Brand & Marketing
When a landing page underperforms, the instinct is to rewrite the headline. Nine times out of ten the real problem is further down the page: an unclear offer, proof that shows up too late, or friction right before the button. Fix those first.

Every round of “our conversion rate is low” feedback starts in the same place: someone wants a punchier headline. I understand the instinct, the headline is the first thing anyone sees, so it feels like the lever. In practice it's the smallest one. People don't bounce because a headline was merely good instead of brilliant. They bounce because they couldn't tell, within a few seconds, whether the page was for them.
Before rewriting a single word at the top of the page, I want to know three things: can a stranger state your offer back to me in one sentence, is there proof before the point where they'd normally start doubting you, and does anything get harder right before the button. Those three answer for most of the gap between “visited” and “converted.” The headline is the last five percent, not the first.
Every page needs one sentence that a stranger could repeat back accurately: what they get, who it's for, and why they should believe it's real. If you can't write that sentence cleanly, the page can't either, no matter how good the individual lines are. This is usually where the actual problem lives. Teams write pages full of good sentences that never quite add up to a clear offer, because the offer itself was never pinned down in a meeting, only implied.
A useful test: read only the H1, the subhead, and the first line under the CTA button, nothing else. If those three lines alone don't tell someone what they're being asked to do and why, no amount of mid-page copy will rescue it. We write that sentence before we write anything else on a page, and we agree it with the client before a single word of “real” copy exists.
Most pages have proof somewhere, a testimonial, a logo strip, a number. The mistake is placement, not absence. Proof that sits at the bottom of the page arrives after the reader has already decided whether to believe you, which means it's reassuring the people who were going to convert anyway and doing nothing for the people who weren't.
The fix is to put your strongest piece of proof right where the natural objection would form. If the claim is “we'll save you time,” the proof needs to show up near the first mention of time saved, not three sections later. If the claim is about reliability, put a number or a named client near the reliability claim, not in a generic “trusted by” strip at the very bottom. Readers raise objections in real time as they scroll; proof only works if it answers the objection at the moment it's raised.
Nobody abandons a form because the copy wasn't lyrical. They abandon it because it asked for one field too many, or the button said something vague, or the next step wasn't obvious. This is the least glamorous part of copywriting and the part with the clearest, most measurable payoff.
| Friction point | What it costs you | The fix |
|---|---|---|
| A form field you don't actually need yet | A chunk of people who'd have continued with fewer questions | Ask for the minimum to start the conversation; collect the rest later |
| A vague button label (“Submit”, “Learn more”) | A moment of doubt about what happens next, right at the decision point | State the outcome: “Get my quote”, “See pricing”, “Start the project” |
| No indication of what happens after they click | Hesitation from people who dislike disappearing into a black box | One line under the button: “We reply within one business day” |
| A CTA that repeats the same words everywhere on the page | Nothing stands out as the one thing to do | Vary the supporting line around each CTA to match what's directly above it |
None of this is clever. It's closer to checking a seatbelt than writing. But it's usually worth more to the conversion rate than a better headline, because it removes a real obstacle instead of adding a nicer sentence next to an obstacle that's still there.
Take a typical early-stage SaaS landing page: headline about “transforming your workflow,” a features grid, testimonials at the very bottom, and a button that says “Get Started.” It reads fine. It converts poorly, because a stranger finishing the headline still doesn't know what the product actually does, the proof shows up after anyone sceptical has already left, and “Get Started” commits to nothing specific.
The rewrite rarely touches the headline much. It tightens the offer sentence (what, who, why believe it), pulls one concrete result up next to the claim it supports, changes the button to name the actual outcome (“See a live demo,” not “Get Started”), and adds a one-line answer to “what happens next.” The page usually looks almost the same. It reads completely differently.
When a client asks us to “fix the copy,” this is the order, because each step tends to expose how much the next one is actually needed:
Skip straight to step five and you'll ship a better-sounding page with the same conversion rate. Work the list in order and the headline often needs barely any work at all, because by then it's just naming a clear, well-proven, low-friction offer.
If a page isn't pulling its weight, send us the link and we'll tell you honestly where it's actually losing people. Copy is one part of a wider branding and marketing job, and we're happy to just look at the one page if that's all you need.
Sahra
Sahra writes the words that actually ship: landing pages, emails, and the copy that sits between a good product and someone bothering to try it.
Turn an idea into a product, brand, or campaign, with one team from strategy to launch.
start a project
Security
Most breaches aren't clever. They're a reused password, an open storage bucket, an API that trusts whatever the client tells it, or a key that's been sitting in git history for a year. Here's what genuinely moves the needle before launch, and what can wait.
1 Oct 2026 · 10 min read

Engineering
Good QA was never about finding every bug. It's about deciding, on purpose and out loud, which bugs are allowed to ship. Here's how that decision actually gets made when the launch date isn't moving.
27 Sep 2026 · 9 min read

Engineering
The best tech stack for a first build is a boring one: proven, well-documented, and easy to hire for. Here's what “boring” actually means in 2026, the traps that sink second years, and a sensible default to start from.
11 Sep 2026 · 10 min read