Most conversion problems are not design problems. They are message problems. A visitor lands on a page, scans the copy, fails to grasp what is on offer or why it should matter to them, and leaves. Time and again, better copy, written in the customer’s own words, organised around their buying psychology, and edited for clarity, outperforms a visual redesign. The good news is that good copy is not a talent you either have or lack. It is a repeatable system: audit what you have, mine the right messages, build a value proposition, sequence a hierarchy, draft, edit, and format for the way people actually read. This guide walks through that whole system.
Start by tearing down what you have
Before writing a word, audit the existing page through three lenses applied in order.
The first is a conversion heuristic that weights the forces acting on a visitor. In plain form it reads: conversion is driven by four parts motivation, three parts value, and two parts incentive minus friction, with anxiety subtracted. Written out, that is C = 4M + 3V + 2(I − F) − A. The weighting is the lesson. Motivation dominates everything, which means you cannot rescue a weak offer with bigger discounts. If the visitor does not want what you are selling, no coupon saves you. So you fix motivation first, then value, and only then turn to friction and anxiety. Score each variable on every page in the funnel, the homepage, the landing page, the product page, the checkout, on a simple scale where zero means not addressed, one means present but poorly executed, and two means done well. The lowest-scoring variable is your first priority, and the heuristic turns a vague sense that something is off into a ranked to-do list.
The second lens is Cialdini’s principles of persuasion, used as a checklist for what the page is missing. Is there social proof in the form of reviews, ratings, and logos? Authority through expert quotes, press, or certifications? Likability through a founder’s voice and a conversational tone? Scarcity, reciprocity, commitment, and a sense of shared identity? A page with no social proof and no authority signal is quietly working against itself, and naming the gaps is half the fix.
The third lens is Claude Hopkins’ century-old rules from direct-response advertising, which still hold. Be specific, because “saves three hours a week” beats “saves time.” Frame everything as a service to the customer. Tell the full story rather than cutting copy just to look minimal, because an online reader who is still reading is self-selected and interested. And write as a salesperson talking one to one, to a single person rather than a demographic. The most common violation here is the vague superlative, the world-class, best-in-class, revolutionary language that says nothing, and every one of those should be replaced with a number, a name, or a concrete outcome.
Mine the words your customers already use
You do not invent the best copy, you find it, in the exact language customers use to describe their problem, their hesitation, and the outcome they want. The process is methodical. Start from three to five keywords your product or category is actually searched by, which anchors you to real intent rather than internal jargon. Search each one and study the ads, which are paid copy already proven to convert, along with the organic headlines and the related questions people ask. Then go to where customers speak honestly, Amazon, G2, Trustpilot, Capterra, Reddit, and niche forums, and read the three- and four-star reviews in particular, because those contain both praise and objection. Copy the exact phrases verbatim into a sheet alongside their source and theme, keeping the raw quotes intact before you interpret them. Finally, categorise and rank by frequency, because the phrases that recur are the ones your audience keeps returning to.
Every phrase you collect belongs to one of three buckets, and the buckets map directly onto the heuristic. Motivation messages capture desired outcomes, pain points, and purchase triggers, and they raise the M in your formula, so they belong in headlines and hero copy above the fold. Value messages capture unique benefits, delightful features, and deal-breaker needs, and they raise V, so they belong in feature sections and benefit bullets. Anxiety messages capture uncertainties, objections, and perceived risks, and they reduce A, so they belong in FAQs, guarantees, and the copy right next to the call to action. Matching the message type to the variable it moves is the whole game.
Go deeper with direct research
Online mining gives you volume, but talking to people gives you depth, and the two kinds of survey serve different ends. A visitor survey reaches people currently on your site who have a problem but may not know your solution, so the question to ask is what one thing would make them sign up today, surfacing motivations and anxieties. A customer survey reaches people who already bought and can evaluate the product, so the question to ask is what almost stopped them buying, surfacing objections and the moments value clicked.
A few rules make surveys actually work. Never use the words survey or feedback in the invite, because both signal effort and low reward, and instead lead with a benefit and keep the invite to one sentence. Time visitor surveys to trigger after thirty to sixty seconds or on exit intent, and send customer surveys seven to fourteen days after purchase, long enough to form an opinion but not so long they forget. For copy mining, one open-ended question outperforms ten rating scales, and it should ask about the moment of decision, what was going through their mind when they decided to buy, rather than general satisfaction. Better still, interview five recent customers and have them walk through their purchase story in order, listening for the switch moment, the “I was using X, then Y happened, so I tried this.” That switch moment is very often your headline.
Build a value proposition that actually differentiates
A value proposition is not a tagline. It is the single clearest answer to why this specific visitor should choose this specific product right now, and it lives at one precise intersection: what the customer wants, what your product offers, and what makes you different. A feature you have that customers do not care about is irrelevant. A want that you and every competitor satisfy is not a differentiator. Only claims sitting where all three overlap belong in your headline.
Stress-test every candidate with two questions. Ask “so what?” of each claim, and if the honest answer is that nobody cares, you have written a feature, not a benefit, so translate it into an outcome. Then ask “prove it,” and if you cannot back the claim with a number, a name, a testimonial, or a demonstration, it is a hollow superlative that needs either evidence or replacement. To choose which benefits to lead with, score each feature by the severity and frequency of the pain it solves, multiply the two, and lead with the highest scores. A feature that fixes a severe, frequent problem earns the headline, while a minor convenience gets deprioritised no matter how proud the team is of building it.
Sequence the message so people follow it
Having the right messages is not enough, because the order decides whether visitors follow your logic or leave before the call to action. The hierarchy is fixed. Lead with motivation, meeting the visitor’s desire or pain. Move to value, showing how you solve it better than the alternatives. Then handle anxiety, addressing the objections that would stop them acting. Only then make the ask, when resistance is already low. The cardinal rule is never to lead with the call to action, because a visitor who does not yet understand why they should act will not click a button simply because it is visible.
How much copy you need depends on awareness. An unaware visitor who does not even know they have the problem needs long copy that educates before it sells. A problem-aware visitor needs medium copy bridging their pain to your product. A solution-aware visitor needs to be differentiated on value. A product-aware visitor mostly needs their anxiety removed and a push to act. And the most-aware visitor, ready to buy, needs almost nothing beyond price, guarantee, and button. This is why a product page reached from a branded search ad can be far shorter than a cold landing page from a display ad. Underneath it all, every page should still tell a complete story in three acts: the why that sets the context and states the value proposition, the try that shows features, benefits, and proof, and the buy that makes the next step obvious and risk-free. Pages that skip straight to the buy, all features and buttons with no motivation, give the visitor no emotional reason to act.
Draft, then edit hard
The first draft is not about quality, it is about transferring your mined messages and your hierarchy into coherent prose. Before writing, lock down the audience’s awareness level, choose the page’s single conversion goal, confirm your value proposition and top three motivation messages, and decide which anxieties belong on this page versus elsewhere in the funnel. Then write for one person in the second person, match the exact vocabulary from your mining, and use the customer’s word even when a slicker one exists, so “hassle” stays “hassle” rather than becoming “friction.” Write the headline last, once the body has made the real promise clear.
A strong draft still needs editing, and a handful of rules do most of the work. Be clear, so every sentence has one obvious meaning, turning “helps teams work better” into “your team closes twice as many tickets a day.” Match the reader’s mindset by mirroring their internal conversation rather than your feature list. Lead with the most compelling benefit instead of a safe generic. Replace adjectives with quantifiable proof wherever possible, so “significantly faster” becomes “renders in 0.4 seconds, six times the industry average.” Paint a concrete picture, trading “reduces stress” for “you close your laptop on Friday knowing the reports are already done.” Pair copy with a relevant visual so neither has to carry the whole load. And cut anything that does not advance motivation, value, or anxiety reduction, starting with the boilerplate company-history intro that should simply be deleted in favour of opening on the customer’s problem. A practical editing pass helps: read it aloud to expose weak rhythm, delete the first sentence of each paragraph to see if it was just a warm-up, flag every adjective and ask whether a number can replace it, and check that every heading conveys a benefit rather than a label.
Format for how people actually read
Copy and layout are not separate disciplines for conversion, because arrangement determines whether the copy is read at all. People scan in an F-shape, across the top and down the left, so the headline, the primary benefit, and the call to action belong in the top-left. Targets that are larger and closer are easier to act on, so the call to action should be unmissable and the headline should dominate the type hierarchy. Use heading size and weight to walk the eye through the motivation-value-anxiety-CTA sequence, and never let a button visually outrank the motivation copy that justifies it. Fewer choices mean faster decisions, so isolate the primary action with whitespace and let secondary options recede rather than offering three equally loud buttons.
Readability has hard floors that quietly kill engagement when ignored: body text of at least sixteen pixels, line height around one and a half times the font size, a measure no wider than about nine hundred pixels, and paragraphs broken every three or four lines. Long lines and tight spacing make reading physically uncomfortable, and visitors stop not because the copy is bad but because their eyes are tired. Directional cues help when used sparingly, a face in an image looking toward the headline pulls the gaze there, and every arrow or line should point toward the next step in the hierarchy. Finally, the squint test: blur your eyes at the page and see what remains most visible, because that is your dominant element, and it should be the primary call to action or the headline. If a stock photo wins the squint test, the button is losing. Wireframe before finalising copy, since the structure constrains the words and the words constrain the structure, and the two are best iterated together.
The system in one pass
Strip it back and the system is the same whether you sell pet doors or enterprise software. Audit the page with the heuristic to find the weakest variable. Mine messages to find the language your customers already use. Build the value proposition at the intersection of want, offering, and differentiation. Sequence the hierarchy from motivation through value and anxiety to the ask. Draft by transferring those mined messages into the structure. Edit against the rules for clarity, specificity, proof, and concrete imagery. And format for scanning so the copy actually gets read. Run a real page through all seven steps and the uplift follows, not because of any one clever line, but because the whole message now meets the visitor where they are and walks them, in their own words, to the decision.
See you soon.
