RAW Chemistry — Funnel-extremes creative strategy (problem-solution TOFU + objection BOFU)

Date: 2026-09-08 (rev 2, compliance commentary removed at operator instruction — RAW's own team owns claims review). Author: producer (Claude). Status: strategy + ad-package spec, pending operator review. Operator brief (from a Sam Perillo video): target the extreme ends of the funnel, skip the middle. TOFU = problem-solution ads built from an avatar framework (ICPs → every problem they experience, even indirectly related → one pain per ad package → product as the direct unique solution; Andromeda matches the ad to the people who feel that problem). BOFU = objection ads (why people abandon checkout: price, warranty, competitor comparison… → ads that dismantle each objection). Catalog retargeting as an optional baseline.

Terms. TOFU = top of funnel: the cold stranger who has never heard of RAW and may not even have named his problem yet; the problem-solution ad. BOFU = bottom of funnel: the man who has already looked, maybe carted, and didn't buy; the objection ad. MOFU = the middle (consideration, retargeting with more of the same), which this strategy skips.

1. What the framework gets right for RAW, and where it needs adapting

Right, and matches RAW's own evidence.

• Andromeda is a retrieval model: it picks which ad to show which person from the creative itself. One problem per ad, stated in

the language of the person who has it, gives the model a clean signal. Blended "premium mood + offer" ads give it nothing to match on. RAW's 20-day audit and today's TOF mechanism audit both concluded the account's gap is a persuasion/messaging gap, not an audience or bid gap (research/tof-mechanism-audit-2026-09-08/REPORT.md, meta-20-day-audit-2026-09-05/REPORT.md).

• Skipping the middle is right for RAW specifically: the repeat rate is ~4.4%, so every ad must pay back on order one. There is no

nurture economy to fund a "consideration" layer.

• RAW's purchase pockets are already avatar-shaped: men 65+ (married, long-term) converted at ~$44 CPA on $444 in the audit window;

men 25–44 are converting in the price test at $22–43 CPA; men 45–64 burned $799 for five purchases. That is three different problems, not one audience.

Adaptations.

• "Position the product as the direct unique solution" means saying the mechanism out loud: RAW is pheromone-infused, it works

with his body chemistry, she picks up on it before she knows why, and it lasts the night. Every package states what RAW is, what it does that his cologne doesn't (the failed alternative is always his cologne), and backs it with 12,400+ reviews, 500,000+ bottles since 2016 and the 60-day guarantee. Claims review belongs to RAW's own compliance team, not to this document.

• "Warranty" objections translate to the 60-day empty-bottle guarantee that is live on the page.
• The register stays hard where the avatar is hard (25–44) and turns human where the avatar is married and older. Same pain

family, different vocabulary.

• Catalog ads: low priority. Sixty percent of last-30-day abandons are single standard bottles at $35–37 from organic/Google traffic,

not the Meta funnel; a DPA can run as a baseline but the objection ads will do the real BOFU work.

2. Evidence base

3. The avatars (ICPs) and every problem they carry

Ranked by evidence of purchase. Problems are listed exhaustively on purpose: Andromeda needs distinct pains to match on, and "indirect" problems are where the untested angles live.

Avatar A — The long-married man (55–72). "She stopped reaching for me."

Evidence: RAW's best-proven pocket (65+: 10 purchases / $444, $44 CPA; 7 of 11 purchases in the Sep 3 read). GODA's newest wave targets exactly this man ("My wife hasn't reached for me first in nine years", "I'm 49. My wife finished twice last night"). UGC-29/-35/-49/-50: wives reacting; "my wife's been just a little more interested." Context: married 20–40 years, kids gone or leaving, routine bedroom, still attracted to his wife, doesn't talk about it to anyone. Problems (direct → indirect):

1. She never initiates any more; he's always the one reaching.
2. Affection has become logistics: a peck, a pat, "night".
3. He feels invisible at home — noticed only when something needs fixing.
4. Sex is scheduled, rare, or gone; nobody names it.
5. He's tried "date nights" and gifts; they buy a nice evening, not desire.
6. He wears the same cologne he wore in 2009; she stopped commenting years ago.
7. Aging body: gut, grey, softer — he suspects that's the reason and it's the one thing he can't fix by Friday.
8. Fear of being ridiculous if he "tries" (new cologne, new shirt) and she notices the trying.
9. Fear his wife's disinterest is about him specifically, not the marriage.
10. Retirement/empty nest: more hours together, less spark; the contrast is louder.
11. He won't buy anything that smells like a teenager's body spray or looks like a sex product on the bathroom shelf.
12. He doesn't trust internet ads; he trusts numbers, guarantees, and men his age saying it plainly.
13. Erectile/health worries in the background (do not touch in copy; it shapes his defensiveness).
14. Money is not the issue; embarrassment is. He'll pay $99 for something discreet that works.

Avatar B — The divorced re-entry man (45–60). "The bar got younger. I didn't."

Evidence: 45–64 is RAW's most expensive cell historically ($160 CPA in the audit) but the price test's 45–64 set converted 4× at $31 in the last 5 days — the age isn't wrong, the message was. Cupids' advertorial hook ("You spent good money to smell expensive. So why does she just say 'nice cologne' and walk away?") lives here. Context: separated/divorced in the last 1–5 years, on apps for the first time, out with younger colleagues, self-conscious. Problems:

1. Apps: matches that die after two messages; he suspects his in-person presence doesn't match his photos.
2. First dates end with "this was nice" and no second.
3. He gets "nice cologne" as a compliment that closes the conversation instead of opening it.
4. He walks into a bar and nobody registers him; younger guys with less get attention.
5. He bought the designer bottle the salesman recommended ($150–$250) and it changed nothing.
6. His ex's disinterest still lives in his head; every non-reaction confirms it.
7. He doesn't know the rules any more — what to wear, how strong is too strong, what women his age or younger want.
8. Terror of coming across as trying too hard or "creepy"; he under-plays everything.
9. Weight, hairline, energy — the body doesn't cooperate with the new life.
10. Time: he has two nights a week, and each blank night costs more than the money.
11. Friends are married; nobody to ask, so he asks the internet at 1 a.m.
12. He's skeptical of anything that sounds like a pick-up gimmick; he'll buy a "product" not a "trick".

Avatar C — The going-home-alone guy (24–42). "Same suit. Same haircut. Same $90 cologne."

Evidence: cheapest CPM in the account ($29–47 in feed); PT 25–44 set converting at $22–43 CPA; GODA's crude tier ("THE ONLY COLOGNE THAT F**KS" 164 days, "INCREASE BODY COUNT" 144 days) proves the register survives; UGC-01/-23/-24/-28 are this guy. Context: bars, clubs, apps, gym, weddings; goes out with friends; spends on sneakers and cologne; measures nights by outcome. Problems:

1. Goes out every weekend, comes home alone most of them.
2. Interchangeable: same fits, same haircut, same Sauvage/Bleu as every other guy in the room.
3. Gets compliments, not pull — "you smell good" from a friend, nothing from the woman he wanted.
4. Texts that get "Delivered." and nothing else.
5. Apps: swipes into the void; matches who never meet.
6. His cologne is gone by the second bar; by 1 a.m. he smells like the room.
7. Sweat/gym: he wants to smell good at the gym without smelling like he tried.
8. Approach anxiety; he stands next to her and says nothing, then blames the night.
9. Spent $200+ on a designer bottle because the ad had a model in it; it did nothing his $40 bottle didn't.
10. Wingman economy: his friend always gets the attention.
11. Wants an edge that isn't a "line" or a pick-up-artist routine.
12. Wants the outcome named bluntly; euphemisms read as weakness.
13. Suspicious of "pheromone" as snake oil, but willing to try for $40 if the proof is loud and the guarantee is real.
14. Wants it to look good on the shelf/bathroom counter; the bottle is part of the flex.

Avatar D — The woman buying for him (30–60). "I want him to smell like that again."

Evidence: For Her / gift-set lines appear in 26 of 90 abandons; UGC-35/-41 ("I make him wear this every single day", "every woman wants their man to smell good"); female purchase cells too small in Meta but the store sells to women organically. Context: wife/girlfriend shopping for a birthday, anniversary, Valentine's; wants him to make an effort without nagging. Problems:

1. He has worn the same cologne since they met; she's bored of it and won't say so.
2. She wants to feel the pull again and can't buy that in a card.
3. Gifts for men are impossible: another wallet, another whisky.
4. She wants him to feel attractive because his confidence is her problem too.
5. She'll buy for him if the ad speaks to her, not to a club kid.

(Indirect avatar — one or two packages, women-targeted creative, softer register.)

Avatar E — The fragrance guy (25–50). "Tell me what it smells like."

Evidence: UGC-09/-54/-11 (notes: bergamot, pink pepper, vetiver, cherry, vanilla, musk). Low volume; useful for objection handling ("what does it smell like?"), not a TOFU lead.

4. Problems with the brand/product itself (all avatars) — these become BOFU objections

1. "Pheromones are placebo / snake oil." (public reviews: "we all smelled the same", "no pheromone effect")
2. "$99 for cologne from an ad?" — price shock; unknown DTC brand.
3. "Which one do I buy?" — Raw vs For Him vs Midnight vs gift set vs roll-on; three-bottle bundle logic unclear.
4. "Will it actually smell good?" — no scent description in most ads.
5. "It'll wear off in an hour like everything else."
6. "Is this an AI ad / fake brand?" — generated imagery, GODA-style claims.
7. "What if it doesn't work — can I return it?" — guarantee mechanics unstated.
8. "Shipping time / where from?"
9. "Do I need three bottles?" / "Why is the single bottle $39.90 but the ad says 78% off?"
10. "She'll know I'm wearing a 'pheromone' cologne" — embarrassment/creepiness.
11. "Amazon/Walmart have it cheaper / knock-offs."
12. "I already have cologne — do I replace it or layer it?"
13. "Is it safe / natural / what's in it?"
14. "Subscription trap?" (Cupids/GODA funnels condition buyers to expect one.)

5. TOFU ad packages (problem → solution)

Rules for every package: one avatar, one named problem in the first line, the failed alternative is his cologne, RAW stated as pheromone-infused cologne, one credible proof, one CTA, offer as support only. Each package ships as 3 executions (three different images/scenes, same words) so Andromeda has variants; 4:5 + 9:16 per execution; one ad per execution, no DCO. Register column: H = hard, W = warm. Format: S = static, T = text/UI, V = video.

Andromeda hygiene for this set: 16 problems, 5 avatars, 3 formats, two registers — that is the diversity the model needs. Ship in waves of 6 packages × 3 executions (18 ads) into one broad purchase ad set; never split by age (let the model match); keep the PT age cells running as the control until the broad set proves itself.

6. BOFU objection packages (dismantle → convert)

Audience: site visitors 14 days, ATC/IC 14 days, video 50% viewers 14 days, Instagram/FB engagers 30 days; exclude purchasers 60 days. Each package: the objection in the customer's words as the hook, the answer, proof, CTA back to the same page.

7. Structure, budgets, decision rules

• TOFU campaign: Sales/Purchase, one broad US men 25–65+ ad set (no age split; Advantage+ audience per the account lock),

wave 1 = 18 ads (6 packages × 3 executions). Provisional $150/day. Keep the PT age-cell ad sets as the control for one week.

• BOFU campaign: Sales/Purchase, one retargeting ad set (windows above), 6–8 objection ads, $40–60/day; catalog DPA at $15–20/day

as the baseline for the store-wide single-bottle abandoners.

• Judge on first-order CPA and revenue ROAS against the price-test kill line (1.0 ROAS on ad spend is the operator's accepted

test threshold; $45 planning CPA), LPV→purchase and ATC→purchase — not CPM, not thumb-stop. Zero-purchase stop at 3× target CPA per ad set after an attribution check; do not restructure on a single day.

• Naming: TOFU | <avatar> | <T#> <problem slug> | <execution> | <ratio> so the ad-level read tells you which problem is

matching. BOFU: BOFU | <B#> <objection slug>.

• Production route: Codex built-in image_gen with the five RAW references via referenced_image_paths, verified per session;

matched-scale bottle check before type; deterministic type in Remotion (new systems per package, never the V85 template); audit_v85 gates; pair gate + blind judge before launch. Copy goes to RAW's compliance team as part of the launch pack.

8. First 30 days

1. Wave 1 (this week): T1, T2 (built), T5, T10, T3, T7 → 3 executions each. B1, B2, B4, B6 for retargeting.
2. Fix T2 line 2 ("you were right there" is unclear → "at the bar till close if you want").
3. Wave 2 (days 8–14): T9, T11, T12, T15, T8, T14; B3, B7, B8, B9.
4. Day 15: read by package, kill packages with zero purchases past 3× CPA, double the executions on any package with 2+ purchases.
5. Days 16–30: T16 video challenger vs reaction control; scent card B4 into TOFU as a variant for Avatar E.

9. Open items the operator should decide or confirm

• The three scent notes for B4 (bergamot, pink pepper, cherry-vanilla/musk) come from creator descriptions; swap in the official notes if different.
• Whether Avatar D (gifting women) gets one package in wave 2.
• Whether the BOFU audience includes the organic single-bottle abandoners (recommended: yes, that's most of the abandonment).