STORYBOARD — Her Labs Were Normal. Her Hair Wasn't.

01Comp library inspiration anchors

[Anchor A] 1674502353517019 — nivara-shop — 59.3 — podcast-ferritin-hair (THE format-match cornerstone) Hook pattern: host Q + expert reveal (numbers-mystery); split-screen/alternating interview On-screen text style: native-subtitles throughout (podcast convention) — carries the mute test for 2p B-roll cuing: visual-proof-bottle-hold; split-screen-simulation; problem-agitation insert Transition: alternating-cut between host + guest; root-cause-reframe Mute-test asset: native subtitles + two-shot/split (no single face carries — subtitles do the work) Best-fit beat type in our script: ALL — this is the exact dynamic×setting×mechanism (hair+thyroid podcast) match. PRIMARY across every beat. Tooling stack inferred: Heygen multi-avatar / split-screen. Matches our target ✓.

[Anchor B] 1091695962938552 — thebbco — 90.2 — partner-intro 2p register Hook pattern: conversational 2p open Best-fit beat type: hook + handoff cadence — SECONDARY. Tooling stack inferred: Heygen 2-avatar. ✓

[Anchor C] 566786836079175 — elavate — 88.7 — Q&A buying-guide Best-fit beat type: the skeptic-Q + proof beats (Q&A structure) — SECONDARY. Tooling stack inferred: Heygen/real 2-person. ✓

No asset_library.md (V3) — product B-roll (bottle-hold) AI-gen or expert holds prop.

02Per-beat storyboard

▸ Beat 1: THE QUESTION 0:00–0:18

Persuasion goal Mute-test (STRICT 2p) + Q&A hook (host Q → expert "doesn't even measure" reveal); numbers-mystery loop; skeptic re-ask Shot type: Two-shot establish → split-screen (host left / expert right) on the Q&A volley Camera: Eye-level, locked; alternating-cut on each line per Anchor A On-screen text: "HER LABS WERE 'NORMAL.' HER HAIR WASN'T." — center, animate in t=1s, hold to t=5s + native subtitles begin t=0 (subtitles carry the 2p mute-test — no single face does) [Anchor A native-subtitles] B-roll cue: none in hook (two-shot + caption + subtitles carry) Lighting: Clean studio two-point per speaker, warm-neutral Transition out: Alternating-cut continues into Beat 2 Avatar-tool prompt: SCENE: clean podcast studio, two chairs, warm-neutral two-point lighting | AVATAR_1 (host): {host_id} — leaning in, curious, eyeline to expert | AVATAR_1_VOICE: {host_voice} curious-peer | AVATAR_1_SCRIPT: "Okay, so a woman's losing her hair. Handfuls of it. But her doctor says her thyroid's fine. What's actually going on there?" | AVATAR_2 (expert): {educator_id} — measured, eyeline to host | AVATAR_2_VOICE: {educator_warm} | AVATAR_2_SCRIPT: "Her follicles are running out of fuel. And the fuel is a form of thyroid hormone her standard test doesn't even measure." | SHOT: two-shot → split-screen | CAMERA: locked, alternating cut | ON_SCREEN_TEXT: "HER LABS WERE 'NORMAL.' HER HAIR WASN'T." center t1-5s + native subtitles from t0 | DURATION 18s Mute-test: PASS (STRICT 2p path) — center caption + native subtitles + two-shot carry sound-off; subtitles do the heavy lifting since no single speaker-face anchors a 2p frame (per Anchor A) Production-limitation cross-ref: ⚠️ Heygen 2-avatar same-frame eyeline can read awkward → use split-screen / sequential framing (Anchor A split-screen-simulation), NOT a single composited two-shot held long

▸ Beat 2: THE TWO HORMONES 0:18–0:55

Persuasion goal Mechanism (storage vs active) via reaction-translation handoff; host voices "fine on paper, falling apart for real" (VOC §7) Shot type: Split-screen / alternating singles; expert carries, host punctuates Camera: Locked alternating-cut; cut to host on his reaction line On-screen text: native subtitles throughout; "storage T4 ≠ active T3" lower-third t=30–36s (≤6 words) B-roll cue: simple T4→T3 split-screen-simulation graphic t=24–40s [Anchor A] Lighting: Default Transition out: Alternating-cut Avatar-tool prompt: Heygen alternating singles | AVATAR_2 (expert) carries mechanism lines | AVATAR_1 (host) reaction-translation cut: "So the report can say normal—" / "fine on paper, falling apart for real." | subtitles on | T4→T3 graphic overlay t24-40s Mute-test: N/A (subtitles maintain sound-off comprehension) Production-limitation cross-ref: sequential framing keeps eyelines clean

▸ Beat 3: WHY THE HAIR FIRST 0:55–1:28

Persuasion goal Follicle mechanism + the two minerals (selenium/zinc); story-prompt handoff Shot type: Alternating singles; expert on follicle explanation Camera: Locked alternating On-screen text: subtitles; "selenium + zinc = conversion" lower-third t=1:18–1:24 (≤5 words) B-roll cue: brief follicle-dormancy graphic t=1:00 [Anchor A] Lighting: Default Transition out: Alternating-cut to skeptic Q Avatar-tool prompt: Heygen alternating | AVATAR_1 story-prompt: "And why does it show up in the hair first?" | AVATAR_2 follicle answer + minerals | subtitles on | follicle graphic t1:00 Mute-test: N/A Production-limitation cross-ref: none beyond Beat 1 framing note

▸ Beat 4: THE SKEPTIC QUESTION 1:28–2:02

Persuasion goal Skeptic handoff pre-empts overclaim ("is this just take-a-supplement?"); expert: "I'd be careful with anyone who promises that" + won't-work-for-everyone + adjunct + autoimmune-iodine Shot type: Two-shot for the skeptic exchange (tension reads better in-frame here), then expert single Camera: Locked; cut to expert for the honest answer On-screen text: subtitles; none other (the honesty beat is sound-on/subtitle) B-roll cue: none (talking-head exchange is the point) Lighting: Default Transition out: Alternating-cut to close Avatar-tool prompt: Heygen | AVATAR_1 skeptic: "is this just 'take a supplement and your hair grows back'? Because women have heard that a thousand times." | AVATAR_2: "No — and I'd be careful with anyone who promises that…" | two-shot → expert single | subtitles on Mute-test: N/A Production-limitation cross-ref: none

▸ Beat 5: PROOF + CLOSE 2:02–2:30

Persuasion goal Past-story proof ("I had a patient last year… baby hairs came back") + EXPERT delivers CTA (per 2p-interview §4a default) + host tag ("Full panel. Feed the thyroid. Love it.") Shot type: Expert single for proof+CTA → cut to host for the tag; product bottle-hold insert Camera: Locked; expert to camera for CTA (credentialed close); host tag two-shot On-screen text: subtitles; "Kindled — link below · alongside your prescription" lower-third t=2:22–2:28 (≤8 words); end-card B-roll cue: visual-proof-bottle-hold — expert holds Kindled dropper naturally t=2:18 [Anchor A bottle-hold]; AI product end-card Lighting: Default warm; product clean Transition out: End card Avatar-tool prompt: Heygen | AVATAR_2 (expert) proof + CTA to camera, holds Kindled dropper (Anchor A bottle-hold) | AVATAR_1 (host) tag: "Full panel. Feed the thyroid. Love it." | subtitles on | end-card overlay post | bottle = real prop OR MARKETING_STUDIO product insert Mute-test: N/A (close) Production-limitation cross-ref: bottle-hold by avatar = mild Heygen hand/prop risk → product can be a separate Marketing Studio insert if avatar prop-hold reads off

03Visual continuity audit

Location: single podcast studio, two-chair, throughout. ✓ Lighting: warm-neutral two-point consistent. ✓ Speakers: host + expert consistent across beats; alternating-singles/split-screen framing consistent (Anchor A). ✓ Subtitles: present every beat (continuity + mute-test). ✓

04Mute-test verification

First 5s: PASS (STRICT 2p) — center caption "HER LABS WERE 'NORMAL.' HER HAIR WASN'T." + native subtitles from t=0 + two-shot. Per the checklist's hook-discipline: a 2p frame has no single anchoring face, so native subtitles + the bold caption carry the sound-off load (Anchor A's exact solution). Survives full 5s (VS5 ✓). Caption 6 words (VS2 ✓).
Native subtitles throughout — the 2p mute-test insurance (every beat readable sound-off).

05Production-limitation cross-ref

production_limitations.md empty → silent-skip. Setting-file/anchor note (not YS6): Heygen 2-avatar same-frame eyelines are imperfect → designed around via split-screen-simulation + sequential alternating framing (Nivara's exact pattern, Anchor A) + avatar prop-hold de-risked to optional Marketing-Studio insert. Production-feasible via Heygen multi-avatar now. This is the testing-grid concept — Heygen makes it the cheapest 2p to produce at volume.

06Tooling prompt index

All beats — Heygen multi-avatar (host + expert), alternating singles / split-screen, native subtitles ON every beat
Product — real prop (expert holds) OR Higgsfield Marketing Studio closeup-product insert (Beat 5)

07Pipeline update block (for concept_pipeline.md)

08Storyboard produced: 2026-06-02 — Her Labs Were Normal. Her Hair Wasn't. (C5) via /storyboard-yapper

Concept ID: C5 (Kindled B.2 batch, testing-grid) Production Target: Heygen multi-avatar (split-screen/sequential, native subtitles) Visual Style: podcast split-screen (Nivara dialogue family) Comp Library Anchors: 1674502353517019 (Nivara 59.3 — exact format match), 1091695962938552 (thebbco 90.2), 566786836079175 (elavate 88.7) Mute-test: PASS STRICT-2p (caption + native subtitles + two-shot; no single face) Production-limitation flags: 1 designed-around (Heygen 2-avatar eyeline → split-screen/sequential); production-feasible now (cheapest 2p at volume)