⚡ Quick answer: Nielsen Norman Group measured it: visitors have time to read at most 28% of the words on a page, and 20% is more realistic. Half the information gets read only on pages of 111 words or fewer — while the average page holds 593. So your copy isn’t being read, it’s being scanned. Which leads straight to the practical conclusions: a headline with a concrete number, social proof with names and figures, and a form with no more than four fields.
Copy is usually the weakest link in a website project. The design gets debated for weeks, then service descriptions get written the evening before launch — and afterwards everyone wonders why the page doesn’t convert.
The Nielsen Norman Group research measured time rather than impressions: how many seconds someone spends on a page, and how many words they can physically read in that window at 250 words per minute.
The numbers came out like this. On average a visitor has time for 28% of the words — and that’s the ceiling, assuming they do nothing else at all. The realistic estimate is closer to 20%. Half the text gets read only on very short pages, up to 111 words.
Each additional 100 words holds someone roughly 4.4 seconds longer. Meaning longer copy doesn’t get read in full — it just gets scanned slightly longer.
That isn’t an argument for writing less. It’s an argument for writing so that the 20% people do notice is the part that matters.
If copy is scanned rather than read, what works isn’t the elegance of your seventh paragraph — it’s whatever the eye catches while scrolling: headings, opening sentences, bold text, captions, numbers.
In practice that means a few things.
A study of 2,000 pages A/B-tested across Q4 2025 to Q1 2026 (minimum 1,000 sessions per variant, 95% significance threshold) produced a clear result: the best-performing hero was a headline built around one big number — plus 18% against a standard image-led control.
The logic is obvious. “Professional turnkey website development” carries no information — everyone writes that. “Corporate site in 4 weeks from $1,500” carries three concrete things at once: what, when, how much.
A simple test: if your headline could be moved to a competitor’s site without a single edit, it’s empty.
The same study measured different trust formats separately, and the spread is telling.
| Format | Conversion lift |
|---|---|
| Named customer count with specifics | +22% |
| A single named testimonial | +14% |
| A strip of client logos | +8% |
| Generic “trusted by thousands” | essentially zero |
One conclusion: specificity wins. Vague phrasing about thousands of satisfied clients now performs exactly like its complete absence — people have learned not to see it.
A working testimonial carries four elements: full name, role and company, the specific problem before, and a measurable result after. “Everything was great, I recommend them” adds nothing.
The most expensive small detail on a service page. Per the same study, a one-field form converts at 12.4%, while forms with six or more sit at 3.1%. Each field beyond the fourth roughly halves conversion.
The question worth asking of every field: is this information needed right now, or can it be asked in conversation? “Company name,” “how did you hear about us,” “budget” — all of these can usually wait.
Load speed cuts the same way: conversion holds at 4.4% under one second and drops to 1.7% at four seconds. We covered how to speed up a site separately — it’s the same funnel viewed from the technical side.
The most widespread service-page structure looks like this: company history first, then our values, then certifications, and somewhere in the middle — the actual service. That’s an order convenient for whoever writes it and inconvenient for whoever reads it.
Someone who opened the page is resolving one question: will this company solve my problem and what will it cost. A founding date of 2014 doesn’t answer that.
The working order is reversed: the client’s problem in their words first, then the solution, then proof you’ve done it before, and only then who you are. In Netloria’s experience, reordering blocks without changing a word of the copy often delivers more than a full rewrite in the old sequence.
| Block | What it answers |
|---|---|
| Headline with specifics | What is this and is it for me |
| The client’s problem | They understand my situation |
| What the service includes | What exactly am I paying for |
| Price or price range | Is this within my budget |
| How the work runs | What happens next |
| Case studies with figures | You’ve done this before |
| Objections answered | What if it goes wrong |
| A form with 2-4 fields | What to do now |
Pricing deserves its own note. People hide it to avoid “scaring anyone off” — and end up scaring more people off: someone who finds no reference point at all usually closes the tab rather than writing to ask. Even a range of “$1,500 to $4,000 depending on scope” beats silence.
A few rules that deliver the most for the least effort.
Replace abstractions with specifics. “Individual approach” means nothing. “We don’t take projects where there’s no access to someone who knows the product” means something.
Cut the officialese. “We carry out the implementation of projects” means “we build.” Longer words don’t add authority, they add friction.
Write numerals, not words. “4 weeks” registers faster than “four weeks” while scrolling — digits catch the eye, words blend into the text.
Read it aloud. A sentence you can’t say in one breath should be cut in two.
Every business has a set of questions clients either ask out loud or hold silently before walking away. Those are what decide an enquiry — not your advantages, but their doubts.
The typical set looks roughly like this: why are you more expensive than that guy on a freelance marketplace; what happens if I don’t like the result; who will actually work on this and won’t you hand it to a junior; what if you disappear mid-project; will I be able to make edits myself afterwards without you. These questions are uncomfortable, which is exactly why they almost never appear on a website — and that’s a mistake, because the competitor who answers them first takes the client before the first conversation even happens. At Netloria we ask clients during the brief to name the three most awkward questions they get, and it almost always turns out none of them are addressed anywhere on their site.
The format of the answer matters. Not “we guarantee quality,” but specifics: how many revision rounds are included, what happens if a deadline slips, who owns the credentials when the work is done.
There’s no universal answer, but there is a working guide: exactly enough to close every question that arises before a decision, and not a word more.
For a simple service with a clear outcome and fixed price, 300-500 words is plenty. For something complex, expensive and risky for the client, 1,500 or more may be needed, because there are more objections in play.
A way to check: write out every question clients actually ask before ordering. If each one has an answer on the page, the length is right. If there’s text answering none of them, it’s a candidate for deletion, however nicely written.
That last one happens more often than you’d think: service pages within one niche frequently differ only by company name, and it’s obvious to a client looking at their fifth site in a row.
A full rewrite isn’t always needed. Often a structural edit is enough, and it can be done quickly.
Start by marking every sentence in the current text that contains something concrete — figures, timelines, names, limits. There are usually surprisingly few. Then move them up: what sat in the second-to-last paragraph becomes a subheading.
Next, delete everything that could be pasted onto a competitor’s site unchanged. That’s the fastest way to halve the length while losing nothing. And finally, add one block that definitely isn’t there: the answer to the most awkward question clients ask you. That block is usually what separates a page that sells from a page that describes.
A separate category is text written “for keywords,” where a phrase like “web development services” repeats eight times. It reads badly and stopped conferring any search advantage long ago.
The modern approach is simpler: write for a human, use the keyword naturally in the heading, the first paragraph and one or two subheadings. Beyond that, completeness of the answer does the work, not phrase density. We covered ranking factors in more detail, and keyword density hasn’t sat near the top of that list for years.
The side effect is pleasant: copy written for people usually ranks better than the over-stuffed version anyway.
A feeling that “it reads better now” isn’t a metric. Watch two things: how many people reach the form, and how many of those submit it. If they reach it but don’t fill it in, the problem is the form or the price. If they don’t reach it, the problem sits higher, in the page structure.
One condition: don’t draw conclusions from a dozen visits. How much data makes a comparison honest is covered in our piece on website metrics.
And change one thing at a time. Rewrite the headline, shorten the form and add testimonials simultaneously, and you won’t know which one worked — which means you can’t repeat the win on other pages.
In Netloria’s work we usually start with the form and the headline: two of the cheapest edits with the largest expected effect. Everything else comes after those are exhausted.
If you have a service page pulling traffic but not enquiries — get in touch. We’re a web studio based in Ukraine, and we’ll look at where people actually stop before rewriting anything.
