one "perfect" video. Quantity, with feedback, becomes quality — at whatever volume your energy actually allows.

How the algorithm thinks

YouTube is a recommendation engine, not a subscription feed — most views come from Browse (home page) and Suggested (sidebar). It tests your video on a slice of likely-interested viewers and watches two numbers: CTR (did they click?) and average view duration (did they stay?). Strong both → bigger test batches, for months or years. Long-form is the only format with a real back catalog.

2026 update: YouTube now reads the video itself (frames and audio) to decide what it is about and who to show it to. Your title, description and tags confirm what it already saw; they no longer define it. Metadata that disagrees with the footage reads as a contradiction, and contradictions are exactly what the clickbait filter looks for. Full chapter: The 2026 shift.

Packaging-first workflow

Decide the title + thumbnail before you record. If you can't write a title you'd click, the video isn't ready to make. This is the single biggest workflow upgrade most creators ever make.

• Titles: under ~55 characters, front-load the interesting part, one clear idea. Curiosity gap > keyword stuffing.
• Write the title as the answer to a real question someone would type or ask ("what class to pick in Guild Wars 2", not "GW2 class guide"). Question-shaped searches carry a fraction of the competition of the head keyword. Keep it human, keep the hook.
• Thumbnails: readable at postage-stamp size — one face/emotion + max 3 words + high contrast. Your model IS your brand asset; use expressions hard.
• The title and thumbnail should each say something different — together they make the question the video answers.

Score titles against real keyword/competition data before you ship, not just gut feel.

Structure that holds retention

• First 30s: confirm the click ("yes, this is the video the thumbnail promised"), then open the first loop. No branded intros.
• Middle: chapter the video as mini-payoffs every 60–90s. Every time a question resolves, open the next one first.
• One clear subject per block. Ask YouTube can now drop a viewer straight into the segment that answers their question, so a video that is cleanly divided beats a ramble on structure alone. Design a few segments strong enough to stand on their own as a Short.
• Descriptive chapter names. "Method 1 / Method 2" starves YouTube of the info it uses to place your video. Name the thing: "Gold farming with the Drizzlewood meta". This is a reversal of the old "don't spoil the chapter" advice.
• Say the subject out loud. The exact game, the exact boss, the exact product, early and whenever it changes. "This game" and "that thing" leave the model guessing. Do not overdo it; natural beats only, retention still rules.
• End: don't wind down — hand momentum to a specific next video.

Watch your retention graph: a cliff in the first 30s is your intro's fault; mid-video dips mark segments to cut in future edits.

What the numbers mean

• CTR 4–10% is healthy (small channels often higher — your audience is warm). Under 3% = repackage; you can swap title/thumb on a live video, and should.
• AVD 40–50%+ of runtime is strong for gaming/cozy content.
• Browse/Suggested share rising = the algorithm has adopted the video. That's the win condition.
• Subscribers are a lagging vanity metric — views-per-video and returning viewers tell the truth.

What actually changed

For years YouTube worked out what a video was about from what you typed: title, tags, hashtags, description. In 2026 it switched to reading the video itself. Gemini watches the frames and listens to the audio and builds its own idea of what you made and who would want it. Your metadata still matters, but its job flipped: it confirms what the model saw. When your title says one thing and the footage shows another, that is no longer neutral, it is a mismatch, and mismatch is what the clickbait filter hunts for.

Two new ways to be found came with it:

• Ask YouTube: viewers type a full question or describe a vibe instead of a keyword. YouTube answers with videos and starts playback at the timestamp that answers the question, not at 0:00.
• Custom Feeds: a viewer describes what they want ("cozy GW2 streams with a chatty VTuber, no screaming") and YouTube builds a home-page feed from that prompt.

Why this is good news for a small channel: on the old search page a broad question returned the biggest, most generic videos. On a prompt surface, the video that matches the prompt exactly wins. Being specific used to cost you reach. Now it qualifies you.

Keyword SEO ➜ prompt SEO (with real numbers)

This is not "title everything as a question." It is: stop planning a video around one big keyword and start listing the set of real questions the video honestly answers. Then title, chapter and script for those.

Search term (vidIQ, Sept 2026)	Searches / month	Competition	Opportunity
guild wars 2 (head term)	314,348	44.5	71.4
what class to pick in guild wars 2	5,224	7.6	70.3
how to craft in guild wars 2	5,381	8.6	70.0
how to get mesmer shield warframe	9,635	7.4	72.7
how to rank up entrati warframe	9,082	8.0	72.3

Read the last row again: the question out-scores the head term on opportunity while carrying one sixth of its competition. You do not need to beat 314k-search competition any more. You need to be the exact match for the 5k-search question, and every game has dozens of those.

These terms are climbing fast. how to get xaku warframe is +164% vs its 30-day baseline. how to get gauss warframe is +84.5%. Question-shaped searches in your niche are not static; they are growing, which means the competition gap widens even further in your favor if you act now.

• Before you plan a video, pull the question-shaped searches for the game or topic (vidIQ keyword tool ➜ questions), and also pull long-tail variants (vidIQ ➜ matching_terms). Sort by opportunity, not raw volume.
• Prefer terms with competition under 20. Note any that are climbing.
• Your hook rules do not change. The title still needs to make a human click; prompt-matching only narrows which words you pick.

Build in segments (Ask YouTube serves timestamps)

• Give every block of the video one clear subject. If the whole thing is one long undifferentiated ramble, YouTube reads it as less specific and serves a competitor whose video is cleanly divided. You lose on structure, not quality.
• Plan long-form and Shorts together: make specific segments strong enough to stand alone as a Short. One shoot, two discovery surfaces.
• A returning-player GW2 video is not "my GW2 thoughts." It is: the build I picked and why ➜ the first hour ➜ what changed since 2023 ➜ is it worth coming back. Four subjects, four chapters, four possible Shorts.

Chapters: name the thing (this reverses old advice)

The old tip was to keep chapter names vague ("Method 1", "The mistake") so you would not spoil the payoff. That is now actively harmful: it hides the information YouTube uses to decide who to show the video to.

• ❌ 0:00 Intro · 2:10 Method 1 · 6:40 Method 2 · 11:00 The mistake
• ✅ 0:00 Why I came back to Guild Wars 2 · 2:10 Gold farming with the Drizzlewood meta · 6:40 Why Mesmer clones break the rotation · 11:00 The one mistake that wasted my first week

Write chapter labels as specific, descriptive phrases a person might search. Still needs the first timestamp at exactly 0:00 and at least three chapters or YouTube will not turn them on.

Say the subject out loud

Gemini transcribes you. Calling your subject "this", "that" or "it" leaves it guessing what the video is about.

• ✅ "Where Winds Meet just changed how the grapple works"
• ❌ "so this game just changed how this works"
• Name the exact game, boss, product, place. Say it early, at natural beats, and whenever the subject changes.
• Do not overcorrect. Narrating like a robot wrecks retention, and retention still decides whether you rank at all. Make it easy for the model, not exhausting for the viewer.

Stop doing / keep doing

Stop:

• Treating the tag field as a ranking lever. It is close to dead. Fill it in 30 seconds for genuinely ambiguous names, then move on.
• Keyword-stuffing descriptions. A description that contradicts the footage now works against you.
• Vague chapter names, and planning only around the one head keyword.

Keep (none of this changed): CTR, average view duration, the first 30 seconds, thumbnail contrast, a title a human wants to click, and satisfying the viewer once they arrive. Gemini changed how you get matched to an audience. It did not cancel having to hold that audience. Any tip here that would cost you retention is being applied wrong.

What is actually live (read before you panic)

Claim	Status, Sept 2026
YouTube reads the video; typed metadata is demoted	Reported and consistent with the clickbait-matching rollout. Safe to act on.
Ask YouTube exists and starts playback at the matching timestamp	Confirmed (Google I/O 2026).
Ask YouTube is available to everyone	No. US only, Premium, 18+, opt-in. Wider rollout is "planned".
Custom Feeds exist	Confirmed, rolling out through 2026.
Custom Feeds are available to everyone	Partly. US, English only, needs watch and search history on, one feed per account, expires after 30 idle days.
"This is how most people will find videos soon"	A forecast, not a fact. Classic search, home and subscriptions still carry most views today.

How to act on that: the chapter, segment, naming and prompt-title changes are free. They help human viewers too and they pay off on classic search right now, so adopt them this week. What you should not do is tear up a working strategy because someone said search is dead. This is additive, not a replacement. Run the Optimizer on your description: it now flags vague chapter names.

How the shorts feed thinks

Shorts is a swipe feed: your short is shown, and the only vote is swipe away or keep watching. The metrics that matter are viewed vs swiped (first seconds) and % watched — over 90% is strong, over 100% (loops) is rocket fuel. Likes/comments correlate but don't drive. Since 2026 a Short is also matched by what is said and shown in it, so a specific subject stated aloud is discovery, not decoration.

Anatomy of a short that works

• 0–1s: the frame itself hooks — motion, a face mid-emotion, or text that starts a question. Never a fade-in, never a logo.
• 1–5s: establish the situation in one sentence. Captions on by default — most viewers watch muted-ish. Name the game or subject in that sentence (2026: YouTube listens, and "this game" tells it nothing).
• middle: one idea only. Cut every breath. 20–35s is the sweet spot for completion.
• last 2s: land the punchline at the end so completion hits 100% — or loop the ending into the opening line.

Shorts → subscribers reality check

Shorts subscribers convert to long-form viewers at a low rate — that's normal, not failure. Their real jobs: algorithmic warmup, a funnel to Twitch/Discord via pinned comments, and reps that teach you hooks. The pipeline: stream → clip → short → pinned comment → Discord → Twitch regular.

Posting rhythm

3–7 per week, roughly one slot per day. Don't dump 5 at once — the feed tests them against each other. When a short pops, make three more with the same skeleton and different meat.

How the FYP thinks

TikTok cares about your account history less than any other platform — each video is tested cold on a batch of a few hundred viewers. It reads completion rate, rewatches, shares, and comments (roughly in that order) and escalates in waves. This is why TikTok can 100× a nobody overnight, and why one flop means nothing.

What works for VTubers

• Same clips as Shorts/Reels — but TikTok rewards rawness; overly-produced content can underperform.
• Trend-jacking: take a trending sound/format and bend it into your niche within 1–3 days of it rising. Trends are free test batches.
• TikTok SEO is real: say your keywords out loud, put them in captions and on-screen text ("VTuber", the game's name). Search is a growing share of views.
• Reply-to-comment videos mint easy content and boost the original.

Rules of the road

• Volume is safe here: 1–3/day doesn't fatigue the FYP — but only at a pace you can sustain.
• Never delete underperformers (they sometimes revive; deletion can hurt account signals). Set to private if embarrassed.
• New accounts: post normally for 1–2 weeks before judging anything — the system is still fingerprinting you.
• Watching your niche's videos as your brand account trains the algorithm on who your peers are.

How Reels ranks

Instagram's #1 ranking signal for reach is sends per reach — how often viewers DM your reel to a friend. Saves and shares outweigh likes massively. The question for every reel isn't "is this good?" but "who would send this to someone?" Relatable > impressive.

VTuber playbook

• Same vertical clips as TikTok/Shorts — pick the ones with a feeling (wholesome, unhinged, "literally me").
• Trial Reels (shown only to non-followers first) are free A/B tests — use them for experimental hooks.
• Collab posts put one reel on both accounts' grids — the cheapest collab format that exists; do these with VTuber friends monthly.
• Stories are for people who already follow you: polls, streams, behind-the-scenes. They drive retention, not reach.
• Hashtags are minor now — 3–5 niche tags max; the content itself is the targeting.

How the 'For You' algorithm actually works

X open-sourced its ranking stack in 2026 (xai-org/x-algorithm) — so this isn't folklore any more. Your post is drawn from two pools: in-network (people who follow you, served fast by a system called "Thunder") and out-of-network (strangers — found by an ML model, "Phoenix", plus interest-cluster matching, "SimClusters"). Everything then hits one transformer that predicts ~20 things you might do — reply, repost, quote, like, profile-click, dwell, watch a video, follow — and ranks by the weighted sum of those odds. A diversity pass ("VMRanker") stops any one account hogging a feed. Translation: reach is downstream of predicted engagement, not hashtags or post time.

What it rewards — and what quietly buries you

• Heaviest positive signals: replies, reposts & quotes, then likes, then profile/post clicks, then dwell time (attention is scored even with no tap — a thumb-stopping first line counts) and video views, then new follows off the post.
• Out-of-network discount: shown to a stranger, your score is multiplied down — so a stranger reacting counts extra, and being reshared by a followed account drops you into their in-network pool with no discount. That's the mechanical reason the reply game works.
• New-author boost: newer / low-impression accounts get a deliberate lift toward being seen — the system wants to find you, so early consistent posting matters.
• Author-diversity decay: your 2nd, 3rd, 4th post in a row to the same viewer each scores lower — don't fire off five near-identical posts; space and vary them.
• Negatives cut twice: "not interested", mutes, blocks & reports subtract from your score AND feed a safety layer (spam + credibility models like Agatha & user-cred). Rage-bait that farms blocks is a reach tax, not a hack.

How small VTuber accounts actually grow here

X is less a discovery engine than the community layer — art, collabs, memes, industry friendships — and the open-source stack shows why the reply game is the #1 lever: engagement from an established account pulls your post into that account's in-network pool (no out-of-network discount), and their audience is your closest-matching interest cluster.

• Reply game: 5–10 genuinely warm/funny replies a day to bigger accounts in your scene — your best joke under a 50k account outreaches your whole timeline. You grow in other people's replies, not your own posts.
• First hour is the test batch: reply to every early reply fast — conversation velocity is exactly what the model measures.
• Media > text, always: native-uploaded clips/art (with captions) earn the video-view + dwell signal plain text can't. Never just link out — off-platform clicks don't feed the good signals.
• Ask, don't announce: a question or a take that begs a reply beats "going live!" — replies are the heaviest action.
• Pin your best intro clip — profile-clicks are a scored signal and visitors decide in seconds.

The daily rhythm that compounds

• 3–5 posts/day, varied (a clip, a shitpost/relatable, model/art, a genuine reply thread) — variety dodges diversity-decay and reaches different clusters.
• Art RTs + tagging artists builds the relationships every VTuber depends on — and each reshare is an in-network jump.
• Build credibility, not just posts: the stack scores account trustworthiness continuously (spam / authenticity / PageRank-style models) — consistent, real, non-spammy engagement is a long-game ranking asset; pods, follow-for-follow and bots are not.
• Protect your peace: chasing every post's numbers is the fast lane to burnout — post honestly, reply warmly, let the compounding happen.

Draft a post against this, or turn any section into this week's posting plan — and run video titles through the Optimizer before you ship them.

The uncomfortable truth

Twitch has almost no discovery: the directory sorts by viewers, so browsers see big streams first. Expecting Twitch itself to grow you is the #1 rookie trap. Twitch's job is retention — turning people your clips found into regulars, and regulars into community.

What actually grows a Twitch channel

• Clips exported everywhere — the stream is a content factory.
• Raids & being raided: raid a similar-size streamer you genuinely like after every stream; it starts reciprocal relationships. Networks of small streamers grow together.
• Collabs: a shared stream with a peer merges two audiences with zero algorithm involved.
• Category strategy: in a giant game you're page 40; in a mid-size cozy/story game you can be top-10 in category. Cozy and story-game categories are friendly ponds.
• Schedule: the same weekly slots — as a flexible template, not a lock. Appointment viewing is how casuals become regulars; a moved stream with a heads-up keeps the trust.

Craft of the stream itself

• Talk through silence — narrate decisions; lurkers hear dead air as "nothing happening".
• Greet new chatters by name once; don't interrogate.
• First 30 seconds a new viewer sees should explain the stream: a good title, overlay context, energy.
• End every stream with a raid + a "we're in Discord" call.

Why it matters

Every platform above rents you your audience — Discord is the only place you own the relationship. A 100-member Discord that's alive beats 1,000 dead. It's also your safety net when an algorithm turns cold.

Running it well

• Fewer channels than you think: general, clips/memes, stream-pings, fan-art. Empty channels read as a dead server; merge until they're busy.
• Be present in general as yourself — 10 minutes of genuine chat is worth more than any announcement.
• Events make members: movie nights, community game nights, watch parties.
• Ping discipline: @everyone for streams/big news only — trust is spent in pings.
• Give regulars small roles/recognition; being personally seen is the entire retention mechanic.

The multiplier pipeline

Highest-leverage system: every stream produces 3–5 clippable moments → each becomes one vertical Short (15–45s, 9:16) → post to YouTube Shorts + TikTok + Instagram Reels. One night of streaming = 15+ discovery pieces across three platforms. Made once, posted everywhere with the same title.

What makes a moment clippable

• Emotion spikes: laughter, screams, stunned silence, soft wholesome beats.
• Instant context: a stranger must get it in 2 seconds without knowing you. If it needs setup, add a one-line caption instead.
• Reactions > gameplay: your model reacting sells harder than the play itself.
• Fails, wholesome moments, "wait WHAT" reveals, chat roasting you — the eternal four.

The workflow

• Captions always; assume muted viewing. Big, clean, 2–4 words on screen at a time.
• Reframe to vertical with the model+action in frame; punch in on the reaction.
• Title the clip like a short (curiosity gap), not like an archive ("Stream highlight #23" is a tombstone). Get the game name in the title or caption; a hook with no subject cannot match anyone's search.
• Post the same clip everywhere — audiences barely overlap; "cross-posting shame" is a myth.

The 10-second pitch

Every discovery surface gives you seconds. A one-line identity a stranger can repeat — [your vibe] + [what you do] + [the feeling you give], e.g. "cozy gacha streamer — story reactions, cover songs, and the softest corner of the internet" — is what makes people able to recommend you. Fill that formula with your character and make it visible within seconds everywhere: bios, pinned posts, banner, overlay.

Consistency toolkit

• Same handle, palette, and face everywhere — recognition compounds across platforms.
• Lore should flavor content, not gate it (nobody does homework to enjoy a clip).
• Emotes/overlays/alerts in your palette make screenshots self-branding.
• Your model's expressions are brand assets — thumbnail packs of strong expressions pay for themselves.

Debuts & redebuts

A debut is a milestone for the existing community, not a growth event — strangers don't attend debuts. Redebuts/model reveals ARE clippable hype moments though: milk them across every platform for a week of content. Planning one? The Debut Checklist walks the whole big day, phase by phase.

How VTuber networking actually works

• Give first, for weeks: genuine replies, clip THEIR moments, show up in their chat. Every "collab?" DM from a stranger is spam; the same ask from a familiar friendly face is a yes.
• Collab ladder: aim within ~½×–3× your size. Peers rise together — your gen of small VTubers is the network that matters in 2 years.
• Community events, game nights and group collabs are wholesale networking — one event = ten relationships.
• Raid trains and recurring collab formats compound harder than one-offs.

The monetization playbook — a strategy-call recap

Field notes from a creator monetization call (VTuber Xena, figures verified by the VOX team): she pulls strong revenue at just 30–50 concurrent viewers — proof that community depth + diversified platforms beats raw size. Her verified stack: ~40K YouTube subs, ~10% Patreon conversion, $2–3K/month on Twitch despite inconsistent streaming, ~$150K lifetime on Throne, now building toward an Amazon Prime deal and an animation studio powered by VTuber models. The blocks below are her framework.

Your community tone IS the revenue engine

• Treat your community like a garden — what you allow to grow is what you harvest. Keep the vibe consistently positive; don't feed negativity or "sad fishing".
• If someone drops heavy content, gently redirect them to real resources (e.g. Stack Up, a mental-health charity) — you're a host, not anyone's therapist.
• Publicly reward the behaviour you want more of, and train trusted regulars/mods to redirect negativity so the space self-filters over time.
• Every "bad egg" you tolerate quietly costs you 9–10 generous fans who leave without a word. Mindset: "if you don't run your community, your community runs you."

Low-effort connection converts (free → paid)

• Keep a living list of under-5-minute engagement moves that don't drain you: 5-second personalised Discord voice notes, good-morning messages, emoji replies to 50 people, kiss-print fan gifts.
• Meet fans where they already are — some won't move platforms, so go to them. Jenna Marbles / Taylor Swift energy: two seconds for you can be the whole world for a fan.
• Include free engagement too — free converts to paid over time and honours the fans who can't pay yet.
• You do not owe every DM a reply — you have permission to ignore and set boundaries so this stays sustainable.

Diversify off Twitch/YouTube (before you're forced to)

Platforms can and DO demonetise creators overnight — always keep a presence on Patreon, Fansly, Fourth Wall, Throne so no single switch can zero your income.

• Patreon: post 2–3×/week (always include some free posts to re-engage lapsed patrons); cap at 3 tiers max to beat decision fatigue (e.g. $1 / $5–10 / $100); use free-tier notifications to remind former patrons you exist.
• Throne safety: set payouts to cash, not physical gifts, so you never accidentally leak your home address via Amazon gift accounts.
• Fourth Wall is well-reviewed (responsive, VP-level engagement). Avoid over-discounting — constant sales train fans to wait instead of paying full price.
• Prefer many small supporters over one whale — diversified income is healthier and sustainable. And shout your platforms out verbally on stream, not just via bots — personal mentions convert far better.

Share your goals publicly (fans fund the dream)

• Talk openly about the long-term dream (animation studio, a farm, a movie) — it aligns your audience and motivates real financial support.
• Give fans ownership: let them name a character or a street in your story, add their name to a thumbnail, or list them in the credits. Invested fans evangelise you into other communities for free.
• Don't force the goal — give yourself up to a month; even just talking about your interests on stream lets the right goal emerge.
• Real example: Xena talked about animation on stream, a Tubi employee turned up in her chat, and that connection became her FilmHub deal. Say the dream out loud and the room helps build it.

FilmHub & Tubi — the long-form licensing play

FilmHub is a free platform where you upload long-form content and streaming services (HBO, Showtime, Netflix, Hulu, Tubi…) can license it. Tubi is investing hard to rival YouTube and is hungry for animated content right now — VTuber-model content fits it perfectly.

• Payout is currently ≈ $12 per 1,000 views on a single video (incentive-based; may normalise toward ~$0.75 in a few years), and Tubi runs on Roku — so you reach the big screen beyond YouTube.
• Trigger: a video must clear 1,000 views to earn meaningfully — so drive eyes deliberately.
• Strategy: use YouTube react content to pull attention → publish original character/series content ~once a week → push it to Tubi → build toward an exclusive series deal.

Give back — generosity compounds

• When you're in a season of abundance, redirect your audience to support other creators or charities — reciprocity comes back around.
• Tell fans to save donations for upcoming big events so giving feels like being part of something, not money thrown into the void.
• Audiences who watch you be genuinely generous are far more likely to rally hard for you when you actually need it.

The realistic stack

• Direct support first: Twitch subs/bits, memberships, Ko-fi — powered by community closeness, not raw reach.
• Sponsors: reachable earlier than you think — indie games (especially cozy/gacha adjacent), VTuber-adjacent brands. Micro-influencer budgets are real; an engaged small stream is worth more than a dead big one.
• Affiliate codes tick along passively — put them in every description.
• Merch works after the community asks for it, not before.

The Rate Calculator answers "what are your rates?" so you never have to guess on the spot.

Rate math that protects you

• Baseline short-form sponsor rate ≈ your average views × $25–50 CPM ÷ 1000, then adjust for niche fit and usage rights.
• Charge more for: usage rights, exclusivity, tight deadlines, multi-platform bundles.
• Everything in writing — deliverables, dates, payment terms (50% upfront for new partners).
• It's always okay to say no to brands that don't fit — one off-brand shill costs more trust than the check pays.

Find deals on X / Twitter (the free pipeline)

Brands, studios and sponsorship managers post open calls on X constantly. Search is a direct pipeline — no marketplace middleman, no follower minimum. Paste these into X search (swap in your game/niche):

• sponsoring streamers · looking for streamers / looking for content creators
• paid partnership gaming · sponsored stream [your game]
• game key creators / creator codes · now recruiting OR open campaign streamers · PR list OR creator program [game/brand]

Filters that cut the noise: min_faves:5 (skips spam) · -is:retweet (originals only, so you reach the actual poster) · since: a recent date (calls go stale fast). Save the good queries as bookmarks and re-run them weekly.

Follow so deals come to you: indie/AA studio accounts in your niche (they DM creators before launch), creator-relations managers (bios say "creator partnerships" / "influencer mktg"), marketplace accounts (Lurkit, Woovit, Noiz, PowerSpike — they signal live campaigns on X first), and aggregator accounts that collect "brands looking for creators" calls.

Replying: be fast (early replies get seen), keep it to one short pitch — who you are (VTuber + niche), your best number, why you fit that brand — and drop a media-kit/rate link, not a wall of text. Test the product first when you can; "not a good fit" is a valid, brand-safe answer.

Marketplaces — sign up once, apply to campaigns

• Lurkit (lurkit.com) — game keys + paid campaigns from indie→AAA studios. Very small-creator friendly, huge with VTubers. Paid Quests need just 250 followers + a 90-day-old account — the best paid entry point for a small creator.
• Woovit (woovit.com) — keys + sponsorship campaigns, low barrier — great for building a track record early.
• Noiz (noiz.gg) — gaming sponsorships paying on views/hours streamed.
• PowerSpike (powerspike.gg) — Twitch-focused campaigns, open to all creators (no manager needed).
• StreamElements Sponsorship Marketplace — sponsor offers right inside your stream dashboard.
• Bigger/curated once you have size: Ader, Aggero/Streamforge; general (non-gaming) boards worth a profile: Afluencer, Aspire, #paid.

Start order: Lurkit + Woovit first (portfolio-builders) → Noiz + PowerSpike (real paid) → CPM platforms once your short-form clips pull views. Marketplaces run in parallel with the X searches: X = direct asks + inbound, marketplaces = steady campaign flow.

The deal types (know what you're being offered)

• Deliverable package — the "typical" deal: a fixed bundle, e.g. 1× 2-hour stream segment + 1× IG story + 1× TikTok/reel + 1-month exclusivity. Most brand deals and marketplace campaigns are this.
• Integration — the sponsor woven into your normal content with agreed talking points, not a standalone ad read. If the talking points don't fit your brand, say "not a good fit."
• Ongoing relationship (the goal) — same brand, repeat deals. Retention beats acquisition: keep using their product after the deal so they want you back. One maintained relationship > chasing a new brand every time.
• Passive / behind-the-scenes — money the audience doesn't obviously see: affiliate placements, referral links, Prime.
• Own the brand instead of renting it — a brand deal pays once; owning the product (your merch, your brand) pays forever.

Who brokers it: self-served (you negotiate, you keep 100%) vs manager-sourced (~10% of deals they land you — never a cut of your whole income). If you can only have one rep, pick a manager over an agent, and trial ~3 months before signing.

How marketplaces actually pay (+ where to start)

Benchmark first: gaming CPMs run LOW — roughly $3–$12 per 1,000 views (finance runs $40–$80). Most gaming deals sit at the low end, so real money comes from volume + repeat deals, not any single post.

• Per-view / CPM — you post, verified views are tracked, paid $X per 1,000 up to a budget cap. Scales with views, not follower count.
• Per-hour-watched (CPH) — the livestream model: paid per average-viewer-hour streaming the sponsored game. Rewards consistent live viewership, not clip virality.
• Flat fee — fixed $ for an agreed package (the default). Submit offer → negotiate → deliver → paid on approval.
• Affiliate / rev-share — a code or link, paid a % of sales. Ongoing trickle, no cap, but only pays if your audience buys.
• Product-only / keys — no cash, but the fastest way to build the portfolio that unlocks the paid tiers.

Quick pick — start where you're strongest: strong short-form views → CPM campaigns · consistent live audience → Noiz/PowerSpike (CPH) · small & still building → Lurkit Quests + Woovit/Lurkit keys · loyal buying audience → affiliate codes.

Price yourself — the rate formula

These are negotiation anchors, not fixed prices. Per-deliverable base = (expected views ÷ 1,000) × your CPM. Pick your CPM honestly from the gaming band: still building → $3–$5 · steady engaged small audience → $6–$9 · proven converter / strong short-form → $10–$12+.

Deal price = sum of every deliverable's base × multipliers (they stack): exclusivity ×1.2–1.5 · usage/whitelisting (they run your clip as a paid ad) ×1.5–2.0 · rush turnaround ×1.2 · multi-month bundle ×0.85–0.9 (the small discount that wins repeat volume).

Your floor — never quote under it: hours to produce × what your time is worth (don't work below a fair part-time wage). Product-only floor = the retail value of the key/product, and only when it builds your portfolio — otherwise it's an unpaid "exposure" deal. If a brand's budget is under your floor → cut deliverables, don't cut your rate. Offer them less content, not a cheaper you.

Worked example (small VTuber): shorts ~5,000 views, ~40 live viewers/2h, IG story ~1,000 views, CPM $5 → 1 short $25 + 1 segment $20 + 1 story $5 = $50 base; + 1-month exclusivity ×1.3 ≈ $65 → round up and quote a range: $75–$120. The per-view math is small at small size — that's why flat Quests + repeat deals out-earn per-view for now. Use the formula to know your worth and sanity-check offers.

Red flags & the yes-checklist

Walk away from: unpaid "exposure" deals · cuts of your overall income (reps take ~10% of deals they land, never your whole revenue) · overly long contracts / no-exit clauses · clashing brands or over-broad exclusivity · friends/family as your manager · anyone asking YOU for money or gift cards (scam — block).

Before saying yes: ☐ understand the integration — does it fit your brand? ☐ do the talking points fit? ☐ test the product yourself ☐ read the contract for exactly what they want you to say.

Negotiation lines that work: "What's the budget for this campaign?" (always make them name a number first) · quote a range, never a single figure · anchor to value, not follower count ("my short-form converts — here's a clip that pulled X views") · bundle up, discount down · silence is leverage. You are the brand — they need you more than you need them.

The weekly ritual

Once a week: what were the top 2 pieces of content, what did they share (topic? hook shape? game?), and what does that suggest for next week? That's it. Daily analytics-checking is anxiety, not strategy — the sample sizes are too small to mean anything.

One extra look (2026): in YouTube Studio, open Reach ➜ Traffic source ➜ YouTube search and read the actual phrases people used. If they are questions, that is your next title format.

The double-down rule

When something pops: make 3 more in that shape within two weeks — same skeleton (hook style, format, topic), fresh meat. Winners reveal audience demand; most creators celebrate and move on, which wastes the signal.

90-day experiments

Judge strategies on ~90 days, not videos on 3 days. Commit, ship, then read the trend line. Algorithms need weeks to learn who your content is for; so do you.

Burnout is the #1 killer of channels

Not algorithms — quitting. Nearly every VTuber who "made it" is mostly someone who didn't stop in year one or two. Build the machine so your worst week still ships something: batch clips on good-energy days, keep a buffer of finished shorts, and let a light week be light without shame. Spoon-aware pacing isn't a compromise on this list — it IS the strategy.

Systems over willpower

• Content pillars → formats → calendar: 3–4 repeatable formats (e.g. "gacha story reactions", "cover song night", "cozy hand-cam") remove the daily "what do I post" tax.
• Batch by mode: one clipping session, one posting session — context-switching is the hidden cost.
• Streams are renewable: they make content, community and clips simultaneously. When capacity is low, stream + clip and skip everything else.
• Comparison hygiene: mute your own dashboards on rest days. Other people's month-one virality is survivorship bias in a trench coat.

Field notes from five creators who've done it — Ludwig, Agent00, Valkyrae, Adapt & Cinna (Streamer University, curated by caiuwus). The platform sections above are the how; this is the mindset, money & business underneath.

Ludwig · Stream FOR the video (the content funnel)

Treat every stream as a recording session: pre-plan concepts (no full script), stream 4–6h (longer just dilutes), then one night becomes many assets — clips → Shorts/TikTok/Reels → a long-form YouTube video → the VOD. Add voice-over in the edit so clips feel made, not ripped. A clip's job is to lead people to the real content, not just rack up views.

Ludwig · The daily $1 / $1,000 / $100,000 idea drill

Every day, write three stream ideas — a $1 one, a $1,000 one, and a $100,000 dream. A good idea passes two tests: it works as a YouTube video and it produces good clips. Be different — your uniqueness is the part nobody can copy. (When someone steals your idea, you just became their ceiling.)

Ludwig · Packaging is the #1 YouTube skill

• Packaging = thumbnail + title + first minute. Make your own thumbnails; they grab the click, they're not art pieces.
• Steal title frameworks that work — research similar top videos, be inspired, don't copy 1:1. There's a playbook; reuse what wins.
• The 4-P intro (clickbait is dead — the first minute is everything): Premise/hook (what is this?) → Problem (what do we solve?) → Plug (optional CTA) → Payoff ("stick around to see…").
• Learn your analytics; mix risky swings with tried-and-true.

Ludwig · Collabs, reputation & your team

• Collab priority: ① have fun (makes everyone want a round two) → ② make it easy to say yes (give premise + when + length) → ③ hype/analytics (least important).
• Don't burn a bridge for an out-of-pocket clip — bad business for a moment of maybe-virality.
• More people know you than like you; you can't control your reputation, but it fades, and it isn't what matters — the people who care about you are.
• You're responsible for anyone you hire: respect, communication, scaling care as you grow. You are the sauce — don't lose yourself to the grind.

Agent00 · Short-form first — the follower-tier split

• 0–250k: ~60% short-form / 40% streaming. Stream less, post more; get a shorts editor if you can.
• 250k–750k: 35% shorts · 25% stream · 15% long-form · 10% ops · 5% business · 10% other.
• 750k–1M: 55% streaming · 25% strategy · 10% business · 10% managing — now you earn the right to stream more.
• Costream everything; don't stream exclusively on a platform that isn't paying you for exclusivity.

Agent00 · The short-form hook trio + trend-jacking

Every short needs one: the critical hook (the premise that makes them stay — your face can be the hook), the caption hook (on-screen text they read in), and the curiosity gap — plus real human connection. Trend-jack by searching your niche, filtered by likes + this week/month, and bending the trend to your content fast. Keep a separate research account to stay current.

Agent00 · Sponsorships & "you are a business"

• Get sponsors via: ad platforms (Famebit) · a direct ask · management inbound · developed relationships · proof-of-concept (show the idea works, then circle back).
• Understand the integration + talking points, test the product, and read the contract. Typical package: 1× 2-hr Twitch+YT segment · 1× IG story · 1× TikTok/IG reel · 1-month exclusivity. Keeping one brand beats chasing new ones.
• You own nothing — a policy change can wipe your reach — so build something you do own (Shopify / brand / product; concept → money infra → hire → rollout). You can't sell content, but you can sell a business.

Valkyrae · Deals, reps & red flags

• Plan A is having a Plan B. When you get lucky, get a manager (and maybe an agent). If you can only have one → manager; trial ~3 months before signing.
• Walk away from: unpaid deals · cuts of your overall revenue (reps take ~10% of deals they land, never your whole income) · over-long contracts · clashing brands / exclusivity you'll regret · friends-and-family as reps · no-exit clauses.
• You are the brand — they need you more than you need them. Say no to anything off-brand; always negotiate; know your worth.

Adapt · Adaptability, Plan B & burnout

90% of it is showing up and trying. The three qualities: consistency (show up when you said — it's not frequency), discipline (show up when you don't feel like it), passion (hold quality above clout). Transfer your audience across platforms and accounts so you're portable. Protect mental + physical health — you can't make the craft running on empty — and grab the opportunities most people let pass.

Cinna · Networking 101 (kill the fear)

Stop chasing the biggest creator — mingle with peers at your level; a rising tide raises all ships. Reach out with a real plan (segments you'd both enjoy), never half-asked. The thing that actually holds people back is fear of looking like a "clout chaser" — that mindset is the failure. Get uncomfortable, be proactive, and know your why. (For the girlies: do the content that's actually you; don't mistake people using your looks for views as respect.)

Algorithm + short-form growth:

• How Social Media Algorithms Actually Work
• I Blew Up A New Gaming Channel With YouTube Shorts

Twitch growth:

• How To ACTUALLY Grow On Twitch In 2026
• How To Grow On Twitch FAST (Under 5 Min)
• This ONE Trick Will Make You A Famous Streamer
• Top Advice for Aspiring Streamers

Stream + chat engagement:

• 7 Easy Tips To Get More Chatters On Twitch
• Small Streamers: Do THIS and Viewers Will LOVE You

Conversion + retention:

• Converting TikTok Audience to Twitch
• Full-Time VTuber in 2026 — Content System Workshop
• Every Small VTuber Makes THESE Mistakes

Pre-publish + YouTube optimization:

• 9 Things To Check Before Uploading (so your video doesn't die in 48 hours) — source for the long-form checklist
• VidIQ — How the YouTube Algorithm Works in 2026
• VidIQ — How to Find the Right Time to Publish on YouTube

Monetization + sponsors:

• VTubers Are Doing Sponsorships Wrong — How To Get Sponsors
• Full-Time VTuber in 2026 — Monetization Section

Eggie's threads:

• Eggie on growth + niching
• Eggie's follow-up thread