|
Morkes and Nielsen's Web Usability Studies
This web page summarizes two recent studies which measured the effects of better writing on the performance of users. Better writing improved the usefulness of a non-technical web site by 124%, and a technical site by 159%. The users of better-written web sites were 800% more accurate than the control users. These studies clearly demonstrate the power of good writing.The usability of the test web sites was determined by measuring the following criteria:
Users in the first study (Morkes and Nielsen 1997) were randomly shown one of five versions of a travel-related web site. Each version had the same basic content, but different writing styles. Those writing styles and the corresponding usability scores were:
- Speed: how long it took users to perform a few simple tasks.
- Recall: how many answers they got right during a memory test afterwards.
- Comfort: how testers responded to subjective questions.
- Accuracy: whether their searches retrieved the right information.
Essentially, the benefits of revision were additive. Note particularly that usability scores rose considerably when researchers cut half the words from the document.
- Promotional: the boasting, cajoling language of the advertising world ("Click here! You won't be disappointed!").
Usability Result: 0
(Considered the control, because this kind of writing is prevalent on the Internet).- Objective: roughly the same content as the control version, but without the boasting and cajoling.
Average: 27% better- Scannable: exactly the same text as the control version, but with more subheadings and bulleted lists.
Average: 47% better- Concise: about half as long as the control document, but containing about the same information
Average: 58% better- Combined: an objective, scannable, and concise version.
Average: 124% betterThe second study (Morkes and Nielsen 1998) revised technical documents already being used on-line at Sun Microsystems. Subjects who read the revised (objective, scannable, and concise) version outperformed their control counterparts in four categories.
On average, the usability of the revised site was 159% better than the original.
- Speed: 80% faster (at performing basic tasks)
- Recall: 100% better (based on their answers to a quiz)
- Comfort: 37% higher (based on their responses to subjective questions)
- Accuracy: 809% better (far less likely to give incorrect answers to search questions)
The shorter version was considered more comprehensive than the longer one (Morkes and Nielsen 1998: "Conclusions," paragraph 2). The concise version packed more information into the same space, bringing more items to the user's notice in a shorter time, making the user think that the page was more informative.
The law of diminishing returns predicts that at some point it will no longer be worthwhile for a writer to invest more energy in a particular document, but web pages have a huge potential audience. Even minute penalties or bonuses (in terms of time, accuracy, recall, and comfort, per user, per visit) pile up quickly.
References
- Morkes, John and Jakob Nielsen. 1997. "Concise, SCANNABLE, and Objective: How to Write for the Web." Sun Microsystems. Internet. <http://www.useit.com/papers/webwriting/writing.html>. Last accessed March 8, 1998.
- -----. 1998. "Applying Writing Guidelines to Web Pages." Sun Microsystems. Internet. <http://www.useit.com/papers/webwriting/rewriting.html>. Last accessed March 8, 1998.
|
Last modified . Site designed and maintained by Engineering Communication Centre Staff. |
 |