Effective Hyperlinks:
What We Can Learn from Books

A Rationale
for Navigation

Communication, like most aspects of our lives, is a time-mediated process. Instances of communication, therefore, are sequences of some kind, even if individual events in these sequences, such as a photograph or even a voice quality, may seem largely nonsequential. The sequence may or may not be imposed on the receiving party: it is strongly imposed by a video segment, less so by a well-structured text, and even less by a graphical representation.

Humans describing a trajectory, whether it be a physical journey or a communication sequence, want answers to questions. They want to know where they are. They want to know where they are going to. Possibly, they want to be reminded of where they come from. Effective communication, no matter its nature or its medium, provides answers.

Humans, embarking upon trajectories for a given purpose, usually look for ways to control their trajectory so as to reach their purpose, or reach it faster. To make decisions, they need to know what the options are and what consequences choosing a given option bears. Effective communication, no matter its nature or its medium, allows informed decisions.

Traditional paper documents, especially the longer and maybe more structured ones such as books, carry a surprising amount of navigational information, both by nature and by contents. As physical embodiments of the communication sequence, they provide numerous clues as to where readers are (for example, near the beginning or somewhere in the middle). Numbered headings and, especially, running headlines or footlines further help readers locate their current position in a structure often rendered explicit by a table of contents. This table of contents, possibly together with a document overview at the end of the introduction, allows readers to make choices about where to go. So does an index, along a different logics. Finally, the references to tables and figures, to bibliographical entries, and to other parts of the document are as many “hyperlinks”.

Document navigation, in other words, is by no means an Internet invention. For centuries, carefully designed paper documents have provided both motivation and means for readers to jump to other parts of the page, to other pages of the document, or to other documents altogether. By allowing such a jump at a click of the mouse, computers simply make navigation faster and easier, yet at the risk of disorienting the visitors.

To allow informed decisions, Web pages must tell visitors what they can expect to find by following links. Traditional paper documents, being limited to explicit and largely conventional links, often offer reasonable clues for informed decisions, a noteworthy exception being the footnote. By contrast, computer documents, which can make a hyperlink of any word or graphical element on the page, can render hyperlinks cryptic.

Guidelines for Web Navigation

Traditional paper documents provide an excellent reference model for designing the navigation of a Web site and thus helping visitors know where they are and decide where to go.

Helping visitors know where they are supposes a reference structure. Though virtual documents could easily assume any conceptual structure, a hierarchical or tree-like structure is probably the easiest one for visitors to perceive and memorize, and for Web authors to signpost, usually in two ways.

• A table of contents, usually called a site map for Web documents, provides visitors with a global view of the Web site's structure.

• A hierarchical path on each page, the equivalent of a running headline or footline, indicates to visitors where the page fits in the site's hierarchical structure.

Because visitors lack the clues readers may gain from the physical nature of a paper document, situating a page explicitly in the site's structure is all the more important. In contrast to running headlines or footlines, which typically situate the page within one or maximum two levels, Web pages best display a full hierarchical path, possibly reflected in the page's URL. Similarly, they usefully remind visitors of what site they are visiting, for example by all displaying the sponsor's logo. (This logo could, for example, be a hyperlink to the home page—an emerging convention adopted by more and more Web sites.)

Helping visitors decide where to go supposes to let them know both what links are available to them and what they will find (or, better, what they will gain) by following a link.

Effective links are highly visible and identifiable as links, without drawing undue attention onto themselves. On scrollable pages, main links appear near the top of the page and may be repeated at the bottom. Verbal links differentiate themselves from other words by their appearance or their position on the page (or both). Graphical links are identified as links by their appearance (suggesting a button) or by accompanying text.

Effective links clearly indicate their destination on their own. They require little surrounding text, if any, and no mouse-over pop-up explanation, even if they can benefit from such explanations as a form of effective redundancy (graphical links, for example, benefit from text alternatives). Three examples of ineffective hyperlinks are (1) the “click here” syndrome, conveying no information in the link itself (the explanation is in the surrounding text), (2) hyperlinked words as part of a sentence, providing no clue as to the destination (unless a clear code has been established), and (3) mystery boxes, revealing their destination only when the mouse passes over them (see my article in IEEE Prof. Commun. Soc. Newsletter 44:2 of March/April 2000).

Three Types
of Hyperlinks

Possible destinations for navigational links may be organized in three types, found in paper documents: start there, continue there, and go there and back.

Start there links help visitors locate from the onset the information they seek. They can reflect the site's structure and thus form a partial or full table of contents, possibly repeated on every page or as one of the panels in multi-panel frames. They can also allow random access to pages, regardless of the site's structure, and thus form an index, either limited to keywords (like the index of paper documents) or allowing simple or complex searches on any word or combination of words.

Well-phrased items in a table of contents go a long way towards helping visitors decide where to start. As in paper documents, however, they may not suffice to allow informed decisions. A site map may therefore usefully be complemented by some text orienting visitors on the basis of their purpose (possibly as explicitly as with if you…, then go to… constructs), just like a table of contents in a paper document does not make purpose statements or preview paragraphs superfluous.

Continue there links help visitors explore a site or pursue a train of thoughts. Effective such links clarify their destination by typically referring to pages or parts of pages identified as such in the table of contents (and thus reachable by other routes, too), much like cross-references in paper documents would. Moreover, they typically appear in expendable parts of the text (as in parentheses), or outside of it, possibly grouped after a paragraph.

Go there and back links help visitors understand contents better by providing side information, such as definitions or explanations, not unlike footnotes or bibliographical references in paper documents. To be effective, however, such links must let visitors know what kind of side information to expect. This is the case for bibliographical references, but very seldom for footnotes: you never quite know what to expect.

Poor hyperlinks are much like footnotes. Instead of allowing informed decisions, they tell visitors, “hey, you might be missing something here—but I am not telling you what.” Most frustrating!