--- name: viral-shorts-course description: Decision framework for creating YouTube Shorts, Reels, and TikTok videos that go viral consistently, distilled from creators with verified analytics track records. --- # viral-shorts-course Skill v2 ## Purpose Decision framework for creating YouTube Shorts (and Reels / TikTok) that go viral consistently. Cross-referenced from 5 creators with verified analytics track records. Only patterns appearing in 3+ sources are promoted to "consensus principles." Conflicting advice is flagged explicitly. --- ## Sources | # | Creator | Credential | Content Analyzed | Views on Content | Duration | |---|---------|-----------|-----------------|-----------------|----------| | 1 | **Jenny Hoyos** | Avg 10M views/short; 600M+ views/yr; personal channel | Interview with Jay Clouse (Creator Science) | 4.6M views | 38:14 | | 2 | **Logan Smith** | 2.3M views on methodology video; 0→10K subs in 3 weeks (own data) | "How to make shorts that go viral every time" + "How to get retention SO HIGH" | 2.4M + 272K | 9:41 + 11:16 | | 3 | **heyDominik** (Dominik Rieger) | Agency founder; 28M views in 45 days for a client from 0 followers; 500+ creators worked with | "I Studied 1,000 Hooks" + "Blew Up Instagram to 30M Views in 40 Days" | 540K + 34K | 20:06 + 20:56 | | 4 | **MrBeast** (Jimmy Donaldson) | 200M+ subscribers; most-subscribed individual on YouTube; production memo internally verified by Business Insider | Leaked 36-page internal production document (Sep 2024) | N/A (internal doc) | ~36 pages | | 5 | **Colin and Samir** | 1.5M subs; 445M views; interview the top creators; run YouTube-native education | "How To Win on YouTube in 2026" (interview with Jay Acunzo) | 29K | 56:14 | **Analytics filter rationale:** - Jenny Hoyos: personal shorts average 10M views, verified by channel analytics - Logan Smith: own experiment data (0→10K in 3 weeks vs 4 months on first channel) - heyDominik: documented client case study, 28M views in 45 days from zero, public breakdown - MrBeast: largest individual YouTube channel in existence; memo describes what actually runs at scale - Colin and Samir: meta-analysts who interview top creators and publish data-backed frameworks --- ## Quick Reference One-page cheat sheet. Every short needs all of these. **Hook formula (first 3-5 seconds, visual + spoken):** ``` [Compelling visual in first frame] + [Open a loop — tell what happens, not how] + [Foreshadow the payoff at the end] ``` Jenny example: Visual = McDonald's location. "McDonald's banned this item" + "so I'm going to make it at home and convince them to put it back on the menu." Logan example: "I want to be the first person to have a snowball fight in the desert" — introduces concept, leaves 3 open questions. **5th grade readability:** Run script through readability-formulas.com. Never use "profit," "business," "finance," "strategy." MrBeast scripts: 1st grade. Your target: 5th grade or below. **Target duration: your data, not a universal rule.** Jenny's is 34s. Logan's data: once per day posting, time-to-peak varies by channel. Default to 34s if no data yet. **Retention benchmark: 90%+ required for virality** (Jenny). MrBeast: losing 35% of viewers in minute one is acceptable. The floor is different for long-form vs short-form. **Mechanism / payoff:** Every short needs a visible reason to watch to the end. Three-step list, countdown, or escalating stakes all work. Always delay the payoff — never pay off in the first 5 seconds (Logan: "the terrible mistake is paying off the topic too quickly"). **Structure in order:** 1. Hook (2-3 seconds, visual + spoken — open loop) 2. Foreshadow the ending (1 line — what they will see by the end) 3. Smooth transition line (not "let's get started") 4. But/Therefore storytelling through the middle (tension builds) 5. Tension peak before payoff 6. Payoff with twist (fulfill expectation AND surprise) 7. Abrupt end right after payoff — no outro **Platform split (not interchangeable):** - YouTube Shorts: 34s, story-driven, rewards rewatches, slightly mature audience - TikTok: 10-20s, dense information, no slow moments - Instagram Reels: visual-first, subtitles required, sharability is the metric; feels like the "cultural operating system of the United States" (Colin and Samir) --- ## Consensus Principles (3+ creators agree) ### C1. The hook is the most important single element **Jenny Hoyos:** "The hook is primarily visual, not verbal." First frame must communicate concept without sound. **Logan Smith:** "The first 5 seconds of your shorts [are] the most important part of your content." **heyDominik:** "Hooks, by far... are the most important part. If you got that down, you can kind of slack off on the rest of the video." **MrBeast:** "This is why we freak out so much about the first minute and go so above and beyond to make it the best we freakin can." **Colin and Samir:** "I think it's like a hover and hook platform... your hook becomes unbelievably important." Consensus: Hook is not just important — it is the single constraint that determines whether the video gets tested. Invest disproportionate time here. ### C2. Never pay off the topic too early — delay the payoff to force completion **Jenny Hoyos:** Mechanism rule — every short needs a visible progress indicator that forces viewers to watch to the end. Foreshadow the ending in line 2 of the script. **Logan Smith:** "The terrible mistake this Creator made was paying off the topic too quickly. Ideally the secret should have been revealed at the end — that way the audience had to keep watching in order to find it out." **heyDominik:** "Open loops and don't close them as long as possible, basically keeping them tangling at all times. And when you do close the loop, you want to open another one." **MrBeast:** The "last-to-leave" format exists precisely because the winner is never revealed early. "Crazy progression" in minutes 1-3 advances the story without resolving it. Consensus: Withholding the payoff is not a trick — it is the structural mechanism of all successful short-form content. The hook opens the loop. The payoff closes it. Everything between them must build tension, not release it. ### C3. Tension and conflict are the engine of storytelling **Jenny Hoyos:** "But/Therefore storytelling, not And/Then." Replace every "and then" with "but" or "therefore" to create tension. **Logan Smith:** "There is something that's important to focus on in storytelling for every niche... tension is the most important instrument of storytelling. As long as you're building suspense in your viewers." **heyDominik:** Build "binge sessions" through cliffhangers — "establishing hooks at video beginning and end to encourage consecutive content consumption." **MrBeast:** Stair-stepping format (escalating stakes) and last-to-leave format are both tension-delivery systems at scale. Consensus: And/Then is passive. But/Therefore is tension. If the script doesn't have a "but" or "however," it's probably boring. ### C4. End abruptly, immediately after the payoff **Jenny Hoyos:** "Abrupt end right after the payoff. No outro, no 'see you next time.'" **Logan Smith:** (implied via payoff structure — the climax is the last element, nothing after) **MrBeast:** "Video endings must always be abrupt to protect retention." Signal conclusion only for prize/payoff buildup. Consensus from 3+ creators: Any content after the payoff is dead weight that destroys the retention graph's final seconds. Cut immediately after the money moment. ### C5. Quantity produces the data you need to improve **Jenny Hoyos:** "Uploaded every day for one to two weeks specifically to get data. Retention graphs are useless without content volume to compare against." **Logan Smith:** "Between two channels I posted as little as twice per week and as often as four times per day... once per day was a sweet spot." **heyDominik:** "Posted twice a day for around one to two weeks" to train the algorithm. Pillar strategy requires volume to generate comparative data. **MrBeast:** (Production scale implies volume — the whole memo is built on iterating from data across hundreds of videos) Consensus: You cannot learn what works on your channel by studying other creators' analytics alone. Volume is how you generate your own signal. ### C6. The first frame must communicate the concept without sound **Jenny Hoyos:** "Sketch your first frame on an iPad before writing a script. If this were a long-form thumbnail and title, would it get clicks?" **Logan Smith:** Hook example ("snowball fight in the desert") — concept is understood visually from the first shot. **heyDominik:** "There's basically no way of making this look good" — analyzing creator whose first frame had visual clutter that prevented the hook from landing. **MrBeast:** "I Spent 50 Hours In Ketchup" — the thumbnail/title visual promise must be visible in the first frame, and production teams must know the exact thumbnail before filming. Consensus: If you mute your short and the first frame doesn't communicate the concept, the hook is broken. Fix the visual before you fix the words. ### C7. Audience specificity and simplicity — know exactly who you're talking to **Jenny Hoyos:** "Avatar is everything." Two nieces, ages 7 and 10, barely speak English. If they'd watch it, it passes. **Logan Smith:** "Think of your intended audience. What do they want to see? What type of videos are already performing well in your space?" **heyDominik:** "You have to be 100%, 95% aware of who your target audience is." Psychographics = "why do people want to watch your content? Why are they crying at night in bed?" **MrBeast:** "Simple enough for 50 million viewers to understand simultaneously." Content must work for the broadest possible audience. Consensus: Audience clarity precedes everything. Research phase (2-3 weeks per heyDominik) is not optional. Content that fails most often fails here before execution. ### C8. Format consistency builds the brand that makes people remember you **Jenny Hoyos:** Consistent structure (hook + foreshadow + but/therefore + twist + abrupt end) applied across all videos. **heyDominik:** Content pillars with similar hook structures train the algorithm and the audience simultaneously. "Binge sessions" only work if the format is recognizable. **Colin and Samir:** "The chance that you would remember a video, especially on a short form feed, it gets a lot higher if you've seen something similar to it from the same creator before... Finding a format is the path to consistency." **MrBeast:** Last-to-leave, stair-stepping, and "who else on YouTube can do this" are recurring formats applied systematically. Consensus: Virality without format is a one-time event. Format is what converts viral moments into a channel. Ask: "can I do this format for 52 weeks in a row?" --- ## Advanced Principles (single source, data-backed) ### A1. Sharability may outrank retention as the real virality signal (Jenny Hoyos) "My hunch — stated explicitly as not yet fully proven: sharability may be more important than retention." One short had 20% share-to-view ratio and 92% scroll-through rate. Retention graphs at 100% don't always predict blow-up. Design payoffs with: "would someone screenshot this and send it to a friend?" ### A2. Readability score is a hard filter, not a style preference (Jenny Hoyos) Scraped thousands of short transcripts, ran through readability-formulas.com. Most-viewed consistently score 5th grade or below. MrBeast: 1st grade. Single words like "business," "profit," "finance," "strategy" push score to 8th grade instantly. Replace with plain-language definitions every time. ### A3. The 3% per second cost rule (Jenny Hoyos) On a 30-second short, one dead second = 3.33% of video length. Trimming one second off the end of a video moved watch time from 83% to 88%, causing the video to take off. Monitor last-second retention religiously. Any dip at the end is a trim target. ### A4. The "wow factor" is a legitimate unlockable metric (MrBeast) "Anything that no other YouTuber can do" creates differentiation that data cannot capture. Lifting a house via crane adds nothing to the plot but makes the video unforgettable. "Who the fuck else on YouTube can do that lol" is the benchmark question for differentiating moments. Not measurable, but real. ### A5. Long-form structure: minute-by-minute re-engagements (MrBeast) For content over 3 minutes: - 0-1 min: Match thumbnail promise, establish urgency. Losing 35% of viewers here is acceptable. - 1-3 min: "Crazy progression" — multiple story beats delivered quickly - 3-min mark: First re-engagement (spectacular) - 3-6 min: Most exciting, simple, fastest content - 6-min mark: Second re-engagement (higher-production) - Back half: Long explanations acceptable This is long-form only. Shorts compress this to a single arc. ### A6. Negative comments as community accelerant (heyDominik) Share negative comments in Stories and create a conversation among followers. "This is one of the most powerful weapons" for triggering community engagement and propelling algorithm. Use sparingly and authentically. ### A7. Content pillars for algorithm training (heyDominik) Set up 4-5 pillars with similar hook structures before posting. The algorithm needs multiple data points to understand who to serve the content to. A 2-3 week research phase before posting is standard. First-party data from pillar comparison replaces guessing. ### A8. Music changes mood and builds tension from thin air (Logan Smith) "The soundtrack you pick for your shorts can be used to make the audience feel different moods during different parts of the short." Using a song with a buildup creates tension independently of the script. Wrong soundtrack can "backfire heavily." ### A9. Hover and hook — first-second retention is the new click-through (Colin and Samir) "As you're on the homepage, you can hover over a video and it auto plays for 40 seconds." YouTube TV auto-previews shift the bottleneck from thumbnail clicks to first-second hooks. "Thumbnails maybe become a little bit less important. Your hook becomes unbelievably important." This is an emerging 2025+ dynamic. --- ## Conflict Map Where creators disagree — with analytics weight assessment for each position. ### CONFLICT 1: Ideal short duration — is there a universal number? **Jenny Hoyos:** 34 seconds is her data-backed optimum (personal analytics). "There is no universal ideal duration." She analyzed her own top 10 performing videos. **Logan Smith:** Does not specify a target duration; focuses on retention graph shape, not seconds. **MrBeast:** Average video length is 13:37 (long-form) — not applicable to shorts. **heyDominik:** Initial pillars used 5-15 second snappy videos as core content type for Instagram/Reels growth. **Assessment:** Jenny's 34s is a personal data point, not a universal truth — she says so explicitly. heyDominik's 5-15s reflects Reels/TikTok platform behavior. The real principle is: analyze your own retention data. 34s is the default if you have no data. Don't apply it mechanically. ### CONFLICT 2: Posting frequency — how much is too much? **Logan Smith:** "Once per day was a sweet spot... four times per day was too much — quality dropped and only half got picked up." **heyDominik:** "Posted twice a day for around one to two weeks initially" to accelerate algorithm training. Then reduced. **Jenny Hoyos:** "Uploaded every day for one to two weeks specifically to get data." Implies short high-frequency burst, not indefinite daily posting. **MrBeast:** Volume is a function of production capacity. Not applicable to solo/small creators. **Assessment:** The conflict is timing, not principle. All agree on a high-frequency burst (1-2x/day) for the first 1-2 weeks to generate data and train the algorithm. After that, Logan's "once per day" is the sustainable floor. The mistake is applying burst-phase logic to the steady-state. ### CONFLICT 3: Is shorts growth transferable to channel growth? **heyDominik:** Shorts/Reels with pillar strategy can generate subscriber conversion and community if hooks create binge sessions. Case study: 100K followers in 45 days. **Colin and Samir:** "I don't believe yet that YouTube shorts specifically dictates or drives culture... I'm not sure that short people in a short form feed are even really thinking about going long form." Shorts as "extra data" that informs long-form, but not a subscriber conversion engine on YouTube specifically. **Assessment:** Both are right in their context. heyDominik is operating on Instagram/Reels which has messaging-driven sharing. Colin and Samir are describing YouTube Shorts specifically, where the lack of native messaging limits virality spread. Platform matters. On Instagram, shorts compound into brand. On YouTube Shorts, they complement long-form but may not drive it. ### CONFLICT 4: What matters more — retention or sharability? **Jenny Hoyos:** "Sharability may be more important than retention" (explicitly flagged as a hunch, not proven). 20% share-to-view ratio on one short. **Logan Smith:** "Just this simple line [the retention graph] will decide if your short gets 10 views or 10 million views." **MrBeast:** Three metrics: CTR, AVD, AVP. Retention-centric. **Assessment:** Jenny has the most granular personal data and has earned the right to challenge conventional wisdom. But she herself flags this as unproven. Logan and MrBeast represent mainstream evidence. Safe operating position: optimize for retention first (it's the known lever), design payoffs that are also shareable (no cost to add this). Do not trade retention for sharability. --- ## Agentic Tools Scan Tools that fit into the viral-shorts workflow. Evaluated on: does it solve a real workflow bottleneck, is it used by creators with proven results, and does it earn its complexity? | Tool | What it does | Workflow slot | Pricing | Verdict | |------|-------------|--------------|---------|---------| | **OpusClip** | AI auto-clipping: extracts viral moments from long video, adds animated captions, "Virality Score" ranking | Long-to-short repurposing; finding best clip in 1-hour video | Free tier; paid from ~$15/mo | WORTH IT — saves 8-10 hrs/video, clips ranked by virality score, 80% ready-to-post. Best ROI tool in this list. | | **Submagic** | Caption generation + animated captions for short-form; word-level highlight + emoji support | Caption styling post-production | ~$24/mo | WORTH IT for caption work specifically. Accurate, fast, word-highlighting matches viral caption style. Use if OpusClip captions don't match your brand. | | **Captions.ai** | All-in-one: script → shoot assist → captions + auto-edit for mobile | Mobile-first production; good for solo creators shooting on phone | Free tier; paid ~$10/mo | SITUATIONAL — best for mobile-first workflows. If you shoot on camera + edit on desktop, OpusClip + DaVinci/Premiere covers this better. | | **Vizard** | Long-to-short repurposing + caption + social scheduling; analyzes emotional peaks | Alternative to OpusClip; scheduling included | ~$14.50/mo | ALTERNATIVE to OpusClip. Slightly simpler, includes scheduler. Use if you need scheduling + clipping in one tool. | | **Klap** | Long-to-short AI clipping; AI speaker reframing; multi-language dubbing | International content; podcaster repurposing | ~$29/mo | SITUATIONAL — strong if you need dubbing for non-English markets. Reframing is best in class. Overkill for single-language creators. | | **Munch** | Clip extraction + trending topic analysis + multi-platform distribution suggestions | Strategy layer — tells you which clips suit which platform | ~$49/mo | NOISE at early stage. Adds value at scale (agency, team). Solo creator: skip until you're posting 10+/week. | | **Vidyo / Quso.ai** | AI clipping + captions + 8-platform scheduling | All-in-one for social media managers | Free tier; paid ~$25/mo | ADEQUATE — not best-in-class at any single thing, but covers the whole workflow cheaply. | | **Vizard** | See above | | | | | **Decipher** | Transcription + semantic search through your own video archive | Finding past clips by topic keyword | Paid per minute | NICHE — only useful once you have 50+ hours of archived content. Skip until then. | | **Eddie AI** | Pro rough-cut generation for interview/doc content; exports to Premiere/FCP/DaVinci | Professional production pipeline for interview-heavy content | Paid (premium) | PRO ONLY — built for documentary producers with Premiere/FCP. Not for typical social creator workflow. | **Top 3 tools that genuinely save time vs noise:** 1. **OpusClip** — if you record long-form content and want shorts from it, nothing else competes on clip discovery + caption + virality ranking. The "8-10 hours saved per video" number is realistic. 2. **Submagic** — for creators who write their own shorts from scratch and need word-level animated captions that match viral style. Single-purpose, does it well. 3. **Captions.ai** — for mobile-first solo creators who shoot on phone. Script-to-caption pipeline is fastest here. **What adds noise:** Munch, Decipher, and Eddie AI are premature at creator-stage. They are agency/team tools that add process before you have the content volume to need process optimization. --- ## How to Apply ### Pre-production checklist - [ ] Audience: write down exactly who this is for. Age, context, what they're anxious about right now. - [ ] Sketch the first frame. Concept communicated muted? If not, re-sketch. - [ ] Write hook line (5 words or fewer if possible). Run through readability-formulas.com. 5th grade or below. - [ ] Write foreshadow line (line 2 of script — what will happen at the end). - [ ] Identify the tension mechanism. What is the "but" that prevents easy resolution? - [ ] Identify the payoff twist. Fulfills expectation AND surprises. - [ ] Write the last line BEFORE you write the middle. Know where you're ending. - [ ] Identify mechanism/visible progress. Three-step list, countdown, or escalating stakes. ### Production checklist - [ ] First frame intentionally composed. Clean visual = concept understood muted. - [ ] Script 5th grade or below. No "profit," "business," "finance," "strategy." - [ ] Hook + Foreshadow in under 4 seconds combined. - [ ] Smooth transition line (not "let's get started"). First action of the story. - [ ] Middle section: every sentence is "but" or "therefore," never "and then." - [ ] Music selected to match mood arc; buildup aligns with tension peak. ### Post-production checklist - [ ] Duration: check your own data. Default to 34s if no data yet. - [ ] Check retention graph after upload. Any dip in last 2 seconds: trim it. - [ ] Scroll-through rate / swipe-away rate: target 85%+ (Jenny benchmark). - [ ] First-frame visual hook test: muted = concept understood. - [ ] Payoff has a twist. Would someone share this with a friend? - [ ] Captions: word-level highlight on key words only. Never full-sentence highlight. - [ ] Abrupt end — nothing after the payoff. --- ## Anti-patterns | Anti-pattern | Why it fails | Who calls it out | |---|---|---| | Paying off the topic in the first 5 seconds | Removes all reason to keep watching | Logan Smith (explicit), heyDominik (loop theory) | | "Let's get started" / "Without further ado" | Breaks pacing — replace with first action line | Jenny Hoyos | | And/Then storytelling in the middle | No tension = no reason to stay | Jenny Hoyos, Logan Smith | | No foreshadow of ending | Viewers leave without a commitment to the payoff | Jenny Hoyos, Logan Smith | | Same video posted to all platforms unchanged | YouTube Shorts and TikTok are different platforms, different audiences | Jenny Hoyos, Colin and Samir | | Outro / "see you next time" | Destroys last-second retention | Jenny Hoyos, MrBeast | | Adult vocabulary in script | "Business," "strategy," "profit" push readability to 8th grade | Jenny Hoyos | | Hook that only works with audio | Viewers scroll on mute; visual must carry the concept | Jenny Hoyos, heyDominik | | Posting 4x/day indefinitely | Quality drops, algorithm picks up only half | Logan Smith | | Optimizing for retention alone without asking about shareability | Retention at 100% does not guarantee distribution | Jenny Hoyos | | No content pillar / format strategy | One viral video doesn't compound into channel growth | heyDominik, Colin and Samir | | Skipping the research phase | Pillars built on guesswork fail; demographics + psychographics first | heyDominik | | Spending money on production that doesn't appear on screen | "If you're spending over $10,000 on something and it won't be shown on camera, seriously think about it" | MrBeast | | Waiting for the perfect idea | Jenny has 1,000 ideas, publishes ~10. Volume creates the data. | Jenny Hoyos | --- ## Examples and Templates ### The Irony Structure (Jenny's garden video — 30M views) ``` [Expensive/annoying thing I deal with] + [Ironic solution that shouldn't work] + [Does it work?] Hook: "[Problem] costs me [amount/time] every time" Foreshadow: "So I'm going to [ironic fix] to [goal]" ``` ### The Expectation + Twist Structure (Jenny's Mother's Day video) ``` Hook: "[Person] has never had [thing]" Foreshadow: "So I'm going to [give them the thing]" Twist: [Delivery goes wrong but reaction is better than expected] ``` ### The HPC Formula (Logan Smith) ``` H - Hook: introduce topic, leave 3 open questions unanswered P - Progression: fulfill the promise made in the hook; show the journey C - Climax: reveal payoff at the end, not the beginning ``` Example: "I want to be the first person to have a snowball fight in the desert." Opens: how, where, is this possible? Progression: getting to the mountains, collecting snow. Climax: the snowball fight happens. ### The Loop Stack (heyDominik) ``` Open loop at the start → never close it fully → when you do close it, open another one → second hook at end creates cliffhanger into next video ``` Used to generate binge sessions across multiple posts. Works in longer content (60s+) where a cliffhanger ending drives traffic to the next piece. ### The Stair-Stepping Format (MrBeast) ``` Start: [$1 version of the thing] Middle: [$100 version → $1,000 version] End: [$100,000 version — the payoff] ``` Stakes increase at every step. Viewer is always oriented toward the next escalation. ### The Pacing Break (Jenny's "I cooked illegally" line) ``` After Hook + Foreshadow, never say "let's get started." Instead: [First action verb from the story] + [one surprising detail about that action] ``` "McDonald's banned this item, so I'm going to make it at home and convince them to put it back on the menu. So I cooked illegally." — flows from foreshadow, adds humor, advances story, zero dead air. --- ## Cross-skill references - `video-editing-pro`: cut decisions, hook timing (1.5-3s), pause rules, retention benchmarks, caption highlighting — execution layer for what this skill defines conceptually - `thumbnails`: first-frame composition principles for shorts = thumbnail principles — what communicates muted communicates at a glance - `social-media`: per-platform duration norms, caption format requirements, posting strategy, Instagram vs YouTube behavior differences - `seo-pack`: hook = curiosity-gap headline; the "but" tension structure = the same mechanism as an SEO hook that withholds the answer - `appie-video-production`: production workflow, normalization, shot setup for Weblyfe context - `content-factory` (when exists): volume strategy, pillar system, frequency recommendations from heyDominik's framework