While a website’s visual design is how your company looks online, the site’s writing is how your company sounds online. Tone of voice and great content are crucial for communicating on the Internet. Nevertheless, the best copywriting is for nothing if users don’t read the text.

As with other areas of user experience, content has to survive a cost–benefit analysis on the part of users:

  • Cost: how much hassle and pain do I have to suffer on this website?
  • Benefits: What’s in it for me, what will I gain if I read this information?

Definitely provide benefits to users. Why else have a website? But you also need to reduce the barriers to using it (i.e., lower the cost). For online copy, the barriers to use fall into 3 categories: legibility, readability, and comprehension, each of which is defined and discussed below.

Legibility

Definition: Legibility is the lowest-level consideration in content usability: it’s whether people are able to see, distinguish, and recognize the characters and words in your text. Legibility is thus mainly determined by visual design, specifically typography.

The main guidelines to ensure legibility are:

  • Use a reasonably large default font size and allow users to change the font size. Tiny text dooms legibility — and remember that what counts as “tiny” differs across people, depending on their visual acuity, which sadly declines with age. Old users need bigger text, but even young users appreciate not having to squint to read the text. (Teens, in particular, often have ghastly posture and don’t sit straight in front of their computer.)
  • Have high contrast between characters and background. Preferably, employ a plain background instead of a busy or textured one, since the latter interferes with the recognition of the fine details in the letterforms.
  • Use a clean typeface. With today’s high-resolution monitors, serif type is fine, but strangely shaped fonts (e.g., emulating handwriting or gothic style) have reduced legibility.

Testing Legibility

If you follow the guidelines above, as well as other best practices for clean typography, you’ll be unlikely to need special tests for legibility. In regular user testing you can watch out for cases where users lean in toward the screen or say something like “it’s hard to make out the text here.” If you notice such behaviors or comments, you can worry and perform special legibility tests, but otherwise you’re probably safe.

The main way to test legibility is to measure reading speed in words per minute for a sample of users, as they read some simple text. Because people read at drastically varying speeds, this is best done as a within-subjects test, where the same test participants test multiple systems. If users are, on average, say, 20% slower when reading from your design than when reading from a reference design, then you know that your site has poor legibility. (See our course on Measuring User Experience for more on within-subjects vs. between-subjects study designs.)

As an example, see the test we did in 2010 comparing reading speeds on iPad, Kindle, and printed books. (But don’t use the old results to decide what tablet to buy today, since both products now have better screens and better legibility.)

Readability

Definition: Readability measures the complexity of the words and sentence structure in a piece of content. The assumption behind this metric is that complex sentences are harder to parse and read than simpler ones. It’s usually reported as the reading level (stated as years of formal education) needed to easily read the text. For example, a 12th grade reading level means that somebody with a good high-school diploma will be able to read the text without difficulty.

The main guidelines to ensure readability are:

  • Use plainspoken words, the shorter the better. Avoid fancy words and made-up terms.
  • Use short sentences. Avoid convoluted sentence structures, especially compound sentences with many subordinate clauses and conjunctions that put a strain on users’ short-term memory, which is a notorioriously weak point. (That last sentence is already pushing it a little :-)
  • Mainly write in the active voice. (Passive voice can be used in rare circumstances when it allows frontloading of key terms.)
  • Aim at an 8th-grade reading level if targeting a broad consumer audience.
  • If writing for an educated or specialized B2B audience, still target a reading level several steps below the audience’s formal-education level. A 12th-grade reading level is often a good target to make text easy for readers with college degrees. (The article you’re reading right now has an 11th-grade reading level, and I hope you find it readable.)
  • Remember that writing at an Nth-grade reading level is a different matter than writing for students who’re currently attending school in that grade. There are special design guidelines for targeting young kids, teenagers, and millennials, and that’s not what we want on a mainstream website for grownup users. We simply want words and sentences that correspond to that level of educational achievement, but written in a mature tone of voice.

Testing Readability

Readability is usually scored by computer. There are many readability calculators on the web and a common formula is built into Microsoft Word: in Word 2013, the Flesch-Kincaid reading level is available under Review > Spelling & Grammar. (In some versions of Word, readability statistics must first be turned on under File > Options > Proofing.)

We actually have a slight preference for a different readability formula called the Fry score, but it doesn’t really matter which formula you use. While there are many readability formulas, they all focus mainly on the length of the words and the length of the sentences. Longer words are usually less common in a language, and longer sentences are usually harder to parse, so these types of metrics make sense. But remember that these counts are only an approximation of what really matters: how hard it is for your readers to make sense of the words and sentences that you use.

Reporting the reading level in terms of years of education is also only an approximation. When we say that a reading level of 12 corresponds to a high-school diploma, that assumes that the high-school graduate was diligent enough to study and master all the grade-level materials and subjects. Many people get through high school despite having poor reading skills, and they would struggle with a text that scored at a 12th-grade reading level.

Comprehension

Definition: Comprehension measures whether a user can understand the intended meaning of a text and can draw the correct conclusions from the text. In the case of instructional or action-oriented content, we also want users to be able to perform the intended actions after reading the text.

The main comprehension guidelines are:

  • Use user-centric language; terms familiar to your audience facilitate comprehension.
  • If targeting a specialized audience (e.g., for a B2B, scientific, or enthusiast/fan/hobbyist site), use the specialized terminology of that field, even if some of these words are difficult for a broad consumer audience, and thus lower your readability score.
  • Use an inverted-pyramid writing style: start with the conclusion or an overview of the main point. People relate better to subsidiary points when they already know the basics.
  • Minimize cognitive load by building on existing mental models and reducing the need for users to remember things from one part of text to another.
  • Pictures or conceptual diagrams can sometimes explain things better than reams of words.
  • Be brief. If you say less, people are more likely to make the effort to understand what you do say.
  • If mobile users are important to you, be even briefer and simplify even more. The smaller viewport hurts comprehension because users can see little context at a glance and can’t easily refer back to previously read information.

Testing Comprehension

The main way to test comprehension of web content is standard user testing, sometimes with small modifications for content testing. The only way we can tell whether people draw the right conclusions from our copy is to have them perform realistic tasks with the site and observe how they interpret the site’s information as they encounter it.

After study participants have performed their test tasks, we can further measure comprehension by giving them various forms of exams. These can range from a simple memory test to more complex questions that truly probe people’s understanding of the material. In one case study, we rewrote web pages about a complex B2B product, and after using the new version of the site, users remembered 65% of the product characteristics we tested for — almost double the 33% of the material similar test users remembered after using the original content.

You can also run a Cloze test on individual pieces of content that are particularly important or which have caused problems in user testing. The Cloze test provides a nice score, so you can rerun the test after a rewrite to see if the text has improved enough.

Content Usability: Beyond Legibility, Readability, and Comprehension

You’d think that the 3 criteria discussed in this article would be enough to ensure great web content. Not so. They are necessary, but not sufficient.

Just because people can read and understand your content doesn’t mean that they will do so. On the average web-page visit, users read only 28% of the words. There’s so enormously much information available on the web that people usually scan instead of read.

Still, users sometimes do read web content, particularly when it includes information of interest to them. The key point is to grab users quickly, and get them interested before they decide to leave, which they often do right away. Headlines are particularly important for fast communication, and the first few words are even more important, given users’ tendency to scan.

Thus, in addition to good legibility, readability, and comprehension, do the following:

  • Write for how users read on websites: clear headlines, scannable layout.
  • Focus on information of interest to users, not on the things you want to promote.
  • Communicate immediately at the top of the page that your content is indeed interesting and useful to users.

Our full-day course on Writing Compelling Digital Copy has hands-on coverage of the issues raised here.