# MANUAL.md — Conversion Copywriting

> The master's-degree reference for the **master-copywriter** agent.
> Discipline: **Conversion Copywriting** — the words on any single persuasion surface.
> Every school, framework, and law below is attributed to the real person or company who originated it. Nothing here is invented. Where a claim is a rule of thumb rather than a measured constant, it is marked as such.

---

## 0. What this document is

This is the operating theory, the science, the architecture, the studied schools, the decision engine, the playbooks, the metrics, the failure rubric, and the wiring diagram for the copywriting master. It is written to be read once end-to-end by a new instance of the agent and then used as a lookup table forever after.

The agent it governs owns **words on a single surface**. It does **not** own funnel structure (that is `funnel-psychology`), and it does **not** own send cadence or list mechanics (that is `email-marketing`). It receives a stage-and-awareness spec and hands back finished, gate-passing copy.

---

## 1. Mandate & mental model

### 1.1 What this master owns

The **WORDS on any single persuasion surface**: sales pages, VSL (video sales letter) scripts, ads, headlines, hooks, hero sections, email subject lines treated as standalone artifacts, landing-hero variants. One surface at a time. The unit of work is a *page* or a *beat sheet* or a *variant set* — never the funnel, never the calendar.

### 1.2 What it does NOT own

- **Funnel structure** — the sequence of surfaces, the offer ladder, the stage map. Comes *in* from `funnel-psychology`.
- **Send cadence and list mechanics** — who gets what, when, how often. Handed *out* to `email-marketing`.
- **The offer itself** (price, guarantee terms, stack contents) — those are given as inputs; the copy dresses them, it does not set them.

The master never invents funnel structure or sending cadence. If a brief is missing the stage/awareness spec, the master **stops and requests it** rather than guessing — guessing awareness stage is the single most expensive error in the discipline (see §3).

### 1.3 The core operating theory (five load-bearing beliefs)

1. **Desire is pre-existing; copy channels it, never manufactures it.** — Eugene M. Schwartz, *Breakthrough Advertising* (1966): "Copy cannot create desire… It can only take the hopes, dreams, fears and desires that already exist in the hearts of millions of people, and focus those already existing desires onto a particular product." Every job starts by locating desire that is already there.

2. **Enter the conversation already in the prospect's head.** — Robert Collier, *The Robert Collier Letter Book* (1931). You do not start the conversation; you continue it. This is the master's *first* instruction and the philosophical root of message-to-market match.

3. **Directness must scale to awareness.** — Michael Masterson (pen name of **Mark Ford**) & John Forde, *Great Leads* (2011), operationalizing Schwartz. The more the prospect already knows, the more directly you may sell; the less they know, the more indirect the entry.

4. **Proof is the bottleneck, not cleverness.** — Gary Bencivenga (*Bencivenga Bullets*; the *Bencivenga 100* seminar). In a market drowning in claims, belief is the scarce resource. Every claim carries proof or it does not ship.

5. **Fluency is misread as truth.** — Adam Alter & Daniel Oppenheimer (2009); Rudolf Flesch (readability, 1946–49). Copy that *reads* easily *feels* truer. Readability is therefore a persuasion lever, not a courtesy — and it is mechanically measurable, so it becomes a gate.

The master's job, reduced to one sentence: **locate the prospect on the Schwartz grid, enter the conversation already in their head, choose the lead type their awareness demands, and carry unquestionable proof for every claim — in a voice that reads so easily it feels true.**

---

## 2. The science — the psychology, behavior, and statistics under the craft

Every rule in this manual traces to a named researcher or a named practitioner-theorist. This section is the warrant.

### 2.1 Eugene M. Schwartz — desire is exogenous (the keystone)
*Breakthrough Advertising* (1966). Two claims: (a) desire is pre-existing and can only be **channeled**, and (b) the message must match the market's current mental state along two axes — **Awareness** and **Sophistication** (the grid, §3). This is a demand-side theory of persuasion: the copywriter is a lens, not a generator. Everything downstream hangs on it.

### 2.2 Robert Cialdini — the principles of influence
*Influence: The Psychology of Persuasion* (1984); *Pre-Suasion* (2016). Seven levers, each a copy device:
- **Reciprocity** — free value first (lead magnets, the "give" before the ask).
- **Commitment & Consistency** — micro-yeses; the mechanical basis of Sugarman's slippery slide (§4).
- **Social Proof** — testimonials, user counts, "join 40,000…".
- **Authority** — credentials, endorsements, and *mechanism explanation* (explaining how it works reads as expertise).
- **Liking** — similarity and candor lower resistance.
- **Scarcity** — deadlines, caps, limited runs.
- **Unity** — shared identity, the "we," the in-group. (Added in *Pre-Suasion*.)
- **Pre-Suasion** itself: the moment *before* the message — the frame and attention channel — shapes receptivity. The empirical cousin of Collier's "conversation in the mind." Langer's *"because"* experiment (a reason attached to a request raises compliance even when trivial) is the lab version of Hopkins's reason-why.

### 2.3 Daniel Kahneman & Amos Tversky — Prospect Theory
"Prospect Theory" (*Econometrica*, 1979); Kahneman, *Thinking, Fast and Slow* (2011). Four copy-critical results:
- **Loss aversion** — losses loom roughly **2×** larger than equal gains. Frame the cost of *inaction*, not only the gain of action.
- **Anchoring** — the first number sets the reference point. State the high "value" before the low price.
- **Framing** — gain vs. loss framing flips choices on identical facts.
- **System 1 / System 2** — copy sells to fast, intuitive System 1 and *reassures* deliberate System 2 (proof, guarantees, specifics exist to quiet System 2's objection).

### 2.4 Chip Heath & Dan Heath — Made to Stick (SUCCESs)
*Made to Stick* (2007). Memorability = **S**imple, **U**nexpected (violate a schema → open a curiosity gap), **C**oncrete (sensory, testable), **C**redible (proof, the "Sinatra test," sticky stats), **E**motional (make them feel; appeal to identity), **S**tories (simulation + inspiration). This is the checklist for whether a *big idea* will survive contact and be retold.

### 2.5 Robert Zajonc — Mere-Exposure Effect
"Attitudinal Effects of Mere Exposure" (*JPSP*, 1968). Repeated exposure raises liking with **no argument at all**. Copy corollary: the **Rule of One** — one idea, repeated across the piece and across the funnel, beats many ideas stated once. (Cross-surface frequency is `email-marketing`'s to schedule; the master's job is to keep the *core claim* singular and repeatable.)

### 2.6 Adam Alter & Daniel Oppenheimer — Processing Fluency
"Uniting the Tribes of Fluency to Form a Metacognitive Nation" (*PSPR*, 2009). Ease of processing (legible type, short words, high contrast, rhythm) is **misattributed** as truth, value, and safety. Adjacent: **McGlone & Tofighbakhsh (2000)**, "Rhyme as reason" (*Psychological Science*) — rhyming aphorisms are judged *truer* than their non-rhyming synonyms. Warrant: fluency (a feeling) is mistaken for validity (a fact). This is the science under the readability gate.

### 2.7 Rudolf Flesch — readability, made measurable
*The Art of Plain Talk* (1946), *The Art of Readable Writing* (1949). **Flesch Reading Ease** (0–100, higher = easier; direct-response targets ~60–80 ≈ 7th–8th grade) and, with **J. Peter Kincaid**, the **Flesch–Kincaid Grade Level** (built for the U.S. Navy). Both are mechanically computable — this is what turns "write simply" into a linter (§8).

### 2.8 Blair Warren — the One Sentence Persuasion Course
*The One Sentence Persuasion Course — 27 Words* (2004). The sentence: *"People will do anything for those who **encourage their dreams, justify their failures, allay their fears, confirm their suspicions,** and **help them throw rocks at their enemies.**"* Five identity-protective moves. The two most under-used in copy: **justify their failures** ("it's not your fault" — removes shame, opens belief) and **help them throw rocks at their enemies** (the villain / us-vs-them frame). Underlying science: motivated reasoning — people move toward messages that protect their self-concept.

### 2.9 Gary Bencivenga — proof as the persuasion bottleneck
*Bencivenga Bullets*; *Bencivenga 100* seminar. **The Persuasion Equation:** *Urgent Problem + Unique Promise + Unquestionable Proof + User-Friendly Proposition (+ Unbeatable Price + Urgency) = Persuasion.* **Proof-elements checklist:** demonstration, the (creative) guarantee, reason-why, specifics, the exposé, explain-the-mechanism, sell-against-type, acknowledge-disbelief, highly-believable source, testimonials. This is the operational form of Cialdini's authority/consensus, and it is the "proof for every claim" law in the mandate.

### 2.10 Adjacent science the master keeps on the shelf
- **BJ Fogg** — *Behavior = Motivation × Ability × Prompt* (the Fogg Behavior Model). The CTA-friction lens: a call to action fails if any of the three is missing. Reduce required ability (friction) and add a clear prompt before blaming motivation.
- **Richard Petty & John Cacioppo** — the **Elaboration Likelihood Model**. Central route (argument/proof) wins with motivated, able readers; peripheral route (heuristics — authority, liking, social proof) wins under low elaboration. Tells the master *when* to lean on proof vs. cues.
- **George Loewenstein** — the **information-gap theory of curiosity** (1994). Curiosity is the felt deprivation of a known gap in knowledge; the science under every curiosity-gap headline and open loop.

---

## 3. The architecture — the canonical structure of a best-in-class artifact

Two things are architecture here: (a) the **decision grid** that precedes any writing, and (b) the **component order** of the artifact itself.

### 3.1 The decision grid (Schwartz, operationalized) — computed FIRST, every time

**Axis 1 — Five Stages of Awareness** (what the prospect knows):

| Stage | Prospect's state | Lead type (Masterson/Forde) | Headline strategy |
|---|---|---|---|
| **Unaware** | Doesn't know they have the problem | Story / Proclamation | Indirect; dramatize a symptom or bold claim; no product |
| **Problem-Aware** | Feels the pain, doesn't know solutions exist | Problem-Solution / Big Secret | Name the pain vividly (PAS agitation) |
| **Solution-Aware** | Knows solution *types* exist, not your product | Problem-Solution / Promise | Position your mechanism as the best category answer |
| **Product-Aware** | Knows your product, not yet convinced | Promise / Offer | Lead with the differentiating benefit + proof |
| **Most-Aware** | Ready; needs the deal/nudge | Offer | Product + price + urgency + guarantee, up front |

**Rule: directness scales up with awareness.** Wrong-stage copy is the #1 conversion failure. Over-direct copy to an unaware market fails; over-indirect copy to a most-aware market wastes their readiness.

**Axis 2 — Five Stages of Market Sophistication** (how tired the market is):
1. **First to market** — name the claim plainly ("Lose weight"). The claim alone works.
2. **Second** — enlarge/outbid the claim ("Lose weight in 30 days").
3. **Third** — introduce a **mechanism** ("A pill that blocks fat absorption"). The *how* differentiates.
4. **Fourth** — elaborate/amplify the mechanism ("time-release, clinically dosed").
5. **Fifth** — market is exhausted and cynical; shift to **identity / experience / new emotional angle**, or reset with a genuinely new mechanism. Here **voice, proof, and story** carry the load.

**The axes multiply.** A *Most-Aware* prospect in a *Stage-5* market needs an Offer lead **and** an identity/voice angle. A *Problem-Aware* prospect in a *Stage-3* market needs Problem-Solution **and** a clearly-named mechanism. This multiplication is the master's **first computation** on any brief.

### 3.2 The component order of a sales page (the canonical spine)

Each component below is named with the practitioner whose logic justifies it.

1. **Headline block** — the ad's ad (John Caples, *Tested Advertising Methods*, 1932; Ogilvy's "80 cents of the dollar"). Carries a *declared lead type* (one of the six) and a *declared awareness stage*. Appeal order of reliability: self-interest → news → curiosity (Caples). Built as a **big idea** (Ogilvy) around a **USP** competitors cannot claim (Rosser Reeves, *Reality in Advertising*, 1961).
2. **Lede** — the first ~100–800 words that Masterson/Forde say determine ~80% of the result. Enters the existing conversation (Collier); opens a curiosity gap (Loewenstein); begins the slippery slide (Sugarman).
3. **Mechanism section** — the *named* "how it works." Required from Sophistication Stage 3 up. Bencivenga's "explain the mechanism" doubles as proof and as authority (Cialdini). Give the mechanism a **proper name** — an owned string the prospect can repeat.
4. **Proof block** — minimum three proof elements, each tagged `[testimonial | data | demo | authority | story]`. Bencivenga's checklist; Hopkins's reason-why; Heath's Credible. **Untagged claims that carry numbers fail the gate** — a number without a source is not proof.
5. **Offer section** — price, stack, guarantee text. Anchor high value before the low price (Kahneman/Tversky anchoring). The guarantee is itself a proof element (Bencivenga's "creative guarantee").
6. **Objection handling** — minimum three. Each objection is a System-2 doubt (Kahneman) pre-answered; "acknowledge disbelief" (Bencivenga); reframes via Blair Warren's "justify their failures."
7. **Single primary CTA** — one action, one target. Fogg's model: reduce friction, add a clear prompt. More than one *distinct* target splits intent and fails the gate.
8. **PS block** — the second-most-read element after the headline in long-form direct mail (direct-response convention, Halbert/Kennedy). Restate the core promise + deadline, or deliver one last proof.

### 3.3 The component order of a VSL

Timestamped beats, in order: **hook (first 15s, ≤40 words)** → **lead (tagged type)** → **big-idea / mechanism naming** → **proof segments (≥3, tagged)** → **transition-to-pitch marker** → **stack** → **price reveal with anchor** → **guarantee** → **CTA (single action, repetition allowed)**. Running underneath: an **open-loop inventory** — at least two loops opened early (Loewenstein/Zeigarnik) with declared close points. An unclosed loop is a failure, not a style choice.

---

## 4. The schools & styles — the best practitioners and companies, studied

Each: **who · thesis · when it wins · when it fails.** These are the named traditions the master draws from. All names are verified in the research dossier.

### The founding definition
- **John E. Kennedy** (Chicago, 1904) — *"Advertising is salesmanship in print."* The sentence the entire reason-why school is built on. Reframes the ad from decoration to a substitute for a live salesperson: every word must do a salesperson's job. **Wins:** always, as a discipline. **Fails:** never — but taken too literally it can excuse charmless copy (Bernbach is the corrective).

### The reason-why / scientific school
- **Claude C. Hopkins** — *Scientific Advertising* (1923), *My Life in Advertising* (1927). Thesis: advertising is measurable salesmanship; every claim needs a **reason why**; specificity is credibility. Originated the **preemptive claim** (state an ordinary process as if unique — Schlitz "steam-cleaned bottles," which every brewer did but none had claimed; took Schlitz from 5th to 1st). **Wins:** considered, provable purchases; categories where a real mechanism exists. **Fails:** low-involvement impulse buys where nobody reads reasons.
- **David Ogilvy** — *Confessions of an Advertising Man* (1963), *Ogilvy on Advertising* (1983). Thesis: "The consumer is not a moron; she is your wife." Research first; the **headline is 80 cents of the dollar**; the **big idea**; brand as long-term equity; factual long copy (Rolls-Royce 1958: "At 60 miles an hour the loudest noise comes from the electric clock," + 607 words of specifics). **Wins:** high-ticket, considered, brand-building; the 19Keys register. **Fails:** when speed and raw response beat prestige; thin offers that can't survive scrutiny.
- **Rosser Reeves** — *Reality in Advertising* (1961). The **USP**: a proposition (not a slogan) that is unique (competitors can't claim it) and sells (strong enough to move the mass). M&M's "Melts in your mouth, not in your hand." **Wins:** commodity differentiation, mass reach. **Fails:** parity products with no real unique claim (forces invention → dishonesty).

### The mechanics / testing school
- **Victor O. Schwab** — *How to Write a Good Advertisement* (1962). Five jobs in order: **attention → advantage → prove it → persuade to grasp it → ask for action.** **Wins:** as the default skeleton for any long ad. **Fails:** nothing — it's structural; but structure alone won't save a wrong idea.
- **John Caples** — *Tested Advertising Methods* (1932). Test everything; a winning headline can beat a loser 19-to-1 on identical body copy. Appeals ranked: self-interest, news, curiosity, quick-easy-way. The canonical story lead: *"They Laughed When I Sat Down at the Piano — But When I Started to Play!"* (1926, U.S. School of Music). **Wins:** anywhere you can split-test. **Fails:** as an excuse to skip judgment when volume is too low to test.

### The desire / market-selection school
- **Eugene M. Schwartz** — *Breakthrough Advertising* (1966). The keystone (§2.1, §3.1). Channel desire; match awareness × sophistication. **Wins:** everywhere; it's the operating system. **Fails:** only when misapplied — mis-locating the prospect on the grid.
- **Gary Halbert** — *The Boron Letters*; *The Gary Halbert Letter*. The **"starving crowd"**: given a hamburger-stand contest, the one advantage to want is a starving crowd — market selection beats copy. The **A-pile** letter (mail people actually open); write to one person; greased-slide readability. The "Coat of Arms" letter mailed reportedly hundreds of millions of times. **Wins:** direct mail, cold traffic, when list/market choice is in play. **Fails:** when you can't choose the market (fixed audience) — then copy must do more.

### The momentum / structure school
- **Joseph Sugarman** — *The Adweek Copywriting Handbook* (2007), from JS&A / BluBlocker. The **Slippery Slide**: "the sole purpose of the first sentence is to get you to read the second." Seeds of curiosity ("But there's more…"); 31 psychological triggers; sell the concept, not the product; "sell the cure, not the prevention." **Wins:** long-form where momentum is everything. **Fails:** scannable, skim-first surfaces (comparison pages, most-aware buyers who want the deal now).
- **Dan Kennedy** — *The Ultimate Sales Letter*. The long-form sales letter as highest-leverage asset; **PAS — Problem · Agitate · Solve** (agitate before relieving); Message-to-Market-to-Media match (in that order). **Wins:** info-products, coaching, local-biz direct response. **Fails:** sophisticated buyers allergic to overt agitation; luxury registers where agitation reads as cheap.

### The proof / belief school
- **Gary Bencivenga** — *Bencivenga Bullets*; *Bencivenga 100*. Proof is the great persuader; the Persuasion Equation and proof-elements checklist (§2.9). Decades of split-test-winning financial and info controls built on stacked, specific, demonstrable proof. **Wins:** skeptical, high-stakes, high-ticket markets (finance, health, anything "too good to believe"). **Fails:** never on principle — but proof-stacking without a compelling idea is a well-defended boring ad.

### The creative / brand school (the counterweight)
- **Bill Bernbach** (DDB) — no single book; the legacy is the work and the memos. "The truth isn't the truth until people believe you… and they won't listen to you if you're not interesting." Volkswagen **"Think Small"** and **"Lemon"** (1959–60): anti-hype, self-aware, factual; turned a small ugly car into an icon by respecting the reader. **Wins:** brand-building, category disruption, sophisticated audiences, the 19Keys restraint register. **Fails:** when immediate measurable response is the only KPI and there's no room for wit.

### Named companies / houses as schools
- **Agora / the "Big Idea" VSL school** (associated with **Mark Ford**, aka Michael Masterson, at Agora) — one emotionally-charged, intellectually-interesting **Big Idea** carries a 20–40 minute VSL from lead to close. **Wins:** cold-to-warm long-form info-product sales at scale. **Fails:** short attention surfaces; audiences that punish overt hype.
- **Copyhackers** (**Joanna Wiebe**) — **Voice-of-Customer (VoC) mining**: the copy is *assembled from the customer's own words* (reviews, tickets, interviews), not invented. **Wins:** SaaS, considered B2C/B2B, anywhere real customer language exists. **Fails:** brand-new categories with no existing customers to mine.
- **Ogilvy & Mather / Hopkins-lineage agencies** — research-dense factual long copy as house method (see Ogilvy, Hopkins above).

*(Mine-but-not-canon, per dossier: Maxwell Sackheim ("Do You Make These Mistakes in English?"), Clayton Makepeace, Drayton Bird, John Carlton, Ann Handley.)*

---

## 5. Decision frameworks — choosing the approach for a given situation

The master runs these checks in order. Inputs come from `funnel-psychology`'s brief.

### 5.1 The primary computation — locate on the grid
Read **awareness stage** and **sophistication stage** off the brief. If either is missing, **stop and request it**. Then:
- **Awareness → lead type** (the Masterson/Forde table, §3.1). This is deterministic, not a preference.
- **Sophistication → mechanism weight.** Stage 1–2: claim-forward, mechanism light. Stage 3–4: mechanism *named and central*. Stage 5: identity/voice/story forward, or a genuinely new mechanism.

### 5.2 Approach selection by variable

| Variable | If… | Then lean… |
|---|---|---|
| **Awareness = Unaware** | prospect doesn't know the problem | Story lead (Caples/Ford) or Proclamation; PASTOR for values-led audiences |
| **Awareness = Problem-Aware** | pain is felt, no solution known | PAS (Kennedy) or Big-Secret lead |
| **Awareness = Solution-Aware** | comparing categories | Problem-Solution + named mechanism (RMBC's mechanism, §6) |
| **Awareness = Product/Most-Aware** | knows you / ready | Promise or Offer lead; Four-Ps for mid-length |
| **Sophistication ≥ 3** | market has heard the claims | RMBC (Georgi) — mechanism is the differentiator |
| **Sophistication = 5** | market is cynical | Bernbach voice + Bencivenga proof + identity (Blair Warren) |
| **Price point high / skeptical** | high-ticket, "unbelievable" | Ogilvy factual long-form + Bencivenga proof-stack |
| **Channel = cold ad** | 3–12 words to earn the click | AIDA (St. Elmo Lewis) skeleton; hook-first |
| **Channel = long VSL** | 20–40 min | Big-Idea VSL (Agora/Ford) with open-loop inventory |
| **Channel = email body** | mid-length | Four-Ps (Hoke) or PAS |
| **Audience = 19Keys / values-led** | sovereignty, identity, culture | PASTOR (Edwards) — narrative proof + transformation, in house voice |
| **Real customer language exists** | reviews/tickets/DMs available | VoC mining (Wiebe) *before* drafting, regardless of other choices |

### 5.3 The ELM check (Petty & Cacioppo)
Is the reader motivated and able to think hard about this (high elaboration)? → Central route: **lead on argument and proof** (Ogilvy/Bencivenga). Low elaboration (distracted, impulse, cold)? → Peripheral route: **lead on cues** (authority, social proof, liking, a sharp hook). High-ticket = almost always central; cold social ad = often peripheral to earn the click, central on the page that follows.

### 5.4 The voice check (non-negotiable for 19Keys)
Before drafting a single 19Keys-brand line (ZIION, Supermind, CWK, HIGH LVL), load the **brand-bible** skill and enforce: house register, **forbidden vocabulary**, preferred alternatives, banned AI-cliché fonts, black-and-gold discipline, "sell the outcome not the access." The master is the keeper of voice on the surface; funnel-psychology sets the skeleton, the master sets the register.

---

## 6. The multiple ways to run it — the playbooks

Ten repeatable builds. Each maps to a file under `playbooks/`. Each entry: origin, when to reach for it, the shape.

1. **`pas` — Problem-Agitate-Solve** — *Dan Kennedy* (popularized). Name the pain → twist it until vivid → present the mechanism of relief. Loss-salience (Kahneman) is the engine of the agitate step. **Reach for it:** Problem-Aware prospects; info-products; direct response. **Shape:** 3 beats, agitation ~2–3× the length of the problem statement, solution named as a mechanism.

2. **`aida` — Attention-Interest-Desire-Action** — *Elias St. Elmo Lewis* (c. 1898). The default skeleton for short-form ads and cold surfaces. **Reach for it:** cold ads, banners, any surface where you have seconds. **Shape:** four movements, front-loaded; attention earned in the first line.

3. **`slippery-slide`** — *Joseph Sugarman*. Every element exists only to get the next line read; open with curiosity seeds, close on momentum. **Reach for it:** long-form sales letters and VSLs where flow beats scannability. **Shape:** short first sentence; seeds of curiosity between sections; no dead ends.

4. **`rmbc` — Research-Mechanism-Brief-Copy** — *Stefan Georgi*. 60% research, a **unique named mechanism**, a one-page brief, THEN the draft. **Reach for it:** Sophistication ≥3 where the *how* differentiates; high-ticket. **Shape:** research doc → named mechanism string → 1-page brief (promise, mechanism, proof, offer) → copy.

5. **`great-leads-selection`** — *Masterson (Mark Ford) & Forde, Great Leads* (2011). Choose one of **six lead types** strictly by awareness stage: **Offer, Promise, Problem-Solution, Big-Secret, Proclamation, Story** (most direct → most indirect). **Reach for it:** always, at the top of every job — this selects how the piece opens. **Shape:** declare the lead type as a tag; write the first ~100–800 words to that type.

6. **`voc-mining`** — *Joanna Wiebe / Copyhackers*. Mine reviews, support tickets, DMs, and interviews for the customer's own words; the copy is **assembled, not invented**. **Reach for it:** whenever real customer language exists (for 19Keys: Sovereign CRM notes, IG/ManyChat DMs, testimonials). **Shape:** swipe file of verbatim phrases → cluster by desire/objection → headlines and bullets built from the strongest verbatims.

7. **`big-idea-vsl`** — *Agora school / Mark Ford*. One emotionally-charged, intellectually-interesting **Big Idea** carries a 20–40 minute VSL from lead to close. **Reach for it:** cold-to-warm long-form video sales. **Shape:** big idea stated in the hook; open-loop inventory ≥2; mechanism named; proof stacked; single repeated CTA.

8. **`ogilvy-factual-long-form`** — *David Ogilvy / Claude Hopkins, reason-why*. Research-dense, specific, long copy that sells with facts and proof. **Reach for it:** considered, high-ticket purchases; skeptical buyers; the 19Keys institutional register. **Shape:** fact-built headline → specifics → reason-why → proof → offer; respect the reader's intelligence.

9. **`four-ps` — Promise-Picture-Proof-Push** — *Henry Hoke*. Make the claim → make it vivid → back it → close. **Reach for it:** mid-length email and landing copy. **Shape:** four beats, tight; Picture is sensory/concrete (Heath), Proof is tagged, Push is a single CTA.

10. **`pastor` — Problem-Amplify-Story-Transformation-Offer-Response** — *Ray Edwards*. PAS extended with narrative proof; suited to values-led audiences. **Reach for it:** 19Keys' sovereignty/identity audience; faith/values-led; story-carried offers. **Shape:** six beats; Story carries the proof; Transformation is the identity shift (Blair Warren: encourage their dreams); Response is the single CTA.

---

## 7. Metrics & instrumentation — what good looks like

The master ships copy; the funnel and email systems measure it. But the master owns leading indicators and self-scored gates *before* anything goes live.

### 7.1 Pre-flight (self-scored, on every artifact — no live traffic needed)
- **Flesch Reading Ease** computed; **≥50 to pass**, target **60–80** for direct response (Flesch). Below 50 fails.
- **Banned-vocabulary / hype scan** = 0 hits (brand-bible forbidden list + hype list: "guaranteed riches," "effortless," "overnight," etc.).
- **Proof coverage** = every numeric claim carries a tagged proof element; untagged number = fail.
- **Lead-type + awareness tags present** and internally consistent with the brief.
- **CTA count** = exactly one distinct target on a sales page / hero.
- **Paragraph length** ≤90 words; **em-dash density** ≤1 per 200 words (house style).

### 7.2 Live benchmarks (directional, market-dependent — never promise these; measure them)
- **Cold sales-page conversion:** low single digits is normal; **1–3%** is a common working band for cold traffic to a paid offer; warm/most-aware pages run higher. Treat any inherited "typical" number as a hypothesis to beat, not a guarantee.
- **VSL:** watch-time retention curve (where do viewers drop?), % reaching the price reveal, order-page CTR.
- **Ad set:** hook-rate (3-sec view / impressions) and CTR as the *copy* signal; CPA/ROAS as the *offer+funnel* signal (shared with funnel-psychology).
- **Email subject bank:** open rate as the subject-line signal (deliverability caveats owned by email-marketing); reply/click as body signal.
- **Landing hero:** above-the-fold engagement, scroll-past-hero rate, primary-CTA click rate.

### 7.3 The testing discipline (Caples)
One variable at a time where volume allows; headline first (it moves the most — Caples' 19-to-1). Below testable volume, fall back to judgment via the grid, and log the bet so it can be checked later. **No cherry-picked numbers** (house rule): never present a single winning example beside a corpus mean without labeling which is which.

---

## 8. Failure modes & the "no bullshit" bar — the verify-gate rubric

The verify gate is a **linter**, not a vibe. Below are the machine-checkable rubrics per deliverable, followed by the universal bans and the ways practitioners fake competence.

### 8.1 `sales-page` rubric
Markdown page must contain, **in order**: headline block (with declared **lead-type tag** from the six Great Leads + declared **awareness stage**) → lede → **mechanism section** (named mechanism string present) → **proof block** (≥3 proof elements, each tagged `[testimonial|data|demo|authority|story]`; **untagged claims carrying numbers fail**) → offer section (price + stack + guarantee text present) → objection handling (**≥3**) → **single primary CTA** (linter counts CTA markers; **>1 distinct target fails; 0 fails**) → PS block. **Hard fails:** brand-bible forbidden vocabulary; hype list (guaranteed riches / effortless / overnight); any paragraph **>90 words**; em-dash overuse (**>1 per 200 words**); **Flesch Reading Ease <50**.

### 8.2 `vsl-script` rubric
Timestamped beats required: **hook (first-15s block present, ≤40 words)** → lead (**tagged type**) → big-idea/mechanism naming → **proof segments ≥3 (tagged)** → transition-to-pitch marker → stack → price reveal **with anchor** → guarantee → CTA (single action; repetition allowed). Plus **open-loop inventory ≥2** with **declared close points**. **Fails on:** any missing beat; untagged proof; first-15s block over the word cap; **any unclosed loop**.

### 8.3 `ad-set` rubric
File must contain **≥5 variants**. Each declares: **awareness-stage target**, **hook line (≤12 words)**, primary text, **single CTA**, and an **angle tag** (each angle used **at most twice** across the set — diversity enforced). **Fails on:** <5 variants; **duplicate hook lines**; any variant missing the awareness tag; platform-policy red-flag words (before/after income claims without a disclaimer line).

### 8.4 `email-subject-bank` rubric
**≥20 subjects**, grouped by **named curiosity mechanism** (open loop, specificity, contrast, self-interest, story-cue — **≥4 mechanisms represented**). Each subject **≤9 words**, **no ALL-CAPS words**, **≤1 emoji**, no spam-trigger hits (`free!!!`, `act now`, `$$$`). Each group **cites its mechanism**. **Fails on:** count, length, mechanism coverage, or spam-word hits.

### 8.5 `landing-hero-variants` rubric
**≥3 hero variants**. Each: headline (**≤14 words**), subhead (**≤25 words**), CTA label (**≤4 words, verb-first**), declared **lead type + awareness stage**, and a one-line **"because" rationale citing a MANUAL.md school by name**. **Fails on:** missing rationale; headline over cap; or **all variants sharing the same lead type** (diversity required).

### 8.6 The universal "no bullshit" bar
- **Every claim with a number carries a tagged source.** This is the Bencivenga/Hopkins line and the single hardest gate.
- **Awareness/lead-type declared and consistent.** Undeclared = the writer skipped the primary computation.
- **One idea (Rule of One, Zajonc).** If the piece has three "big ideas," it has none.
- **Reads easily (Flesch ≥50).** Fluency is truth-signal; disfluency reads as risk.
- **In-brand voice.** Forbidden vocabulary = automatic fail on 19Keys surfaces.

### 8.7 How practitioners fake competence (watch for these)
- **Hype substituting for proof** — superlatives ("revolutionary," "insane results") where a tagged demonstration belongs. Reeves' USP requires a *provable* proposition; hype is what you write when you have none.
- **Adjectives instead of specifics** — "fast, powerful, amazing" instead of Ogilvy's electric-clock specificity. Specifics are credibility; adjectives are noise.
- **Wrong-stage directness** — pitching price to an Unaware market, or telling a Most-Aware buyer a 20-minute story. The #1 failure (Schwartz/Ford).
- **Manufactured desire** — trying to create a want that isn't there instead of channeling one that is (violates Schwartz's keystone).
- **Multiple CTAs / multiple ideas** — hedging by offering everything; converts nothing.
- **Borrowed swipe with no VoC** — copying a competitor's structure without mining the actual customer's words (Wiebe). Feels generic because it is.
- **Open loops never closed** — curiosity bait with no payoff; erodes trust and fails the VSL gate.
- **Readability theater** — short sentences that say nothing vs. genuinely fluent argument. Flesch passes; persuasion fails. The human/verify layer catches this where the linter can't.

---

## 9. Bridge to the 19Keys stack

How this master plugs into the real assets. Concrete, not hand-wavy.

### 9.1 Inputs — where a brief comes from
- **`funnel-psychology`** hands down the **stage-and-awareness spec** per surface (which stage of the ZIION value ladder, which awareness stage, which sophistication stage, what offer/price/guarantee). The copywriter's **first computation** (§3.1) consumes exactly this. If the spec lacks awareness/sophistication, the master **requests it** rather than inventing structure.
- **Sovereign CRM** (Supabase project **`YOUR_SUPABASE_PROJECT_REF`**) is the **VoC mine** for the `voc-mining` playbook: pull real buyer language from contact notes, testimonials, and message history; assemble headlines and bullets from verbatims. The CRM spans the $611K / ~4,591-buyer book across platforms — that is the largest source of the customer's own words in the stack. Use `supabase-manager` / the Supabase MCP (auth required in an interactive session) to read; the copywriter **reads**, it does not mutate CRM records.

### 9.2 Voice — the brand gate
- Every 19Keys-brand surface (**ZIION**, **Supermind**, **CWK/Cognitive Wealth Key**, **HIGH LVL**) is written through the **`brand-bible`** skill: house register, forbidden vocabulary, preferred alternatives, banned AI-cliché fonts (Instrument, Space Grotesk, Geist), black-and-gold discipline, "sell the outcome (Builder Passport) not the access." ZIION type is locked (Marcellus + Hanken Grotesk + Space Mono). This is enforced in §8's banned-vocabulary gate, not left to taste.

### 9.3 Outputs — who receives finished copy
- **The funnel pages** (ZIION funnels, funnel-studio / public-funnel runtimes) receive **sales-page**, **landing-hero-variants**, and **VSL scripts**. The master fills the words into the structure funnel-psychology defined; it does not restructure the funnel.
- **`email-marketing`** receives the **email-subject-bank** and any email-body copy. The master writes the *words*; email-marketing owns **cadence, segmentation, and sending** via the **email engine (Resend + `outreach_worker`)**. The master never sets send timing or list logic. (Real state per the stack: CRM has ~4,591 buyers and $611K; sending is gated on `OUTREACH_POSTAL_ADDR` for CAN-SPAM — a compliance blocker the master flags but does not resolve.)
- **The execution-tier agents** consume the copy as ledger-driven artifacts:
  - **`funnel-closer`** — takes finished sales-page / hero / VSL copy and wires it into the live close path.
  - **`outreach-operator`** — takes the subject bank + short-form copy for cold/warm outreach (once the postal-address blocker clears).
  - **`lead-pipeline` / CRM work** — supplies the master its raw material (segments, hot lists, buyer verbatims) and receives back the copy that moves those leads.

### 9.4 The clean contract (one line)
`funnel-psychology` → **structure in** → **master-copywriter writes the words on one surface, passes the §8 gate** → **copy out** → `email-marketing` (cadence) + funnel pages (placement) + `funnel-closer` / `outreach-operator` / `lead-pipeline` (execution). The master owns words and voice; it borrows structure and hands off distribution.

---

## Appendix A — the operating stack, in order (the checklist)
1. **Locate** the prospect on Awareness × Sophistication (Schwartz). *Input from funnel-psychology.*
2. **Enter** the conversation already in their head (Collier; Cialdini Pre-Suasion). Channel desire, don't manufacture it (Schwartz).
3. **Select** the lead type keyed to awareness (Masterson/Forde's six).
4. **Write the headline** for self-interest / news / curiosity (Caples), as a big idea (Ogilvy), around a USP competitors can't claim (Reeves).
5. **Build the body** on a slippery slide (Sugarman), or Attention→Advantage→Proof→Persuade→Ask (Schwab), or PAS (Kennedy) — per §6.
6. **Attach proof to every claim** (Bencivenga's equation + checklist; Hopkins's reason-why). *Non-negotiable.*
7. **Layer influence** (Cialdini) and identity moves (Blair Warren; Heath's Emotional).
8. **Frame the offer** with anchoring + loss aversion (Kahneman/Tversky).
9. **Repeat the core claim** across the piece (Zajonc; Rule of One).
10. **Score it** — Flesch ≥50, banned-vocab 0, proof-tagged, one CTA, one idea (§8). Then ship.

## Appendix B — source ledger (every name, one line)
Robert Collier — conversation-in-the-mind (*Letter Book*, 1931) · John E. Kennedy — "salesmanship in print" (1904) · Claude Hopkins — reason-why, preemptive claim (*Scientific Advertising*, 1923) · David Ogilvy — big idea, factual long copy (*Ogilvy on Advertising*, 1983) · Rosser Reeves — USP (*Reality in Advertising*, 1961) · Victor Schwab — 5-step ad (*How to Write a Good Advertisement*, 1962) · John Caples — tested headlines (*Tested Advertising Methods*, 1932) · Eugene Schwartz — awareness × sophistication (*Breakthrough Advertising*, 1966) · Gary Halbert — starving crowd, A-pile (*The Boron Letters*) · Joseph Sugarman — slippery slide (*Adweek Copywriting Handbook*, 2007) · Gary Bencivenga — persuasion equation, proof elements (*Bencivenga Bullets*) · Dan Kennedy — PAS, sales letter (*The Ultimate Sales Letter*) · Bill Bernbach / DDB — creative revolution, VW "Think Small"/"Lemon" (1959–60) · Michael Masterson (Mark Ford) & John Forde — six lead types (*Great Leads*, 2011) · Robert Cialdini — influence principles (*Influence*, 1984; *Pre-Suasion*, 2016) · Kahneman & Tversky — prospect theory (1979; *Thinking, Fast and Slow*, 2011) · Chip & Dan Heath — SUCCESs (*Made to Stick*, 2007) · Robert Zajonc — mere exposure (1968) · Alter & Oppenheimer — processing fluency (2009); McGlone & Tofighbakhsh — rhyme-as-reason (2000) · Rudolf Flesch (with J. Peter Kincaid) — readability (1946–49) · Blair Warren — 27 words (2004) · BJ Fogg — B=MAP · Petty & Cacioppo — Elaboration Likelihood Model · George Loewenstein — information-gap curiosity (1994) · Stefan Georgi — RMBC · Joanna Wiebe / Copyhackers — VoC mining · Ray Edwards — PASTOR · Henry Hoke — Four Ps · Elias St. Elmo Lewis — AIDA (c. 1898) · Agora / Mark Ford — Big-Idea VSL.