Write conversion copy on direct-response fundamentals

01 Diagnose the layer before you touch a word. Copy is not positioning.

Positioning, messaging, and copy sit at three altitudes with three different decision rights. Positioning decides which mental box the product occupies. Messaging is the grayscale substance: capability, primary benefit, primary use case, differentiation. Copy is messaging dressed in brand voice and channel context. Jumping from a vague positioning idea straight to generated headlines produces lines that sound interesting and carry no truth. Positioning, messaging, and copy are three distinct layers, collapsing them produces hollow output

"Positioning ≠ Messaging ≠ Copy. ... Without a grounded message, I was just picking copy that sounded 'interesting' but lacked substance or truth."

· Aatir Abdul Rauf, LinkedIn, 2026-04-10

So the first move is to check which layer is actually broken. If reps say different things about the same product on the same day, the messaging is missing and no amount of copy polish will fix it. If the page reads clearly but the offer does not land, the problem is upstream of copy entirely. Market choice beats offer strength beats persuasion. Market and offer beat funnel optimisation Do not optimize a headline for a starving-crowd problem.

02 Mine customer language first. The phrase is the hook; the synthesis is the description.

Before writing, collect the words buyers already use. Reviews, support tickets, sales-call transcripts, Reddit threads, onboarding chats. Momoko Price works in reverse from most teams: collect the language, find the recurring emotional themes, build the copy upward from verbatim fragments. The copy sounds native to the audience because it literally came from them. Customer language first, positioning built upward from it, not the reverse

"Your customers' words should become your copy. Messaging-first optimization combines analytics rigor with emotional authenticity."

· Momoko Price (synthesized from operator's published work)

Adrienne Barnes calls the highest-value extraction the relational keyword set: the exact words customers use for the product, which should overwrite internal jargon in every line of copy, every SEO term, every sales script. Internal teams converge on "platform" and "solution." Buyers say "the thing that stops me from getting yelled at on Monday morning." Substitute the buyer's vocabulary and relevance jumps because the buyer recognizes their own language. The exact words customers use should replace internal terminology in every piece of copy The frontline conversation is the substrate; the desk is downstream. Frontline customer contact is the PMM substrate

The trap with AI here is specific. Deep-research synthesis smooths verbatim phrases into clean themes and strips out the exact words that make a buyer feel seen. Mine the sources directly and keep the phrases whole. AI synthesis smooths over verbatim buyer phrases: mine Reddit and review sites directly, mirror the words, do not paraphrase

"Resonance comes from mirroring, not paraphrase."

· Nicolas Cole, Build and Launch a VOC Landing Page, 2026-05-07

Collect 50 to 100 phrases before writing a single headline. The phrases that repeat across multiple buyers, unprompted, are your headline candidates.

03 Diagnose awareness level. Match the opening to where the reader is, not to the product.

Eugene Schwartz's rule still governs: copy cannot create desire, only channel desire that already exists. The job is to find the strongest existing desire the product connects to and write to that, not to the feature list. Awareness level decides how directly you can name the product. The further from awareness, the more indirect the open.

Diagnose by channel. A cold social ad reaches unaware buyers. A retargeting ad reaches product-aware or most-aware buyers. A comparison page reaches product-aware buyers actively evaluating. The same product needs five different openings for five audiences. A most-aware headline means nothing to an unaware reader, and an unaware open bores a most-aware buyer who wants the price.

04 Pick one pole. Toward pleasure or away from pain, never both in the same piece.

Every buying decision reduces to a single axis: the reader is moving toward pleasure or away from pain. Copy that activates neither pole is filtered as noise. Every buying decision reduces to one polarity, moving toward pleasure or away from pain. Copy that activates neither doesn't convert.

"People buy for exactly one reason: to move closer to pleasure or further from pain. Everything else is noise."

· Cole Schafer, Sticky Notes — Pain-Pleasure Polarity, 2024-01-15

Diagnosing the pole is a research job, not a guess. Read the interviews and reviews from step 2. Relief language ("finally," "no more," "we can stop") signals a pain buyer. Aspiration language ("imagine," "the kind of team that," "now we can") signals a pleasure buyer. Write to the language you find. Most B2B copy lives on the pain pole because the thing the buyer is trying to escape carries more emotional load than the thing they are moving toward. The failure is oscillating. A headline that mixes "escape the chaos" with "find new growth" delivers neither. Commit to one.

05 Spend a fifth of your time on the headline. Make it a complete argument, not a hook.

In a scanning context the headline is often the only element the reader sees. So it cannot be a hook that leans on the body to land the argument. It has to complete the persuasion alone, for the 95% who never scroll. Body copy amplifies a case the headline already made. Every headline must function as a complete persuasive argument, in the age of infinite scroll, the headline is often the only element a reader sees

"Every headline must function as a complete persuasive argument in itself."

· Eddie Shleyner, VeryGoodCopy — Headline as Complete Argument, 2024-02-01

The headline carries most of the weight, so it gets most of the deliberate effort. Schafer's discipline is to block separate time for it. Spend 20% of total writing time on the headline alone, it carries 80% of the persuasive weight

"Spend 20% of your total writing time on the headline alone. If a page takes five hours to write, the headline gets a full hour of that time."

· Cole Schafer, Honeycopy — the 20% Rule for headlines, 2026-03-03

Clarity wins over cleverness every time. A headline that asks the reader to interpret has already lost, because in a 1-to-2-second scan the reader allocates zero decoding time. The B2B instinct to sound sophisticated reads to the buyer as "I have to think about this," which is the same signal as "this is not for me." Clarity beats cleverness, always, a headline that requires interpretation is a headline that fails

"Clarity beats cleverness, always. A headline that requires interpretation is a headline that fails."

· Eddie Shleyner, VeryGoodCopy — Clarity Beats Cleverness, 2024-02-01

Then pick the archetype that matches the funnel stage, and do not mix archetypes in one line. Three headline archetypes, Flirting (curiosity), Direct (clarity), Pain-based (problem-recognition), pick the one that matches funnel stage

Write ten versions before choosing. The first is rarely the sharpest.

06 Put the reader in the sentence. Start with "you," not "we."

The most common B2B copy failure is company-side framing. "We are proud to announce." "Our award-winning team." The reader does not care how proud you are. People don't want to know how proud you are, they want to know how you'll change their life

"People don't want to know how proud you are of your company. They want to know how you'll change their life."

· Dave Harland, The Copy Cabin — Benefits Over Pride, 2024-03-01

The discipline is mechanical and consistently violated: audit every sentence that starts with "we" or "our" and ask whether it serves the reader or the company. "We support 14 languages" is fine because it implies "you can use this." "We are proud of our 200 customers" is not, because it serves the company. This is not anti-feature. Features earn their place by answering "what will the buyer be able to do that they cannot do now?" State the capability, then the benefit. Sell to the buyer's mindset, not to product features

07 Make every abstract claim concrete. Vivid images survive to the moment of decision.

Abstract claims do not survive the gap between exposure and purchase. Concrete images do, because the buyer recalls them at the moment they decide. "Save time" is abstract. "Stop spending Saturday morning rebuilding the spreadsheet" is concrete and gets recalled days later. Buyers who do not remember the claim do not act on it. Abstract claims disappear from memory; concrete images persist, vividness creates memorability

"Vividness creates memorability. Abstract claims disappear from memory; concrete images persist."

· Eddie Shleyner, VeryGoodCopy — Vividness Creates Memorability, 2024-02-01

This is also the engine of a good case study. Donald Miller's structure makes the customer the hero and the product the guide, and Joel Klettke's rule is to land the metric inside the emotional moment, not in a stat block. The customer is the hero. Your brand is the guide. If you confuse, you lose. Information in narrative form is 22x more memorable. Integrate the metric into the moment.

"Stories are 22x more memorable than facts alone. But the best conversion copy combines story with data."

· Joel Klettke / Case Study Buddy (synthesized from operator's published work)

The concrete image has to come from the buyer's actual life, drawn from the interviews in step 2. Invented detail that does not match the buyer's reality damages trust faster than a vague claim does.

08 Stack proof in layers and interleave it. One proof type is a claim; three is evidence.

Single-proof claims fail against the buyer's default skepticism. The Halbert sequence builds evidence in layers: state the benefit, give a specific number, name a real customer and outcome, add a third-party validation. Round numbers from nobody read as invented. "40% faster" from no source is a claim. "38% faster from a named study" is evidence, because the specificity signals someone measured it.

For B2B buyers, the highest-trust layer is customer ROI evidence. It is also the hardest to get on the record, so the team that solves the supply problem wins disproportionate trust. Customer ROI evidence is the highest-trust proof point for B2B buyers

"UserEvidence surveyed 619 B2B buyers and 51% said it's the most trustworthy form of evidence."

· Jason Oakley, LinkedIn, 2026-04-10

Proof is most persuasive adjacent to the claim it supports, not segregated into a testimonials block at the bottom. Pair each pain claim with its own solution claim and its own proof, per beat, rather than running all pain, then all solution, then all proof. Pitch a vision as pain → solution → proof, interleaved per beat, not three sequential acts

"Words will only get you so far... Figma practices what it preaches in terms of the future being visual communication."

· Mihika Kapoor on Lenny's Podcast, 2026-04-28

A testimonial section earns its place only when it answers three buyer questions: did this solve a problem like mine, did it work for people like us, and do I believe this is real and not marketing fluff. A testimonial section must answer three questions to drive conversion.

09 Build the hero on the five-second trinity. Lead with capability, not vision.

The homepage hero is the one artifact always one click away from every prospect, so it has to hold the positioning under scrutiny. Lead with the trinity: the specific use case, the alternative the buyer is using today, and the result the product produces, readable in roughly five seconds. B2B homepages must communicate use case, alternative, and result in five seconds

Vision-led copy ("reimagine how teams work together") fails because the buyer cannot tell whether the product is for them. Lead with capability, what someone actually does with the product. "Build pipeline" is an aspiration. "Call 300 leads a day without manual dialing" is a capability, and the capability converts.

Run the comprehension test before you trust the hero. Miller's caveman version is the blunt form. Could a caveman understand your homepage?, three questions, no marketing vocabulary

"Could a caveman look at your website and immediately know what you offer, how it will make their life better, and what they need to do to buy it?"

· Donald Miller, Building a StoryBrand — The Caveman Test, 2017-10-10

For B2B, optimize each section to send the reader to the next section, not to convert on the spot. Buyers consume across multiple visits, then convert weeks later. Optimizing each section for "convert now" reads as desperate and triggers procurement skepticism. Conversion isn't the goal of a B2B landing page, consumption is. Send them to the next section.

10 Write CTAs that are honest and specific. The action on click matches the button.

A generic CTA ("Learn More," "Start Now") makes the user guess what happens next, and the guess is friction. A specific CTA that names the outcome in the buyer's language removes the guess. Nikolai Golos reports specific, outcome-promising CTAs converting far better than generic ones. Specific CTAs that promise a clear outcome convert 269% better than generic ones. Mengqi Pei reports a sticky, solid CTA lifting signups by removing the scroll-back friction between intent and action. Sticky CTAs remove friction and boost conversion by 28%.

Then make the verb honest. "Book a demo" if it is a demo. "Start free trial" if it is a trial. "See the product" if it is a video. The reader who clicks expecting one thing and finds another does not convert, and the mismatch costs trust you do not get back.

11 Protect the voice from the polish pass. Voice is the one thing AI cannot replicate.

In a market flooded with competent AI prose, the only defensible advantage is a voice that sounds like it could only come from one source. In an AI-flooded content market, voice is the only defensible advantage, distinct, authentic, sounds like one source

"the only defensible advantage is a distinctive, authentic voice that sounds like it could only come from one source"

· Ann Handley, Everybody Writes — Voice as Defensible Moat, 2024-04-01

Voice is not a style guide. It is a function of one specific person's perspective, word habits, and lived experience, which is exactly why AI cannot produce it. AI defensibility comes from non-AI moats Harland's operational version: write the way you specifically speak, not the abstract median of how people speak. Voice quirks aren't bugs, they're the only thing AI cannot replicate Handley's direct-content rule: write as if to a single subscriber, use "you" liberally, and cut anything with a whiff of "Dear Valued Customers." Write as if to a single subscriber, use "you" liberally and remove anything with a whiff of "Dear Valued Customers"

The danger of an AI polish pass is not that it makes the draft bad. It makes the draft generic, because the model is the average of all written prose and sands every deliberate rule-break back to expectation. AI prose can't violate expectation because it IS expectation, protect the smallest deliberate rule-break from every polish pass

"AI prose can't violate expectation because it is expectation. It's the average of everything."

· Ann Handley, What AI Would Delete From Great Writing, 2026-05-03

The operating rule: mark the smallest signature quirk before the polish runs, then verify it survived. Keep a human verifying the voice on the marked elements, not just the whole piece. Verification, not execution, is the irreplaceable human job Schafer's three-pass process builds this in: pass one sets the pole and the headline, pass two injects the voice, pass three integrates and cuts anything that entertains but does not persuade. Three sequential passes, pole + headline first, voice second, integration third, single-pass writing collapses persuasion and personality into mush Single-pass writing collapses persuasion and personality into mush.

12 Run the honesty test before you ship. Would you say it to the reader's face?

Two old tests catch the manipulation that survives because the writer never imagined being seen. Ogilvy's family test is the floor. Never write an advertisement you wouldn't want your own family to read, the family test as ethical filter

"Never Write an Advertisement Which You Wouldn't Want Your Own Family To Read."

· Ogilvy, Confessions of an Advertising Man, p. 87 (1963)

Schafer's face test is the descendant: would you say this line to the reader, in person, without flinching? If you would soften it or retract it across the table, it fails. Would you say it to the reader's face without flinching?, the test that separates copy from manipulation

"The test: would you say this to the reader's face, in person, without flinching? If not, it is manipulation."

· Cole Schafer, Sticky Notes — The Honesty Test, 2024-01-15

Both tests trace to the same standard: the work is judged by what it sells, not by what the team admires. Advertising is selling, not art, when I write an advert I don't want you to find it 'creative'; I want you to find it so interesting you buy the product

"When I write an advertisement, I don't want you to tell me that you find it 'creative'. I want you to find it so interesting that you _buy the product_."

· Ogilvy, Ogilvy on Advertising, p. 7