It’s 11pm, and you promised yourself the new homepage would go live this week. You open a blank doc, type “The all-in-one platform for”, and delete it. You try a feature instead, then a customer quote, then you check Slack. If you’re a technical founder at an early-stage B2B tech company, handling the marketing because nobody else is, you know this evening. And somewhere around the third deleted headline you open a new tab and ask ChatGPT how to write homepage copy.
You probably take that blank page as proof you can’t write. But the words usually stall because a decision one or two layers above them never got made. That’s good news: a missing decision is far easier to fix than a talent you think you lack.
The homepage also carries more weight than it did even a year ago. In March 2026, Gartner reported that 67% of B2B buyers prefer a buying experience without a sales rep. So a homepage that explains you well does a lot of your selling to people you haven’t met yet.
The difference between positioning, messaging, and copy
Positioning, messaging, and copy get used interchangeably, but each one names a different piece of work with a different job.
Positioning is a decision. It settles which category you compete in, who you’re for, and why you win against whatever that buyer would do otherwise. It lives inside the company: in a doc, in the founder’s head, in the way both cofounders answer “so what do you do?” at a birthday party. A buyer never reads your positioning directly.
Messaging turns that decision into things you say out loud. It’s the small set of claims, proof points, and phrases that every page, deck, and sales call draws from. It’s still an internal tool, but it’s written for a buyer’s ears.
Copy is the specific words on a specific page that get a specific person to act: the hero, the subhead, the button label. It’s the only layer a buyer ever sees.
Dale Carnegie wrote in The Quick & Easy Way to Effective Speaking that people judge us on “what we say and how we say it”, and that split maps onto the bottom two layers. Messaging decides what you say. Copy decides how you say it on one page, to one reader, in the space a hero section gives you.
Each layer feeds the one below it
Positioning feeds messaging, and messaging feeds copy. That order only runs one way, so polishing a lower layer can’t settle an upper one. You can rewrite a hero forty times, but if the team hasn’t decided who the product is for, every rewrite is a fresh guess at the positioning dressed up as a headline.
So when someone asks me to write a homepage, I start from the positioning, and any copywriter worth their salt does the same. Sometimes it’s already done. Sometimes you work it out together with someone on the team, and some copywriters settle it themselves. But whoever does it has to step out of the office and talk to people, because you can almost never settle positioning on the founder’s input alone (unless the founder has recently done a lot of customer and market research themselves).
The order matters even more once you look at how buyers decide. The 6sense 2025 Buyer Experience Report, which surveyed nearly 4,000 B2B buyers, found they make first contact with a seller about 61% of the way through their buying journey, and they’re the ones reaching out 79% of the time. The vendor that eventually won was already on the day-one shortlist 95% of the time.
So most of the deciding happens while you’re not in the room. Your website and your deck have to tell the story on their own, and your sales calls have to tell the same one. That only works when all three are cut from the same messaging, and the messaging only holds together when it’s cut from one positioning. Get the top layer settled and the website starts doing the explaining you’d otherwise do on a call.
What the right order looked like at ONAH
Sophie, the founder of the HR inbox platform ONAH, wanted to move fast in March 2026 but wasn’t sure what came next. She had just put a lot of time into her own research, so her input gave me enough to settle the positioning on. I pushed for the website first, with Wednesday 25 March as the target, and it went live that week in time for Supernova. (I wrote about how that week unfolded in writing a positioning statement when you have three customers.)
The positioning came first. ONAH was still pre-launch at the time (it has been running pilots since August 2026), so the decision was about where the product should stand: routine HR email handled end to end, so HR people get their time back for the work that needs their expertise. That decision set the category, the buyer, and the reason to care, and you can still read it almost word for word further down the site: “The only HR inbox automation platform that resolves routine emails end-to-end from verified sources.”
The messaging turned that decision into one line anyone at ONAH can say out loud: “Your HR inbox runs itself. You run HR.”
The copy came last. The hero reads “Stop managing HR mails. Start doing HR work.”, and a section further down says “With Onah, your inbox no longer sets the agenda.” Both carry the same message, shaped for their spot on the page: the hero names the frustration, the section names the result. With the two layers above already settled, each line was easy to judge, because it either said what the messaging said or it didn’t.
Sophie was surprised the copy was AI-written. But AI-written or human-written, copy is only as specific as the layers above it.
Why your website copy isn’t converting when you write it first
When the copy comes first, the founder (or the intern hired to “do the marketing”) sits down to write headlines on top of a position nobody has decided. Without a category decision, the safest choice is the widest one, so the hero becomes “the all-in-one platform for modern teams”. Without a settled reason to win, the page lists features, the one thing everyone on the team agrees on.
The result is accurate and interchangeable. Swap your logo for a competitor’s and nothing on the page reads wrong.
Two beliefs keep this going. One says a good product explains itself, so the homepage only needs to describe it. The other says an intern can handle marketing. An intern can write decent sentences, but a brief that says “do the marketing” hands them the call on what the company is and who it’s for, and that call belongs to the founder.
Rob Fitzpatrick describes a version of this in Write Useful Books. He got unstuck on a book’s scope by deleting rather than writing more beautifully, once he asked “What does my ideal reader already know and believe?” That’s a positioning question, and it’s the one to ask when your headline drafts keep dying. Stop editing the sentence and go up a layer.
Why positioning is the hardest layer to settle
Positioning is a decision rather than a deliverable, and deciding feels uncomfortable. Picking a category rules other categories out, and picking a buyer means some people won’t see themselves on your homepage.
In 1957, C. Northcote Parkinson invented a finance committee to make a point. On its agenda was a £10 million contract for an atomic reactor, and it passed in two and a half minutes, because the members assumed the people working on it understood it. Next came a £350 bicycle shed. That one took forty-five minutes and might save £50, because everyone can picture a bicycle shed.
Put a homepage draft in front of a team and you get the same meeting. The category, the buyer, and the reason you win get a nod. The hero headline gets the forty-five minutes, because everyone has an opinion about words and far fewer people want to own the category call.
ZYGO shows what owning that call looks like. In the first week of March 2026, while I was building its homepage, the team spent real time debating acceptable terms for the software before landing on “robotics software platform”. It’s a short phrase, and it deserved the time, because I think of category as the bucket a buyer drops you into before they read another word. Move the bucket and every headline, comparison, and sales slide moves with it. So if your positioning still shifts every few months, you can still write copy. You just need a working decision, written down, that you’re willing to update.
How to tell which layer needs work
Finding the layer that needs work takes three questions. Work through them in order and stop at the first one that gets a shaky answer.
- Does your team agree on the category, who it’s for, and why you win? Ask your cofounder and your first hire to write one sentence each about what the company does, separately. If the sentences describe different companies, the work sits at the positioning layer. (I haven’t run this exact test with a client yet, and I’m trying it with one for the first time in late September 2026.) My positioning audit walks through it step by step.
- Do your homepage and your sales pitch tell the same story? If the team agrees internally but the deck leads with one thing and the website with another, the work sits at the messaging layer. Buyers notice: in a Gartner survey released in June 2025, 69% of B2B buyers reported inconsistencies between what a supplier’s website said and what its sellers told them. A messaging guide your team can use is where you fix that.
- Is the message clear, but people don’t click, or the tone feels off? Then it’s copy, and that’s the cheapest layer to work on. Test two headlines, cut the subhead in half, rewrite the button so it says what happens next.
Start at the top and the words get easier
Decide the positioning, turn it into messaging, then write the copy. Done in that order, the 11pm blank page turns into a translation job: you’re putting a decision you already made into words for one page and one reader.
The upside reaches past the homepage, too. Ramli John makes the point in Product-Led Onboarding that clear positioning and messaging help you “screen out people who shouldn’t be signing up”, so you can spend your time on the ones who should. A buyer who isn’t a fit reads a sharp homepage, recognises it’s built for someone else, and moves on without booking a call. The people who do book have already seen themselves on the page.
It’s also the order I work in. When I start with a company, onboarding gathers what the founders know and tells me where to research. The ICP gets filled in from that research, and the messaging builds on the ICP and whatever else the research turned up. Only then do I build the live assets on top: the homepage, the outbound sequence, the sales deck. If your homepage draft keeps stalling at the headline, start one layer up. And if you’d like a second pair of eyes on which layer that is, I’m happy to take a look.