Copywriting is writing intended to persuade someone to take an action: request a quote, book a call, subscribe, buy. It is the text on a homepage, a service page, an ad, a form, an email. Its job is not to sound impressive; its job is to make a decision easier by telling the reader clearly what is on offer, what it is worth, and what to do next.
It is distinct from content writing, which is meant to inform and build trust over time. A guide explaining how permitting works is content writing. The page asking you to start a project is copywriting. Most business sites need both, and the common failure is applying the wrong register: service pages that read like essays, or educational articles that read like sales letters.
Copywriting is mostly research
The part that distinguishes good copy from competent-sounding copy happens before any writing. Copy that works is assembled from things the business and its customers actually said, not invented at the keyboard.
The sources worth mining: sales calls and the objections that come up in them, the questions people ask in inquiry forms, support tickets, reviews of you and of your competitors, and the words customers use to describe the problem before they know the industry term for it. A single line from a review, “I just wanted someone who would call me back,” is worth more than an afternoon of brainstorming headlines.
Competitor copy is worth reading too, mainly to find what everyone is saying so you can say something else. When six firms in a market all claim quality, service, and experience, none of those words distinguish anyone.
What actually persuades
Specificity. “Most projects are live in six to eight weeks” says something; “fast turnaround” does not. Numbers, timelines, constraints, and named limitations all read as true because vague claims are what people expect.
The outcome, then the mechanism. People buy the result and then want to know how it works. Lead with what changes for them, support it with what you do.
Clarity over cleverness. A headline that requires a second read has already lost. Wordplay is worth it only when the plain version is still obvious.
Evidence. Named clients, real numbers, quoted customers, work that can be looked at. Unsupported superlatives are noise.
Objections answered on the page. Cost, timeline, what happens if it goes wrong, what is not included. Addressing the uncomfortable question directly is more persuasive than avoiding it, because the reader is thinking about it either way.
Honesty about fit. Saying who the service is not for makes every other claim more credible, and filters out inquiries that were going to end badly.
Structure on a page
Most people scan before they read. Copy should survive being skimmed.
The headline states what this is and who it is for, in plain words.
The subhead adds the detail that makes the headline concrete.
The opening confirms the reader is in the right place, quickly.
Subheadings carry the argument on their own. Read in sequence with nothing else, they should still make sense.
Short paragraphs, because walls of text get skipped regardless of quality.
Proof placed near the claims it supports, not quarantined in a testimonials section at the bottom.
One clear action, repeated where a reader might be ready for it. Competing calls to action split attention and reduce the chance of either.
The call to action
The call to action is where copy either converts or does not. What helps: naming the specific action rather than a generic verb, stating what happens next so the click feels low-risk, and reducing the perceived commitment. “Book a discovery call” tells someone what they are agreeing to. “Learn more” tells them nothing and commits them to nothing, which is why it rarely works.
One primary action per page, stated the same way each time, outperforms a menu of options. A reader given three equally weighted choices frequently makes none.
Copywriting for search and AI
Commercial pages have to work for a reader deciding whether to contact you and for machines deciding whether to show or cite the page. These pull in the same direction more often than people assume.
Keywords belong in copy where they occur naturally, particularly in the headline, an early paragraph, and the page title, because those are the words your buyer uses too. Forcing them past that point makes copy worse for both audiences, and search engines have not rewarded density for a very long time.
The genuinely useful overlap is structure. Stating the offer plainly near the top, using honest headings, and putting specifics in the text rather than in an image is what makes a page easy to skim, easy to extract a direct answer from, and easy for an AI system to represent accurately. Copy built around vague positioning gives a machine nothing concrete to say about you, which is its own kind of invisibility.
One concrete rule: prices, service areas, timelines, and contact details belong in text. In an image or loaded by script, they may be missed entirely by crawlers that do not render JavaScript.
Where AI fits, and where it does not
AI tools are genuinely useful in copywriting for generating options to react against, restructuring a draft, tightening long sentences, catching inconsistencies, and producing the tenth headline variant when the first nine are stale.
What they cannot supply is the raw material: what your clients actually said, which objection kills the most deals, what your firm does differently, and which claims you can stand behind. Copy generated without that produces text that is fluent, plausible, and interchangeable with every competitor’s, which is the opposite of what copywriting is for.
The sensible division is to do the research yourself, use AI to accelerate drafting and revision, and keep final judgment on claims and voice with the person who will have to answer for them.
Common mistakes
Writing about yourself. Opening with company history when the reader wants to know whether you solve their problem.
Industry language. Terms your buyers do not use, or use to mean something else.
Claims without evidence. Every competitor says the same words; only proof distinguishes.
Burying the offer. If someone has to scroll to learn what you do, most will not.
Too many actions. Competing buttons split attention.
No objection handling. Silence on price or timeline reads as evasion.
Writing once and never revisiting. Copy is testable, and the first version is rarely the best one.
Testing rather than arguing
Copy decisions are frequently settled by whoever is most senior in the room, which is a poor method. Headlines, calls to action, form length, and the order of sections are all measurable, and disagreements are cheaper to resolve with a test than with a meeting. That is the domain of conversion rate optimization, and it is where copywriting stops being a matter of taste.
The caveat worth stating: a test needs enough conversions to mean anything. On a site with a handful of inquiries a month, judgment and customer research will get you further than underpowered experiments.
Copy that does its job
We write site and campaign copy from customer research rather than from imagination, as part of our content marketing services, and test the parts worth testing through our conversion rate optimization services. If your site describes what you do but nobody inquires, book a discovery call.
