For each font we show users' average reading speed in words per minute (WPM). (Impossible, because there are probably thousands of good fonts, not even considering the even larger number of bad fonts which we sadly do see employed on websites from time to time.) These 16 typefaces were measured in the study. And it would have been bigger again in a hypothetical (but impossible) experiment that measured people reading texts in all available fonts with sufficient legibility to be considered for practical business design. If reading speeds had been measured for all 16 fonts for each user, the difference between the fastest and slowest font would have been likely even bigger. This is a huge difference, considering that each user was only measured on 5 fonts. On average, any given participant read 35% faster in his or her fastest font (314 WPM) compared to that same person’s slowest font (232 WPM, on average). The first of these findings is not always true for web content, and the second is rarely true for web users. High comprehension indicates that (a) the content was easy relative to the users’ reading skills, and (b) people read the texts carefully. However, this last metric was not very interesting, as the score was very close to 90% in all conditions.

Comprehension score: percent of questions answered correctly after the user finished reading a passage.

Reading speed in words per minute (WPM).

Subjective user preference: which font did each person like the best?.

Long-form reading, for example reading of a book.“Interlude” reading: reading of short passages of prose like the ones used in the study and found on most websites.Glanceable reading of a few words, as found in notifications and on tiny screens like smartwatches.The authors note that there are 3 types of reading, of which they only studied one: Even so, I still think the findings about the relationship between fonts and reading speed are of interest. This difference in reading behavior between web users and study participants does raise the question whether findings would be different under more realistic web-usage conditions. (Complete reading was encouraged by asking users a few comprehension questions after they had finished reading each passage.) In contrast, study participants did read the entire texts in a linear fashion. Thus, in real life, users only read 28% of the words on a webpage, or about 166 words. While navigating the Internet, users mostly scan pages for useful information, as opposed to reading all the text word-for-word. However, each user only read texts in 5 of these fonts during the main reading-effectiveness test. The texts were shown in 16 different fonts, with appropriate experimental controls for things like order effects.įonts included classic typefaces (Times, Helvetica, Garamond), typefaces designed for computer use (Calibri, Arial), and typefaces specifically designed for legibility (Noto Sans, Montserrat) no “wild” fonts (think ) were included. (This article is written at a 12th grade reading level, but it targets professionals, not the general public.)

The test stimuli were at an approximate 8th grade reading level, which matches our recommendation for web content targeted at a broad consumer audience. The participants were asked to read several short passages of text each passage had 300–500 words (by comparison, this article contains 2,623 words and the average web page contains 593 words). Shaun Wallace from Adobe and colleagues conducted a reading-speed study with 352 participants. A large new study of the best fonts for online reading is ultimately disappointing, because it doesn’t answer the most burning question: what font should you use for your website? But it still provides many intriguing findings, including the striking conclusion that there is no single answer to this question.

Fonts online