Fernando DRB — Cheat Sheet

Belief Is the Unit of Value

"Belief is the only thing that compounds." Attention is rented, clicks are incidental, conversions are downstream effects.
Identity rescue is the real conversion event. Before the ad they hold three beliefs at once: something is wrong · I tried reasonable solutions · they failed. The silent verdict is "I am defective." The problem mechanism says you were right to try — they were aimed at the wrong cause. Shame drops, hope returns, and they now want your claim to be true.
You are teaching causality, not facts. Simplistic explanations beat technically accurate ones — understanding beats precision.
Test of a won belief: they can re-explain their problem to a friend in your words. That is why root cause, compression, starvation, blockage recur — they are transmissible.
You don't create desire. The problem, the emotions and the dream outcome already exist in their head. "The only 'new' thing added is your product and the mechanisms."
Big idea = a one-sentence summary of how your mechanisms bridge the big problem to the big promise. Gundry: supplements plug the holes in your gut, you lose weight. ProDentim: good-bacteria gummies heal teeth and gums. Miami MD: strengthen the foundations of your skin, no more wrinkles.
Structure beats words. "A great structure with average words will consistently outperform brilliant copy inside a broken framework. Copy is the skin. Structure is the skeleton."
The AI trap (his 2026 warning): output at record levels, win rates shrinking. "When the base ad is always made by AI, your chances of creating winners drop." Keep AI for iterations, swipes and volume — hand-write the base statics and VSL scripts. Human leverage is "everything that precedes the writing."

Research → 7-Figure Angles

"I spend more time on research than on writing copy when launching a new product test."
1. Google "[problem] explained" and read the top 2-3 articles, especially medical ones written in layman's terms. Note the stated cause, the common remedies, how it feels, and what is physically happening inside the body.
2. Perplexity: "what is the real / hidden / overlooked root cause of [problem]?" Scan studies until you find a mechanism that is simple, easy to visualize, and different from the mainstream belief.
3. Go one layer past the obvious. "Back pain isn't caused by bad posture — bad posture compresses the discs and pinches the nerves. The true root cause is lumbar compression, not posture."
4. Avatar language from Reddit, Amazon reviews and forums. Harvest long journey stories and exact phrases: burning pain, shooting pain, hobbled around, folding laundry, walking up stairs. A whole angle can be built from one activity the problem takes away.
5. Dump every note into ChatGPT and ask for the top 10 most unique and compelling angles. "Any one of these little insights is a gold nugget that can easily become a 7-figure angle."
Broad angle = big problem + avatar. Sub angle = mechanism, or a specific desire / specific problem.
Higher win rate than novel research: swipe angles from the FB Ad Library, and broaden "competitor" to anyone solving the same problem for the same avatar — for arthritis, swipe supplements, red-light devices and creams.
Non-ecom swipe source: buy a $49-149 finpub front end (Agora, Paradigm, Stansberry) and get battered with back-end promos. "Finpub promos are BIG on proof" — the best available training in big ideas and gripping leads.

VSL Architecture & Write Order

A VSL is not an ad — it is an entire sales funnel compressed into one video. The equation: New State of Mind + New Information = New Action.
The 11 points: hook → lead → problem mechanism → dismiss alternatives → introduce the solution → credibility / creation story → solution mechanism → testimonials → end result → skepticism + risk reversal → CTA.
Write inside-out, never top-down: the close first (most formulaic, and it defines the destination) → mechanisms → lead → hook last, because only then do you know exactly which loop to open.
Get structures by transcribing 3-5 competitors already scaling to your avatar and turning their transition points into a 10-point bullet outline. Structures = more creative room, lower win rate, bigger winners. Templates = higher win rate, smaller wins. "The less confident you are in your skills, the more templates you should use."
Never mix writing and editing — "like trying to accelerate and brake at the same time." Draft in flow without judgment, then edit in a separate analytical pass. Minimum 2-3 drafts with time between them.
Conciseness drill: take every two sentences and combine them into one. Repeat.
Target a 5th-6th grade reading level. Field test: rewriting a VSL from 10th to 5th grade gave +37% watch time, 2x conversion rate, ROAS 1.5 → 3.1.
The #1 mistake is naming the product in the first 30 seconds. Delay the reveal to at least 50% of the video — 60-80% in sophisticated markets (Gundry's standard is 70-80%). "The product reveal becomes a relief, not a pitch."
Scene pacing: cut every 1-3s in the first 15 seconds, 2-3s out to the 60-second mark, 3-5s through the educational body, back to 2-3s in the close. A subtle fade every 3rd-4th scene makes rapid cutting read as deliberate.
3D animation is proof, not decoration, and it is mandatory on every mechanism. Same knee-massager script: no animation = 18% hold past 30s; with animation = 32% and +78% conversions. If the script says "crushed under 100 pounds," the animation must literally show downward force.
Emotional congruence is non-negotiable. Agitation: dark, cold, gray, somber, actors who look like the avatar, a red flash for "stabbing." Solution: bright, warm, outdoors, upbeat — and show the product being used inside the happy moment so the emotion anchors to it.
Static end screen, last 10 seconds, after the audio stops (the silence is intentional): the largest bundle front and centre, one high-contrast button, an animated cursor clicking it, logo, one urgent line.

Problem & Solution Mechanisms

Three pillars of a problem mechanism: unique (creates new hope — it explains why everything they tried failed), believable (proof language + a three-part explanation + a neutral documentary tone), compelling (metaphor + 3D animation + emotional tie-in).
Name the villain: Cervical Compression · Cartilage Drought · Barolumbar Pressure · Neuropathic Nerve Erosion · the Benjamin Button Protein.
Name the hero too: cervical traction · vascular revival · Double Compression Zone Technology · Osmotic Absorption Technology · Flash Beam Technology.
Instant-simplify device: introduce the technical term, then kill it in the same breath. "Cervical traction sounds complicated, but it's just a fancy way of saying gently stretching your neck at the right angle." Then chain with "That means…"
Metaphor bank: a bowling ball on the neck · 100 lbs pressing on spinal discs · the sciatic nerve as a highway with a collapsed bridge · inner-ear hairs as seaweed drifting in the tide · joints as an engine with no oil · collagen as scaffolding in a collapsing building.
Take the blame off the viewer — modern lifestyle ("you cannot stop using your phone"), a societal crisis ("the stress pandemic"), or bad authority advice ("your doctor is wrong"). Lifting guilt restores hope, which restores attention.
The belief ladder — the whole VSL in five lines: the pain is real → it is caused by [problem mechanism] → the only fix is [solution mechanism] → the only product that delivers it is yours → therefore the only path to relief runs through you.
Extended mechanism for sophisticated markets: the root cause of the root cause. Gum disease isn't bad bacteria, it's the lack of good bacteria. Knee pain isn't cartilage wear, it's cartilage drought from blocked blood vessels. Biotin isn't wrong — the delivery is; pills never reach the scalp. Anchor the new layer to something they already accept (the gut microbiome) so it reads fresh but believable.
Dismiss alternatives because they miss the root cause, then stack the secondary knocks: too expensive, too inconvenient, wrong delivery method, actively harmful.
Multi-function products: sell each function one at a time (EMS → heat → red light → vibration), each with its own "that means" chain, instead of listing them.
Negative future pacing must be specific. "It will get worse" is nothing. "You may need high-risk spinal surgery that costs thousands and leaves you immobile for weeks" sticks.

Hooks & Leads

The hook's only job is to arrest attention long enough for the lead to work. It does not sell. Land it in 3-8 seconds, and never resolve the emotion in the hook — amplify it. Tension is what sustains attention.
Four levers: curiosity (the master key), fear, desire, shock. Pair curiosity with one primal emotion.
Hook scenes / tangential visuals: curiosity within context — honey poured on a banana, sludge inside a dishwasher, a candle-and-bottle heater. Indirect but thematically congruent, so it slips past the "this is an ad" filter.
Mini-pitch hooks (authority + mechanism + big problem + outcome + timeframe) work best in less sophisticated markets: "Nurse reveals 60-second trick to eliminate toenail fungus almost overnight."
His current top-performing hook production tricks: multi-coloured on-screen captions (red, green, blue, yellow — not just white) · 3D animation inside the hook ("always very strong") · split-screen hooks with animation on one side and a human on the other · a photo slideshow/collage for the first few seconds for a viral, social-proof feel.
Absolute ban on stock footage of people in hooks. Only UGC, documentary/news clips, or 3D animation. Invisible stock (a city skyline) is fine; a polished actor pretending to be in pain is not.
Six lead types: root cause · problem agitation · myth-busting · story (start at the climax, hero worse off than the viewer) · value · open-loop.
The indirect lead — his self-declared secret weapon. Capture curiosity with something only slightly related, promise nothing, agitate nothing (Gundry: the health benefits of peeling potatoes before boiling them → weight loss), then hand off: "But if you really want to know the most effective weight loss method I've ever used, listen up." It works because you haven't asked them to believe anything yet, so advertising skepticism never switches on.
Open loops: give what it is NOT, not what it is. "This doctor shocked social media when she revealed we've been wrong about the root cause of [problem] for 30 years — and it turns out it's not caused by [old root cause]." They get something, but not everything.
How to write leads properly: write the BODY first, highlight the most valuable and shocking lines, then build the lead out of those as open loops. "Hooks are highly effective when they are compact, concise, direct open loops."
Congruence test: write the last line of the hook and the first line of the lead as one sentence. Reads naturally = congruent. Reads as a jump = rework it.
Volume standard: 5-10 hook variations per VSL. Full VSLs launch with a minimum of 3 leads x 3 hooks = 9 ads, split one batch per lead.

Awareness & Sophistication

Two separate dials. Awareness sets what you say. Sophistication sets how bold you can say it. Running one script across all audiences is the single most common cause of a good VSL underperforming.
L1 Unaware — create the problem. Slow-build stories that sneak the awareness in, plus big bold claims (competition is minimal, so you have permission).
L2 Problem aware — agitate by dimensionalizing: physical ("a stabbing pain that shoots down your lower back when you bend") → real-life consequence ("you can't kneel to play with your grandchildren") → emotional ("trapped in your own body"). Test three flavours: fear-heavy, frustration-heavy, embarrassment/loss-heavy.
L3 Solution aware — stop selling the problem. Lead with a named solution mechanism plus novelty ("newly discovered"), then systematically dismiss the alternatives.
L4 Product aware — a comparison fight. Differentiate on the named mechanism, dismiss competitors specifically (cost, cheap build, greasy residue), load proof density, move fast to the close.
L5 Most aware — hide the pitch inside a story, delay the reveal to 70-80%, and go one mechanism layer deeper.
Diagnostics: great hook rate + terrible hold rate = the lead is pitched at the wrong awareness level · high CTR + no sales = claims too aggressive for the sophistication level · fatigue across all angles at once = the market moved up a stage.
When aggressive claims stop working, that is not a copy problem. The fix is a deeper mechanism and a more indirect lead — not a punchier headline.
The 3rd-person compliance rule (make it default): keep negatives in 3rd person, positives in 1st. Not "if you have back pain, you're a prisoner in your own body" but "Women with back pain say it's like being a prisoner in their own body." "The less you use the word 'you,' the more compliant you are with FB and the FTC" — and it rarely costs conversion rate.
Attach benefits to the mechanism, not the product. Not "our supplement gives you more energy" but "research shows this molecule can increase energy levels." The viewer draws the connection; you stay compliant.
In subtle mode, sell the secondary benefits first — affordability, convenience, ease of use — before the big health claim.

Proof & The Close

Proof is seasoning, not the main event. "Proof must confirm a desire, not attempt to create one." Dumping testimonials, charts and clinical studies to carry belief by volume backfires — it reads as staged.
Three layers, in this order. Authority early and light (Gundry mentions being a heart surgeon up front — just "this is worth listening to"). Scientific/logical in the middle, after agitation and after the mechanism. Social late, at the close, when they are emotionally ready.
No trials of your own? Prove the mechanism instead — cite the red-light or EMS literature. And always translate: not "downregulates progerin expression" but "women who used this looked five to ten years younger within 90 days."
Testimonials must be raw and must acknowledge skepticism: "At first I thought this was another gimmick. But now I can walk three miles pain-free." Specificity beats adjectives — "a huge, embarrassing neck hump from years at the computer" beats "neck pain."
The close, not the hook, moves the most money. "Attention without action is just giving money to Facebook."
Repeat the CTA 5-7 times by the end. Put the risk reversal on the button itself ("Start My 90-Day Risk-Free Trial") and bake the discount in ("Get Up to 51% Off").
Name the guarantee ("The Results or It's Free Promise"), make it almost unfair (90/120/180 days, no return required), and always give a reason why it is that bold. Then reframe the decision: "the only decision you make today is to say maybe."
Crossroads close: two vivid futures — worsening condition, surgery, invisibility vs freedom and relief. Loss aversion makes not buying feel like the risky choice.
Scarcity needs a logical reason why — a rare herb harvested twice a year, hand assembly at 500 units a week. Without the reason it reads fake.
Value reinforcement: anchor against the expensive alternative ($200 device vs $3,000 of physical therapy), ladder the hypothetical price ($400 → $200 → $150 → $99), give a reason-why for the low price, and daily-cost frame it ("less than $3 a day"). 50-51% off is the ecom standard.

Ad-Type Doctrine (What's Working)

Three top-of-funnel types in ecom/supplements right now: image ads → listicle or advertorial · VSLs · performance shorts. Plus DR UGC. "The future of performance marketing is more variety of creatives top of funnel and strong BOF retargeting-style ads for offer-aware traffic."
The 5-10 minute VSL is the gold standard — "crushes across niches and levels of market awareness." 2-3 minute mini-VSLs run the same structure condensed, for markets that don't need much education. The 30-60 minute monster VSL is showing up less in feed. He runs VSLs to advertorials on $100+ AOV offers.
Performance short = 30-60 seconds, the same VSL structure extremely condensed, every sentence carrying a DR element. Visuals are either a mashup of viral-looking action clips with minimal b-roll, or a spokesperson holding most of the scenes. Use it where a VSL can't be profitable — low AOV, or markets that won't sit through long form.
Educational / documentary copy is his preferred lane (Gundry MD, NeBroo, Miami MD): no personal emotional spokesperson, no telling them how they feel — "just information flowing through a structure," held together by open loops and incomplete information.
Emerging format: long-form indirect story ads whose hook is an emotional story completely unrelated to the product ("I almost said no to the boy who saved my husband's life"). Use the story hook to lower CPMs, then transition into the proven DR body — "Here's what he told me…"
The cheap angle test — 10 mini-VSLs, not 10 funnels. Build one general problem-focused advertorial or listicle, then 10 mini-VSLs (1-3 min) for 10 angles. ABO, 10 ad sets, one per angle, $50-200/day each, 3 hooks per angle. AI writes all 10 scripts in under 2 hours; an editor produces all of them in 3-5 days because the scripts repeat scenes.
"The goal isn't to get something wildly profitable; it's to see which 1-2 angles out of 10 perform best. There is no way to lose; you win by narrowing it down." Then build the full 5-15 min VSL with hook/lead variations, a customized advertorial and a congruent PDP — for the survivors only.
Keep the same mechanism across all 10 angle tests, unless a new mechanism is what you're testing.
Multi-page delivery: run the same creative from an authority page, a brand page, and a neutral community page ("Back Pain Support Network"). Reported: brand-only ≈1.7 ROAS, authority ≈2.3, and adding community pages gave 30% longer creative lifespan plus lower CPM.

Image Ads: The Qualified Curiosity Click

The inversion: unlike a VSL, the image ad shares none of the selling job. Its only job is a qualified curiosity click. Which is why image-ad traffic needs a long informational advertorial with a long close — short pages starve it.
Clickbait, but qualified. Image, headline and copy all point at the problem, never the fix. The product is never named or shown.
Four styles: pure weird ingredient (a tomato in a pot with a Red Bull can, ACV mixed with pickles) · split-screen half weird / half problem — often beats pure weird because it adds self-identification on top of curiosity · problem-relevant but weirdified (purple slime dripping from an ear, a spine with glowing red animation) · organic lifestyle native post, which lowers CPM because Meta reads it as organic.
His own best format: a native-looking photo enhanced with AI showing something unorthodox — "rubbing baking soda with a palm leaf into the knee" — or a fully AI-generated image that is bizarre but still relevant to the problem.
Copy formula: Hook → Problem → Failed Attempts → Discovery → Cliffhanger → CTA. Medium length, 200-350 words, a first-person avatar story starting mid-action, pivoting on a recommendation from someone they trust.
Headline 6-10 words, conspiracy / exposé framing: "The Back Pain Hoax" · "Whistleblower Exposes Truth About Back Pain" · "Scientists Stunned by Memory Trick." Button CTA is always "Learn More."
Tease the fix as a weird specific: "a five-minute nighttime trick," "a strange ice method," "a simple jelly bean butter fix." The weirder and more specific, the higher the curiosity — and never name the product.
Always resize or slightly modify a scraped image before uploading, so Meta doesn't dedupe it against whoever else is running the same file.
AI generation sequence: paste the strategy doc first → add up to 5 visual references → brainstorm 10 ideas, three separate times (asking for 30 at once collapses quality) → generate one at a time at 1080x1080 → iterate on lighting, age, camera angle. No logos, no text overlays, keep the imperfect native look.
His longest-running images stayed live 3-4 months, and he has scaled offers to seven figures on images alone.

The Batch Engine — Exploit / Explore / Core

A batch is one ad set with one creative mission, named by number, not description. "A105" ties together the script doc, the Drive folder, the audio, the editor output, the assembled ads, the ad set and every tracking sheet. Ads are A105-V1 / V2 / V3; the winner is logged as A105-V1.
Five creative types by effort vs win rate: Hook Scenes (same hook audio and script, new visuals only — lowest effort, highest win rate) · Hooks · Hooks + Leads · new VSLs (1-2 a month max — highest effort, lowest win rate, biggest upside) · Image Ads.
Track win rate BY ad type so you can reallocate — hook scenes might run 70% where hook+lead batches run 50%.
Batch composition: hook scenes = 1 winning hook script → 5 visual variations (min 3 approved) · hooks = 3 scripts (AI writes 20, copy chief picks 3) · hooks+leads = 1 new lead + 3 congruent hooks · new VSL = 3 leads x 3 hooks = 9 ads, split into three batches, one per lead.
Output target: 4 batches per product per week, 3-5 ads each ≈ 12-14 ads/week/product.
Variations are only ever created from validated winners. Never spend production on variations of an unproven hook or lead.
EXPLOIT (small variations on a winner): winning image → 3 batches of that image + 1-2 batches of same-thesis, different-content copy. Winning video → 3 batches of new hooks + 1 batch of hook scenes + 1 reskin if it hasn't had one, aggregating all the top hooks.
EXPLORE (bigger swings on a winner): 3 new leads with 3 new hooks each; or keep the message and lead idea identical and re-execute it in a new style and structure — 1st-person UGC vs authority spokesperson vs TED talk.
CORE (new territory): a new pain point + avatar combination, sometimes requiring a new mechanism. Crepey skin for older women → for single women → arms / neck / thighs / face / hands / belly → cellulite reduction (new mechanism) → body sculpting (new mechanism).
The routing question: "Is the idea big enough to make a whole new body, and possibly a lander and funnel, for? If not, put it in a lead or hook."
Hook stacking: on a proven VSL, stack several winning hooks back-to-back in front of the winning lead. "I do this all the time. Very effective, and I highly recommend" — Golden Hippo brands run it on everything.
Image loop: once an ad copy wins, pump a fixed number of image tests against that copy every week until the copy burns out. No iterations on the images themselves — instead use the winning copy to inspire brand-new copy with the same thesis (beauty + romance stories: "my ex-boyfriend didn't recognize me… and three different men asked for my number in one night").
Track cost per winner, not just win rate. "$20k spent over 4 weeks, 4 winning ads = $5k per winning ad." Cost per winner is what actually decides when you can raise volume and grow the team.

Reviving a Dying Winner

Diagnose the shape of the drop first. A sudden death on a high-spend winner (over $100k) is usually a compliance flag from Meta's AI, which penalizes the ad with lower-quality traffic at the same CPM — not audience fatigue.
Compliance-flag fix, VSLs: small unnoticeable changes to bypass the algorithm — run the video at 0.95x speed, alter the audio pitch, swap a few scenes. Nothing about the copy changes.
Compliance-flag fix, image ads: test a high volume of new images against the proven copy, rewrite the hook and lead, and add a story to the top of the existing ad body.
A slow decline instead of a cliff is real fatigue. Climb the ladder cheapest-first: hook scenes → new hooks → reskin → new leads → new body. Only escalate when the cheaper rung fails.
Reskin = the one variation that changes everything visible and nothing written: all-new visuals, voice change, script untouched. Aggregate the top-performing hooks into it.
The 7-day rule: if a winner declines for 7+ consecutive days, shut it off. You can retry a week later; if it's still weak, retire it. Stay ahead by always running new hooks and angles in the background.
The "power up" pass for a mediocre-but-not-dead VSL: simpler language, more concise problem descriptions, a better spokesperson, fresh leads and hooks. He took a 6-minute VSL from 1.5 → 3 ROAS this way, with CPM and CPC both dropping over 50%. "None of these things were huge changes… but they brought out the power of the big idea."
Check the funnel before you rebuild the ad. Advertorial CTR 30-40% · PDP CTR 50% price-shown / 80% price-hidden · checkout 8-12% at ~$100 AOV. "You are not losing clicks — you are losing revenue you already paid to acquire." (Kill and scale ladders live on the Media Buying Hub.)
Don't diagnose off soft metrics in testing: "I don't recommend using metrics like hook rate and hold rate as KPIs in the testing phase. Only ROAS." Use them only to compare creatives against your own winners' baseline — his floors are 40%+ hook rate and 20%+ hold rate, where hold = watched 15 seconds or more.
Ads that show early promise then die after a few weeks are usually not long-term winners — but test the price, the offer and the advertorial before you re-test ads.

Advertorial & Funnel Copy

Purpose: mimic a third-party news or blog article so subconscious pattern recognition lowers their guard. "The more aware the market, the better advertorials work" — which is why nearly all financial and supplement offers use them.
Three above-the-fold elements, in order: headline (+ subhead), hero image (+ caption), lead. The headline is "the most important part BY FAR."
Headline levers: curiosity (new discovery · real root cause revealed · taking the US by storm) · benefit/outcome · authority ("XYZ University just revealed," "[Expert]: '[claim]'") · timeframe ("…in 14 days").
Lead with what they GET. Not "This new viral product has helped 12,700 Americans get relief from back pain" but "Get relief from back pain with this new viral product that 12,700 Americans SWEAR is the 'spine savior'."
Congruence rule: the advertorial headline must match the ad's messaging and big idea, or you pay for the click and lose it to the disconnect.
Hero image: something must be happening. Three that work — before/after split-screen · the problem on one side and a strange image of the problem mechanism on the other · news-style.
Story lead, 7 beats: first-person emotional agitation → a friend or authority recommends it → skeptical but tries it → immediate results → longer-term results → daily-life benefit → better than ever. Use "she's like me but WORSE" — the hero must be further down than the reader.
The body is formulaic and near-identical across all advertorials: problem mechanism → solution mechanism with proof → an authority discovers or invents the product → the product is put to the test and works → sold with benefits, promises and reasons-why → close (social proof → urgency → scarcity → risk reversal → crossroads → clear CTA → bundle push with a reason why → 1-2 sentence restatement → CTA #2).
Advertorial vs listicle: "generally advertorial is best for VSL." If a listicle is already winning on cold traffic, keep the variables separated — otherwise a weak advertorial kills good VSLs. Listicle CTR usually runs higher (60%+).
The PDP flips the focus: the VSL and advertorial were about the problem; the product page is about the product, and its job is to channel all the tension onto the button. "I've taken funnels from losing money to 25% profit margin just by changing the button text, hero image, and product title/headline."
Price reveal decision: show the price if the VSL already delivered the entire sales process; if not, use "See Offer" / "Check Availability" so they arrive at checkout more committed.
Offer economics — bundles, order bumps, upsell slots and take rates — live on the Media Buying Hub.

⚡ When Stuck

Big winner died overnight → assume a compliance flag, not fatigue: 0.95x speed, pitch shift, swap a few scenes, relaunch. Don't rewrite the copy.
Winner sliding slowly → climb cheapest-first: hook scenes → new hooks → reskin → new leads. 7 straight declining days = shut it off.
Great hook rate, dead hold rate → the lead is aimed at the wrong awareness level. Write a new lead, not more hooks.
Everything fatiguing at once across all angles → the market moved up a sophistication stage. Go one mechanism layer deeper, not louder.
Clicks but no sales → audit advertorial CTR, PDP CTR and checkout conversion before touching the creative.
Sitting on 10 angle ideas → don't build 10 funnels. 10 mini-VSLs into one general advertorial, $50-200/day each, keep the top 1-2.
Output climbing but win rate falling → too much AI base copy. Hand-write the next batch of statics and VSL scripts; keep AI for iterations.
Can't find a new angle → Google "[problem] explained," then Perplexity for the hidden root cause, then Reddit for the exact words they use.
Blank page on a VSL → write the close first. Then the mechanisms. The hook is last.
Lead won't open cleanly → write the body first, highlight the most shocking lines, and build the lead out of those as open loops.