wp.breakpoints

Designing for users who don’t read.

You are probably not going to read this whole thing. Good. It was built for that.

You’re going to skim this. Most people will read the bold lines, look at one chart, and leave. That’s not an insult to you or to me. It’s how the web has worked for almost thirty years.

So this essay is built the way we build pages. If you only read the bold, you still get the argument.

This isn’t new, and it isn’t your phone

People were scanning before feeds existed. In 1997, Jakob Nielsen tested how people read web pages. 79% scanned. 16% read word by word.1 That was dial-up. Nobody had a notification to blame.

A decade later he ran the math on real browsing data and found that on an average page, people read at most 28% of the words. He thought 20% was more likely.2 Chartbeat looked at 2 billion visits in 2014 and found 55% spent fewer than 15 seconds actively on a page.3 Their 2026 number for average engaged time is 26 seconds.4

You’ve also heard that humans now have an 8-second attention span, shorter than a goldfish. Don’t repeat that one. When the BBC tried to trace it, there was no research behind it.5 People watch three-hour podcasts. They have plenty of attention. They’re just not spending it on your second paragraph.

Skimming is triage, not laziness

People scan to decide whether to read. That’s a fair way to treat a page you didn’t ask for. They are sorting, quickly, for one thing: is this for me, and where’s the part I need?

The famous F-shaped heat map gets misread here. Designers saw it and started arranging content along an F. The researchers who found it say the opposite. The F is what eyes do to a wall of unformatted text, and they call it bad for users and bad for businesses.6 It’s a symptom. You don’t design for it. You design so it never shows up.

They have names for the better patterns. Layer-cake, where eyes hop from heading to heading. Spotted, where someone hunts for one word, a number, a link. And commitment, where a person decides it’s worth it and reads every line.6 The job is to make the first two easy, so more people get to the third.

And it’s not only busy people. In the latest US adult skills survey, 28% of adults scored at the lowest literacy levels, up from 19% six years earlier.7 The UK government’s writing guidance cites research on legal documents where 80% of people preferred the version in clear English.8 Plain isn’t dumbing it down. Plain is how you get understood.

The 124% rewrite

Shorter, scannable, less salesy copy more than doubled usability. In that same 1997 study, Nielsen rewrote one page three ways and measured each against the original.1

Usability gain from rewriting the same pageNielsen, 1997
Objective, no hype+27%
Scannable layout+47%
Half the words+58%
All three+124%

Look at the first bar. Just taking out the marketing language, with nothing else changed, made the page 27% more usable. Brand fluff has a measurable cost. That one should sting a little. It stings me.

Here’s a made-up example. You have read the first version a thousand times.

Before · 52 words

At Acme, we are passionate about delivering innovative, best-in-class solutions that empower forward-thinking organizations to reimagine what is possible. Our comprehensive platform brings together a robust suite of powerful tools, designed from the ground up to help teams of every size streamline their workflows and drive meaningful results across the entire business.

After · 24 words

Acme is scheduling software for clinics. Patients book online. Staff see the whole day on one screen. Most clinics are live in 2 weeks.

The second one is less than half the length and it’s the only one that tells you what the company does. Nobody loses the first one in a meeting, though. It offends no one, because it says nothing.

There’s math behind cutting, too. Nielsen found that every extra 100 words on a page buys about 4.4 more seconds of attention.2 That’s enough time to read roughly 18 of them. Add 100 words and you’ve mostly added 82 words of scenery.

Now the machines skim too

A summary is the new first screen. Search engines and assistants read your page so the person doesn’t have to. And people don’t finish the summary either. In one study of Google’s AI Overviews, the median person scrolled through about 30% of the overview itself.9 Pew found that people ended their browsing session 26% of the time after a search with a summary, against 16% without one.10

So your page now has two readers in a hurry. A person scanning for their answer, and a model pulling one sentence to stand in for all of it. Both reward the same thing: a clear statement, early, in plain words.

How we design for it

Headings carry the argument. If someone reads only the headings on a page, they should be able to explain the product. I ask for headings that explain it at a glance, not headings that sound good.

roofrabbit.ai
Roof Rabbit homepage. The headline reads: Quote first. Sell first. Paid first.
Six words, and you know what it does and why you’d want it.Roof Rabbit, built by Wild Places

Lead with the result. Each capability starts with what the customer gets, then how. “Get paid faster” before “integrated invoicing.”

Start with the visitor’s question. For a stadium, the wireframes open with the two things everyone arrives with: what event, what section. The story of the building can wait.

Put the meaning in the first two words. Scanners take in roughly the first 11 characters of a link or a headline.11 “Pricing for teams” works. “Explore our flexible options for pricing” doesn’t.

Use numerals. Write 23, not twenty-three. Digits stop a wandering eye.12

One idea per block. Short paragraphs. A strong grid. Remove the widow. Typography isn’t decoration here. It’s how the eye finds the next thing.

Don’t blame the phone. People understand content on a phone about as well as on a desktop. Hard text just takes longer.13 So the fix is easier text, not less of it.

Fair point: some people do read

Layer it. The buyer who is about to sign reads everything. Legal reads everything. The person comparing you against two competitors reads the pricing page like a contract. I’m not telling you to delete the depth.

The page works in layers. The top layer is for the skim: headings, numbers, one line each. The next layer is for the committed: the detail, the proof, the fine print, one click or one scroll away. Cognitive load stays low for the person passing through, and the person who wants more can get all of it.

We run one test before anything ships. Read only the headings and the bold. If you can’t tell what the company does and what to do next, the page isn’t done.

The ruleClarity over cleverness. Write for the skim. Reward the read.

Sources

  1. Jakob Nielsen, “How Users Read on the Web,” 1997. nngroup.com
  2. Jakob Nielsen, “How Little Do Users Read?” 2008. nngroup.com
  3. Tony Haile, “What You Think You Know About the Web Is Wrong,” Time, 2014. time.com
  4. Chartbeat, global audience insights, Q2 2026. chartbeat.com
  5. Simon Maybin, “Busting the attention span myth,” BBC News, 2017. bbc.com
  6. Nielsen Norman Group, “F-Shaped Pattern of Reading on the Web: Misunderstood, But Still Relevant,” 2017. nngroup.com
  7. National Center for Education Statistics, PIAAC 2023 results, December 2024. ies.ed.gov
  8. GOV.UK, content design guidance on clear language. guidance.publishing.service.gov.uk
  9. Kevin Indig, “The first-ever UX study of Google’s AI Overviews,” Growth Memo, 2025. growth-memo.com
  10. Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results,” July 2025. pewresearch.org
  11. Jakob Nielsen, “First 2 Words: A Signal for the Scanning Eye,” 2009. nngroup.com
  12. Jakob Nielsen, “Show Numbers as Numerals When Writing for Online Readers,” 2007. nngroup.com
  13. Nielsen Norman Group, “Reading Content on Mobile Devices,” 2016. nngroup.com