vibrance
Book a call

Writing for websites: how people read online and what follows for businesses

Most people do not read websites word for word; they skim them. This article summarises research on reading on screen, explains what readability formulas can and cannot do, and derives rules for copy that answers questions, gets found and contains no challengeable promises.

by Flint Richter, SalesPublished on 6 min read

When planning a website, most people think about design and technology first. For visitors, though, it is the text that decides whether they understand what a business offers and whether they call or write. At the same time, text is the part of a website that businesses are most likely to supply themselves. This article summarises what is known about reading on screen and derives rules from it.

Reading on screen is skimming

The early studies

In 1997, Jakob Nielsen and John Morkes investigated in several studies how people read web pages. Their finding was clear: 79 percent of test users skimmed new pages; only 16 percent read word by word [1]. People looked for keywords, headings and lists and only read more closely where they expected something relevant.

In one experiment, the authors rewrote the same content of a tourism website in different ways and measured how well test users coped with it. Compared with the original version, the concise version was 58 percent more usable, the scannable version 47 percent and the objective, promotion-free version 27 percent. Combining all three properties achieved 124 percent higher measured usability [2]. The figures come from a single, small study of a particular type of site and should not be read as a general benchmark. The direction of the effect, however, has been confirmed again and again in later research.

Reading patterns

Eye-tracking studies by the Nielsen Norman Group later described the so-called F-pattern: readers first take in the top lines, then a shorter section further down, and scan the rest along the left edge. An update from 2017 stresses that the F-pattern is not a goal but a symptom of text without structure, and that it also occurs on smartphones. Well-structured pages with meaningful subheadings lead to other, more efficient patterns, such as deliberately jumping from heading to heading [3].

What readability formulas can do

Flesch and Amstad

The best-known formula for measuring readability was developed by Rudolf Flesch in 1948. From the average sentence length and the average number of syllables per word, it calculates a score between roughly 0 and 100, where higher scores mean text that is easier to read [4]. Toni Amstad adapted it for German in 1978, because German words are longer on average:

Readability = 180 − average sentence length − 58.5 × average syllables per word

Scores above 60 are considered easy to understand for a broad audience, scores below 30 difficult [5].

Limits

Readability formulas only measure surface features. Text made of short sentences and short words can still be incomprehensible if it contains unfamiliar jargon, jumps between thoughts or fails to answer the reader’s question. Conversely, specialist text for specialists may be appropriately difficult. The formulas are useful as a rough indication of where sentences have become too long; they are no measure of quality.

WCAG contains a related recommendation in criterion 3.1.5: if text requires reading ability more advanced than lower secondary education level, a more easily understood supplement should be offered. The criterion belongs to the highest level, AAA, and is therefore usually not mandatory, but it describes a sensible goal [9].

Copy that answers questions

Ginny Redish describes web copy as a conversation in which visitors ask a question and the page answers it [7]. For businesses, these questions are usually predictable: what exactly do you offer? For whom? Where? Roughly what does it cost? How long does it take? How do I reach you? A page that answers these questions in the order in which they are asked is useful to people and understandable to search engines.

Google’s recommendations describe exactly this approach: content should be written primarily for people, come from first-hand experience or expertise and leave visitors feeling they have achieved their goal. Google views content written mainly to rank in search engines critically [8].

Rules for practice

Structure

  1. Most important first. Every page and every section starts with the core message; details follow. Anyone who reads only the first lines should have learned the essentials.
  2. Meaningful headings. “Replacing your heating system in Fürth: process and grants” is more helpful than “Our services”. Anyone who reads only the headings should know what the page is about [3].
  3. Short paragraphs, one point per paragraph. On screen, long blocks of text are off-putting.
  4. Lists for enumerations, tables for comparisons, for example of packages or opening hours.
  5. One page per service, if a business offers several clearly distinct services. That way each page can answer one question completely.

Language

  1. Cut. Steve Krug recommends getting rid of half the words on each page, then half of what is left; what he mainly means is introductory chatter and self-praise [6].
  2. Specific rather than general. “Appointments within two weeks” says more than “fast scheduling”; “fixed price after an inspection” more than “fair prices”.
  3. The customers’ language. Use jargon only if the audience uses it; otherwise explain.
  4. Active and direct. “We will call you back within one working day” rather than “A callback will take place promptly”.
  5. Consistent form of address. Formal or informal, but throughout.

Call to action

Every page should say what can happen next: call, book an appointment, send an enquiry. A single, clearly named main action per page is more effective than several equally weighted ones.

Business web copy is a commercial practice and subject to competition law. Under § 5 of the German UWG, misleading statements are prohibited, for example about characteristics, prices or qualifications [10]. In practice, this means:

  • No claims to be the best without evidence. “The best hairdresser in town” or “market leader” can be challenged if they cannot be shown to be true.
  • Figures and promises must be accurate. Anyone promising “a reply within 24 hours” has to keep to it.
  • Reviews and awards only if genuine and with a source.
  • Verifiable statements instead of superlatives. Specific, verifiable information is not only legally safer but, according to the studies above, also more effective: the objective version of the text performed better than the promotional one [2].

Checklist for an existing page

  1. Does the first paragraph answer what is offered, for whom and where?
  2. Do the subheadings, read on their own, make a sensible summary?
  3. Are there paragraphs of more than four or five sentences that can be split?
  4. Are there specific details instead of general adjectives?
  5. Is the next step clear on every page?
  6. Are all promises and figures verifiable and up to date?

Conclusion

People read websites selectively. Copy that is concise, well structured and objective is demonstrably used better than lengthy promotional prose. For businesses, this means answering their customers’ questions in the order they ask them, being specific and doing without unsupported superlatives. That makes copy clearer, easier to find and less open to legal challenge at the same time.

Frequently asked questions

Do people read websites word for word?
Mostly not. In studies from 1997, 79 percent of test users skimmed new pages; only 16 percent read word by word.
What is the Flesch formula for German text?
According to Amstad's adaptation: 180 minus average sentence length minus 58.5 times average syllables per word. Scores above 60 are considered easy to understand, but the formula only measures surface features.
May I claim to be "the best" on my website?
Only if it can be shown to be true. Misleading statements are prohibited under § 5 of the German UWG; specific, verifiable statements are safer and also performed better in studies.

Sources

The numbers in square brackets in the text refer to this list. Online sources last checked in November 2026.

  1. [1]Nielsen, J. (1997). How Users Read on the Web. Nielsen Norman Group. www.nngroup.com/articles/how-users-read-on-the-web/
  2. [2]Morkes, J. & Nielsen, J. (1997). Concise, SCANNABLE, and Objective: How to Write for the Web. Nielsen Norman Group. www.nngroup.com/articles/concise-scannable-and-objective-how-to-write-for-the-web/
  3. [3]Pernice, K. (2017). F-Shaped Pattern of Reading on the Web: Misunderstood, But Still Relevant (Even on Mobile). Nielsen Norman Group. www.nngroup.com/articles/f-shaped-pattern-reading-web-content/
  4. [4]Flesch, R. (1948). A new readability yardstick. Journal of Applied Psychology, 32(3), 221–233. doi.org/10.1037/h0057532
  5. [5]Amstad, T. (1978). Wie verständlich sind unsere Zeitungen? Doctoral thesis, University of Zurich.
  6. [6]Krug, S. (2014). Don't Make Me Think, Revisited: A Common Sense Approach to Web Usability (3rd ed.). New Riders.
  7. [7]Redish, J. (2012). Letting Go of the Words: Writing Web Content that Works (2nd ed.). Morgan Kaufmann.
  8. [8]Google Search Central: Creating helpful, reliable, people-first content. developers.google.com/search/docs/fundamentals/creating-helpful-content
  9. [9]W3C, Web Content Accessibility Guidelines (WCAG) 2.2, Success Criterion 3.1.5 Reading Level. www.w3.org/TR/WCAG22/#reading-level
  10. [10]German Act against Unfair Competition (UWG), § 5 Misleading commercial practices. www.gesetze-im-internet.de/uwg_2004/__5.html

About the author

Flint Richter, Sales

Guides businesses from the first conversation to the quote and writes here about visibility, copy and the commercial side of a website.

More insights

Questions about your own site?

In a free thirty-minute call we look at it together.