StoryCraft · Omni-Video Framework
Omni-Video Framework
The anatomy of every video that ships through the Omni system — long-form, medium-form, short-form, entertainment, documentary, and testimonial. Each format has a fixed skeleton: hook, story, proof, CTA. This doc gives you the shot list, gear settings, color recipe, pacing rules, and distribution plan for each one.
The Universal Omni-Video Skeleton
Every video, regardless of format, is built from the same five beats. The ratio between them is what changes by format.
| Beat | Long-form | Medium-form | Short-form | Testimonial | Documentary | Entertainment |
|---|---|---|---|---|---|---|
| Hook | 0–15s | 0–10s | 0–3s | 0–5s (result stated) | 0–20s (question) | 0–2s (pattern break) |
| Setup / Context | 15–60s | 10–30s | skip | The problem | The world | Character intro |
| Body / Story | 60% of runtime | 45–75% of runtime | 3–20s | The turning point | Interviews + B-roll | Tension escalation |
| Proof / Payoff | Case, data, demo | Result, demo, guarantee | Reveal | Named result + metric | Resolution / insight | Punchline |
| CTA | End card + mid-roll | Single clear next step | Caption + comment prompt | Soft — brand mention | Subscribe / next episode | Follow for more |
Long-Form Video
YouTube evergreens, webinars, explainers, deep dives. Purpose: authority + search discovery + long dwell time. Runtime 7+ minutes.
Shot list & structure
Optimize for retention curve. Every 30–45s needs a visual or pacing change or the audience drops.
Shot List
- Cold open (0–15s) — biggest promise or a curiosity gap. NO logo bumper.
- Talking head A-cam — 50mm equivalent, eye-level, chest up, background 6ft+ behind.
- B-cam / wide — cut every 20–30s for movement.
- B-roll — one clip per key point. Product, process, or metaphor.
- Graphics / lower thirds — for stats, names, chapter titles.
- Screen recordings — for tutorials; zoom to focal element.
- End card (20s) — next video + subscribe.
Gear & Settings
- Resolution / fps — 4K 24fps (cinematic) or 4K 30fps (talky).
- Aspect — 16:9 (1920×1080 min).
- Shutter — 2× fps (1/50 at 24fps).
- Aperture — f/2.8–f/4 for A-cam, f/5.6 for B-roll.
- ISO — native (100–800 depending on body).
- Audio — lav OR shotgun 6" off frame. Record at –12dB peak.
- Light — key at 45°, fill at 1/2 power, hair/rim for separation.
Color / LUT Recipe
- Base LUT — neutral Rec.709 conversion of your log profile.
- Contrast — +8 to +12; crush blacks slightly.
- Saturation — +5; skin tones protected via HSL.
- Temperature — warm bias (+150K) for interviews.
- Grain — light 35mm grain overlay for cinematic feel.
Pacing / Edit Rules
- Cut on breath or thought change — never on dead air.
- Jump cut to compress ums, but preserve human rhythm.
- Chapter markers every 60–120s (YouTube retention lift).
- Music bed low (–24dB), duck under dialogue.
- Pattern interrupt every 30s: cut, zoom, graphic, or B-roll.
Thumbnail Rules
- One human face, one big emotion, one high-contrast object.
- Max 3 words of text, 60pt+ bold.
- Legible at 120×68 px (mobile feed size).
- Not the exact promise of the title — creates curiosity gap.
Distribution
- Primary — YouTube (SEO-driven title + description).
- Secondary — website resource page, sales enablement, email nurture.
- Cutdowns — 3–5 short-form clips per long-form asset.
Common Mistakes
Logo bumper up front
Kills retention in the first 5 seconds. Save branding for the end card.
Talking head for 8 minutes straight
No B-roll = no reason to keep watching. Break every 30s.
Weak thumbnail + strong title
Thumbnail does 80% of the click. Invest there.
CTA only at the very end
Most viewers leave before it. Plant a mid-roll ask too.
Medium-Form Video
Landing-page explainers, VSLs, product demos, and sales stories. Purpose: education + conversion. Runtime 2–7 minutes — long enough to make the case, short enough to hold attention to the CTA.
Hook · Problem · Promise · Proof · CTA
One idea, one offer, one next step. Every section earns the right to the next.
Shot List
- Cold open (0–10s) — lead with the biggest promise or the most painful problem.
- Talking head or voiceover — direct address to camera; keep eyes near lens.
- Problem agitation — 1–2 relatable scenarios the viewer recognizes.
- Solution / promise — the clear outcome your product or service delivers.
- Proof block — demo, testimonial clip, metric, or guarantee.
- Offer stack — what they get, what it's worth, why now.
- CTA card (10–15s) — single button, phone number, or booking link on screen.
Gear & Settings
- Aspect — 16:9 primary; 1:1 or 4:5 for social placements.
- Resolution / fps — 1080p or 4K at 24fps or 30fps.
- Audio — lav mic preferred; voiceover track mixed clean at −6dB.
- Captions — burned in for social; optional on landing pages.
- Background — clean, on-brand, non-distracting; depth behind subject.
Color / LUT Recipe
- Base — clean Rec.709 with subtle contrast.
- Contrast — +8 to +10; keep shadows clean.
- Saturation — +5; skin tones natural.
- Temperature — neutral to warm (+100K).
- Text overlays — high-contrast, large enough to read on mobile.
Pacing / Edit Rules
- Cut every 5–10s to maintain energy without feeling frantic.
- Use B-roll to illustrate every claim; never let talking head run alone.
- Lower thirds for names, stats, and key takeaways.
- Music bed low (−22dB), duck under voiceover.
- Show the CTA on screen for the full final 10s.
Placement
- Primary — landing pages, sales pages, VSL funnels.
- Secondary — email nurture, proposal follow-ups, paid social.
- Cut-downs — 15–30s proof clips for ads; 60s summary for social.
Conversion Rules
- One offer, one CTA. Multiple asks dilute action.
- State the promise in the first 10s and repeat it before the CTA.
- Address the top 1–2 objections in the proof block.
- End with urgency or risk reversal — guarantee, limited slots, bonus.
Deep Dive — The VSL (Video Sales Letter)
One viewer, one problem, one offer, one action
A VSL is a sales letter performed on camera. It replaces a long sales page with a single narrated argument that moves a cold or warm viewer to book, buy, or call. Standard runtime 4–7 minutes; short VSLs for paid traffic run 2–4 minutes.
VSL Script Structure
- Pattern interrupt (0–10s) — the promise, the stat, or the pain, said plainly.
- Callout — name exactly who this is for ("If you run a home-service company doing $1M+…").
- Problem & cost of inaction — what it costs them every month they wait.
- Failed solutions — why what they already tried didn't work; removes blame.
- Mechanism — the unique "how" behind your result. This is the heart of a VSL.
- Proof — case numbers, screenshots, testimonial cutaways.
- Offer & stack — deliverables, value, price framing.
- Risk reversal — guarantee, pilot, or no-contract term.
- Close + CTA — one action, stated three times, with the button on screen.
Production Spec
- Delivery — teleprompter at lens height; conversational, not read.
- Visuals — alternate talking head with screen recordings, charts, and B-roll every 6–10s.
- Text — on-screen kinetic text for every key claim and number.
- Audio — lav or shotgun, voice at −6dB, music bed −24dB, no music under the close.
- Player — large play button, captions on by default, no autoplay with sound.
Funnel Placement
- Cold paid traffic — 2–4 min VSL above the fold on a dedicated landing page.
- Warm / retargeting — 4–7 min full VSL with the booking calendar below the player.
- Pre-call asset — send in the confirmation email so the call starts at the offer, not the pitch.
- Cut-downs — hook + mechanism (30s) and proof block (15s) become the ad creative.
VSL Metrics
- Play rate — % of page visitors who press play. Target 35%+.
- Hook retention (30s) — target 60%+ still watching.
- Offer-point retention — % who reach the offer section. Target 25%+.
- VSL → booking rate — target 5–12% of players on warm traffic.
- Fix in this order: play rate (thumbnail/headline) → 30s hold (hook) → offer reach (pacing) → conversion (offer & CTA).
Common Mistakes
Multiple CTAs
"Call, email, visit, download" kills conversion. Pick one next step.
Slow problem setup
If the viewer doesn't feel the problem in 30s, they bounce.
No proof
Claims without demo, testimonial, or metric feel like hype.
Running past 7 minutes
Unless it's a deep demo, medium-form that drifts becomes long-form that loses.
VSL with no mechanism
Without a unique "how," the VSL is just claims. The mechanism is what makes the offer believable.
Hiding the VSL CTA
The booking link must be visible under the player the whole time, not revealed at the end.
Short-Form / Reels / TikTok / Shorts
Vertical 9:16, 0–2 minutes. Purpose: reach, follower growth, top-of-funnel awareness. The algorithm rewards watch-through and rewatches.
Hook · Loop · Payoff
The first 3 seconds decide everything. The last frame decides if they rewatch.
Shot List
- Hook frame (0–1s) — visual pattern break OR bold on-screen text.
- Verbal hook (1–3s) — question, contrarian claim, or curiosity gap.
- Body (3–20s) — 3–5 tight cuts, each <3s.
- Payoff / reveal — the answer, transformation, or punchline.
- Loop close — final frame visually matches opening frame so it rewatches seamlessly.
Gear & Settings
- Aspect — 9:16 (1080×1920). Shoot vertical or crop with room.
- Resolution / fps — 1080p 30fps or 60fps (for slow-mo cuts).
- Safe zones — keep text and faces inside center 80% (top/bottom get UI overlays).
- Audio — phone lav or Rode Wireless Go. Music from platform library for algorithmic boost.
- Captions — burned in. 85% watch muted.
Color / Filter Recipe
- Base — bright, punchy, phone-native look.
- Contrast — +15 to +20 (feed-stopping).
- Saturation — +10 to +15.
- Sharpen — moderate; avoid halos.
- Filter preset — CapCut "Fresh," Instagram "Clarendon-lite," or a custom LUT tuned for skin.
Pacing Rules
- Cut every 1.5–3 seconds max.
- Add subtle zoom/pan on static shots to fake motion.
- Front-load payoff clues to keep watch time.
- End on a loop or a question in caption ("agree?").
Distribution
- Primary — TikTok + Instagram Reels + YouTube Shorts (post native to each).
- Secondary — Facebook Reels, LinkedIn (business-context edits).
- Cadence — 3–5x per week per platform to feed the algorithm.
Sound-On / Sound-Off Design
- Visual + captions must carry the story alone.
- Music/sfx should enhance, not deliver, information.
- Use trending sounds only when they fit — not for their own sake.
Common Mistakes
Horizontal video reposted vertical
Wastes 50% of the frame. Algorithm demotes it. Shoot native vertical.
No captions
Muted viewers scroll past. Burn them in.
Hook is a slow intro
"Hey guys, today we're going to talk about…" = dead. Skip to the promise.
Text in top/bottom 10%
Covered by platform UI. Keep it center-safe.
Entertainment-Style Video
Character-driven, comedic, or emotionally engaging content. Purpose: brand affinity and shareability. Sells nothing directly — earns attention that later assets convert.
Character · Tension · Payoff
Story rules the frame. Production polish is secondary to timing and delivery.
Structure
- Character intro — establish who and what they want in ≤3s.
- Tension / conflict — obstacle, misunderstanding, or setup.
- Escalation — 2–3 beats that raise stakes or absurdity.
- Payoff / punchline — visual reveal, twist, or callback.
- Button — one final reaction shot or tag.
Gear & Settings
- Same as short-form; skip cinema polish.
- Handheld / natural = feels less scripted.
- Two-shot + reaction cuts drive comedy timing.
- Music low, sfx accented (whoosh, ding, record scratch).
Trend-Jacking Rules
- Only jack a trend if it fits your character/voice.
- Use the trending sound; write your own visual/text.
- Post within 5–7 days of the trend peaking.
- Don't force a product mention — let it live in the caption.
Delivery Rules
- Deadpan > overacting.
- Cut on the beat of the joke, not after.
- Rewatch loop = watch-time compounding.
- Recurring character or bit builds an audience.
Common Mistakes
Sales pitch dressed as comedy
Audience smells it and swipes. Entertainment first, brand second.
Trend without a twist
Copying = invisible. Add your angle.
Joke too late
Payoff after 20s in a Reel = missed. Compress.
Documentary-Style Video
Interview + B-roll storytelling. Purpose: trust, depth, brand mythology. Works for founder stories, customer journeys, "behind the scenes."
Interview cinematography & story arc
Real moments beat scripted ones. Your job is to capture them and cut ruthlessly.
Interview Setup
- Framing — subject looks slightly off-camera; interviewer sits next to lens.
- Lens — 50mm or 85mm equivalent, f/2.0–f/2.8 for shallow depth.
- Lighting — soft key from window or diffused LED at 45°; negative fill for depth.
- Audio — lav on subject + boom overhead as backup. Record room tone 60s.
- Two cameras — A-cam medium close-up, B-cam wide or tight cutaway.
B-Roll Priorities
- Hands working / process shots.
- Environment establishing shots.
- Product or artifact close-ups.
- Ambient candid moments (no eye contact with camera).
- Slow motion for emotional beats (60–120fps).
Color / LUT Recipe
- Look — filmic, muted, slightly desaturated.
- Contrast — +5, lift shadows to 5–8 (matte look).
- Saturation — –5 to –10 overall; boost oranges/teals via HSL.
- Grain — moderate 16mm or 35mm overlay.
Edit / Pacing
- Let silence breathe. Reaction beats matter.
- Use natural sound (nat sound) between edits — no sterile cuts.
- Lower thirds on first appearance only.
- Score sparingly; music should serve emotion, not fill space.
Common Mistakes
Leading questions
You get soundbites, not stories. Ask open, then wait.
Wall-to-wall music
Kills authenticity. Let nat sound carry.
Over-directed subject
Stiff = untrustworthy. Warm them up off-camera first.
No B-roll plan
Talking head only = boring. Shoot 4× more B-roll than you think.
Testimonial Video
Customer speaks; brand listens. Purpose: proof, objection-crushing, conversion lift. This is the highest-ROI creative asset most brands underinvest in.
Problem · Turning point · Result
Structure the interview to yield this arc, then cut to it.
Question Framework
- Who — name, role, business.
- Before — what problem/frustration existed?
- Trigger — what made them look for a solution now?
- Objection — what almost stopped them from buying?
- Turning point — first "aha" moment with the product/service.
- Result — specific outcome, ideally with a number.
- Recommendation — who else should use this and why?
Shot List
- Talking head (interview A-cam).
- B-roll of them working / at their business.
- Product / service in use.
- Result visualization (screenshot, metric, before/after).
- Optional: happy customer / team candid.
On-Screen Trust Markers
- Lower third: full name, title, company, location.
- Metric callout when result is stated (animated number).
- Logo of their business (bottom corner).
- Runtime: 60–90s for social; 2–3 min for landing pages.
Delivery Recipe
- Open on the strongest result soundbite — not the problem.
- Then flash back to the problem/objection.
- Middle: journey + turning point.
- End: recommendation + brand tag.
Common Mistakes
Scripted testimonial
Reads fake. Interview conversationally, cut for tightness.
Vague result ("really helped us")
Push for a number. "3x leads in 60 days" > "helped us."
No objection addressed
Missing the "I almost didn't…" beat costs conversions.
Poor audio quality
Bad sound = untrusted. Never skip the lav.
Color / LUT Cheat Sheet by Format
| Format | Look | Contrast | Saturation | Grain | Notes |
|---|---|---|---|---|---|
| Long-form YouTube | Cinematic-natural | +8 to +12 | +5 | Light 35mm | Warm bias interviews |
| Medium-form / VSL | Clean conversion-focused | +10 | +5 | None | Skin tones true, text legible |
| Short-form / Reels | Punchy phone-native | +15 to +20 | +10 to +15 | None | Sharpen +moderate |
| Entertainment | Match trend / raw | +10 | +8 | None | Timing > polish |
| Documentary | Filmic matte | +5, lifted blacks | −5 to −10 (HSL boost oranges/teals) | Moderate 16/35mm | Muted, emotional |
| Testimonial | Clean corporate-warm | +8 | 0 to +5 | None | Skin tones true |