Messaging and Copywriting

The gap between a product people want and a product people buy is usually words. Say what it is, for whom, in the language they already use, not the language you invented. Your best copy is a sentence you overheard in a user interview.

The concept

Messaging is the promise; copywriting is how you phrase it everywhere it appears, headline, App Store description, onboarding, buttons, emails, error states. Most products under-invest here and then blame the funnel.

The single most important idea: write in your customer's words, not your own. You've been staring at your product for months; you describe it in terms of its features and your internal names. Your customer describes it in terms of their problem, in plain language. The closer your copy is to how they talk, the less translation they have to do, and translation is friction.

That's why the market-research and interview chapters (3, 4) insisted on verbatim quotes. Those quotes are your copy.

📐 Your best copy is mined, not written. The words that convert are the ones your customer already used to describe their problem, pull them from interviews and reviews rather than inventing your own. Every layer of translation between their language and yours is friction you added.

The message hierarchy

 POSITIONING     who it's for, what category, against what (Chapter 6)
     ▼
 CORE MESSAGE    one sentence: the promise
     ▼
 HEADLINE        the promise, made vivid, above the fold
     ▼
 SUPPORTING      the proof, the how, the objections handled
     ▼
 MICROCOPY       buttons, labels, empty states, errors — the daily texture

Everything flows down from positioning. If your headline doesn't fall out of your positioning statement easily, the positioning isn't sharp yet.

Benefit, not feature

The classic rule, still true: sell the outcome, not the mechanism.

Lead with the benefit; back it with the feature as proof. People buy the outcome and reassure themselves with the mechanism.

Show, don't tell: even in copy

Instead of telling people to feel something or do something, show them what's true and let them conclude it. This is more persuasive than instruction because the conclusion becomes theirs:

• A directive ("do this daily") creates an obligation the reader can fail.
• A vague suggestion ("you might consider…") carries no information.
• Evidence ("people who did X achieved Y") cannot be failed and cannot be argued with, the reader draws their own conclusion, which is why it sticks.

This applies far beyond marketing: onboarding, empty states, upsells, and error messages all persuade better with evidence than with commands.

Voice, and what you refuse to say

Define your voice in three physical adjectives, "warm, direct, unhurried", not abstractions like "professional." Then write down the words you never use (Chapter 36: some are review-risk; others are just off-brand). A forbidden-words list keeps copy consistent across every surface and every contributor.

Microcopy is the product's manners

Buttons, labels, empty states, and error messages are copy too, and they're read more often than your landing page. "Something went wrong" is worse than "We couldn't save that, check your connection and try again." Every error should say what happened and what to do next (Chapter 11).

The AI-writing tell

AI drafts copy fluently and generically, the same em-dashes, the same "in today's fast-paced world," the same rule-of-three cadence, the same inflated adjectives. Readers increasingly recognise it, and it reads as nobody wrote this. Use AI to draft, then cut it to something a specific human would actually say. Your customer's verbatim language is the antidote.

📐 Best practice

Write in your customer's words, mine your interviews and reviews for verbatim phrasing.

Derive the headline from the positioning.

Lead with the benefit, back it with the feature.

Show evidence instead of giving instructions.

Define voice in physical adjectives and keep a forbidden-words list.

Treat microcopy as product copy, every error says what and what-next.

One core message, consistent across every surface.

Cut AI fluency to human specificity.

Test copy on strangers, five-second comprehension (Chapter 6).

Read it aloud. If you wouldn't say it, don't ship it.

💀 Common mistakes

Your words, not theirs. Feature-speak and internal names the customer has to translate.

Feature-first copy. The mechanism before the outcome.

Directive copy that creates an obligation the reader can fail.

Vague "elevate your workflow" filler that says nothing.

Inconsistent voice across surfaces.

Neglected microcopy, raw error codes, "something went wrong."

Clever over clear. Wit that costs comprehension.

Generic AI copy shipped unedited.

Never testing copy on anyone who isn't you.

Copy that overpromises what the product delivers, the fastest route to churn and bad reviews.

The professional workflow

 1. START FROM POSITIONING (Chapter 6)

 2. MINE VERBATIM LANGUAGE from interviews, reviews, support (Chapters 3-4)

 3. WRITE THE CORE MESSAGE — one sentence, the promise

 4. DERIVE THE HEADLINE — vivid, benefit-led

 5. SUPPORTING COPY — proof, how, objections; evidence over instruction

 6. VOICE — three physical adjectives + a forbidden-words list

 7. MICROCOPY — buttons, empty states, errors (what + what-next)

 8. CUT AI FLUENCY to human specificity; read aloud

 9. TEST on 5 strangers (5-second comprehension)

10. KEEP IT CONSISTENT across every surface

Tools, websites & costs

Copywriting costs almost nothing, it's a thinking-and-editing skill. The highest-value input is your interview transcripts.

Alternatives & trade-offs

Clarity vs cleverness. Clear always beats clever when they conflict, but a distinctive voice is a real asset once clarity is secured. Earn the wit; don't lead with it.

Emotional vs rational appeals. Consumer usually converts on emotion (identity, aspiration, relief) with rational proof as reassurance. B2B converts on rational outcomes (time saved, revenue) with emotional undertones (looking good, reducing risk). Match to your buyer.

Short vs long copy. Short suits obvious value and impulse decisions; long suits considered purchases where objections must be handled. Match length to how much convincing the decision needs.

In-house vs hired copywriter. Early, you write it, you know the customer's words and the product's truth. A professional copywriter helps later for polish and conversion optimisation, and can't replace your proximity to the customer at the start.

Write-then-test vs test-then-write. With traffic, A/B testing headlines beats arguing about them. Without traffic, five-second comprehension tests are the cheap substitute.

Checklist

• Copy uses my customer's verbatim language
• The headline derives from positioning
• Copy leads with benefit, backs it with feature
• I show evidence rather than giving instructions
• Voice is defined in physical adjectives
• A forbidden-words list exists
• Microcopy says what happened and what to do next
• The core message is consistent across every surface
• AI drafts are cut to human specificity
• Copy tested on 5 strangers
• Nothing overpromises what the product delivers

📓 Case Study: the complaint that became the copy

Project: SOLIS. Two messaging decisions worth pulling out, one that generalises well, one that shows copy discovered through iteration.

⭐ The core message came directly from a customer's verbatim complaint. Market research surfaced the category's universal grievance, in real users' words:

"Most Stoic apps show you a quote, ask you to journal, and charge $50/year… they skip the hard part… more flavor than substance."

That single overheard sentence contained the entire message. It told the product what to be (substance, structure, actual work, not another quote app) and how to say it. The positioning that fell out, "the only structured, gamified course, with visible proof of change", is a direct answer to a complaint written by the target customer. The best copy wasn't written; it was mined.

The taglines followed the benefit-not-feature rule: "Build the mind of an emperor" and "Master the beast within" sell an outcome (who you become), not the mechanism (a 90-day curriculum).

The reveal-card tone war, copy found through three rejected versions. Each lesson ended with a closing line, and it took three iterations to find the right register, a good illustration of "show, don't tell":

1. Directive, "Take a twenty-minute walk today without your phone." Rejected: a directive creates a pass/fail moment. "People might be lazy… people might not like being told." An instruction the reader can fail is the same deficit the whole product was designed to avoid.
2. Gentle suggestion, softened to "you might consider…" Rejected: "it's more Seneca but now it's not helping the user." In removing the pressure, it removed the information, atmospheric and empty.
3. Sourced fact, "nearly half of self-made millionaires woke three hours before their workday." Kept.

The winning form is the chapter's principle exactly: a fact cannot be failed and cannot be argued with. It doesn't tell the reader to wake early; it shows what people who got what they want actually did, and lets them draw their own conclusion, which is why it lands.

A forbidden-words list existed, for a real reason. Because the product touched a sensitive domain, the messaging kit listed words to avoid (review-risk medical-claim language) and words to use instead (Chapter 36), a voice guardrail that kept copy consistent and out of trouble.

🚩 The critical limitation: none of this copy was tested on a real audience. No five-second comprehension test, no A/B test, no landing-page experiment. The headline, the taglines, the tone, all well-reasoned, drawn from genuine customer language, and entirely unvalidated. The complaint-to-copy pipeline is a strong method; whether this execution converts is unknown, because it was never put in front of anyone with an intent to act.

Lessons

1. Write in your customer's words, not yours. The best copy is mined from interviews and reviews, not invented.
2. The universal complaint is often the whole message. Answer it directly.
3. Lead with the benefit (who they become), back it with the feature.
4. Show evidence, don't give instructions. A fact can't be failed or argued with; a directive can be both.
5. Copy is often found through rejected versions. Directive → suggestion → evidence was three iterations to the right register.
6. Define voice in physical adjectives; keep a forbidden-words list.
7. Microcopy is read more than your landing page. Every error: what happened, what to do next.
8. Cut AI fluency to human specificity.
9. Test copy on strangers. Reasoned-from-real-language is still a hypothesis until someone with intent reacts to it.

Next: Chapter 44: Pre-launch: Audience and Waitlists →