August 2003

Is your website a usable tool, or a designer's dalliance?

So what makes a website work? How can you measure what it is intended to deliver, and can you measure how well it delivers? User-testing is an often neglected component of website development, and yet it is often the most crucial. Virtually all other forms of communication (particularly print-based) have had years of extensive and rigorous testing to determine what constitutes useful and meaningful communication. Little has been done in the domain of websites (and none that can be regarded as definitive work by which all websites should conform), and the field is still evolving.

But without stepping into the shoes of those who evaluate performance attributes and user-evaluation of websites, site stewards (the designers, the managers, the authors, and the creators) can sit down and work out how to best tackle the issue before any building - or renovation - takes place. This preliminary 'foundation work' is required for any website - new or old - before any serious evaluation can occur. Key 'foundation work' issues that should be considered are as follows:

Content is king (but avoid chunking)

A site with poor content is of no use to anyone even if all else is perfect. Content for the web is different, and is viewed differently, to that of print-based communication. People read on-screen material with distraction, and often with impatience. Content which might seem blindingly obvious is often missed entirely through the frustration of wanting information, and wanting it NOW! Documents that work in print invariably fail on the web, partly because less content is able to be viewed all at once on-screen, compared with in print. Thus, the end-user often misses the context, as well as some of the content.

Also, the grouping of information presented in a printed document (in the form of pages, or even chapters, or volumes) is often entirely missed via the web. 'Grouping' of information, even with hyperlinks, is not easily comprehended by the user. Here is an example: if I were to present to you ALIA's Annual report 2001 as a printed document, you would absorb, by the look and feel of the document, the simple fact that it is a dozen or so pages of information, and a quick flick through the pages would tell you that since much of it is tabled material, that it would not take too long to digest. As everyone knows, the web is different. A webpage announcing the contents of the same document gives no clue as to how much time is required to read it, or how much involvement is likely to be required to digest it. Even if the contents page breaks the document into 20 links, it cannot describe how much text is at the end of each link (in other words, no indication of how granular the successive pages might be).

Good web design and construction dictates that users are given as much contextual information as possible, and that each page presents a single, well-defined concept or a single area of focus. It is often difficult to avoid 'chunking' concepts together on a page, but for comprehension, pages should remain 'unchunked' and focussed.

Usability and accessibility

A recent survey of the home pages of nearly 700 federal government websites have shown that none had valid HTML, and only 36 were able to pass any form of accessibility test. Whilst this might seem appalling, it is clear that many websites (and not just government sites) pay little attention to accessibility by users. However, if a site is coded with valid HTML, then the chances of it being accessible are dramatically improved. Where does one determine HTML validity? The W3C's validation service at http://validator.w3.org/ is the place to begin. If the page assessed uses valid code, then accessibility is improved.

'Usability' does not depend on good code alone. A site must have clear navigation, clear content, decent metadata, colour independence, and must support low bandwidth, low literacy, low technology demands, and be useful to both external and internal search engines, for it to be considered a site of excellence.

Clear navigation requires a well-constructed information architecture to underpin the site. Visitors must be able to navigate without an understanding of the organisational hierarchy: the key is to offer the visitor a comprehensible site structure with plenty of visual cues as to their location and context. 'Breadcrumb trails' come to the rescue here, allowing visitors to move further back up the information tree to a place where they can get a better view of the forest.

Colour independence is becoming much more relevant lately. The predilection of site designers to style links with omitted underlining is absolutely baffling many users (and me, too, frankly). This is further compounded by the use of colour for non-linked text. Many users have trouble determining what constitutes a link and what simply constitutes a fancy heading. Bizarre as it might seem, there are people who believe that underlining of links actually decreases usability! Some site designers have gone to incredible lengths to make links more obvious, having first rendered them almost invisible by styling them without underlining. The use of iconic arrows and other inline images to represent links is akin to putting a sign on a door stating that 'This is a door', having first painted it to make it appear to blend into the surrounding walls.

Low bandwidth, and low technology demands are proving to sort good sites from bad quickly enough these days. Many sites are reverting back to a simpler approach to both design and delivery, with less bells and whistles (which are often seem as needless gimmickry), thus making life easier for those with low bandwidth and/or low technology. Surprisingly, there are still designers pushing the envelope, possibly out of a desire to impress a client who is unaware of what impact such sites might have on end-users. With much of Australia - and elsewhere - still on a dial-up connection (and a rusty one at that), designers would be better off learning how to deliver more, with less. It actually works. The notion, though, of making a 'text-only' site is offensive to many: treating a select group of users (usually disadvantaged) to text-only material is subtracting from the user experience, not enhancing it. It is possible to make a site a rich experience for all users, without having to resort to the 'text-only' cop-out.

Low literacy is a challenging area too. Descending into a mass of bullet-points is no solution either. Content should be made applicable for the intended visitors, many of whom might be of lower literacy levels than expected, or with English as a second language.

Clear navigation, site maps, linked tables of contents, clear and simple content, and clear page structure assists in making the site easier to use, comprehend, and absorb. This is especially relevant since many users do not actually 'read' content, but instead 'skim' the information, looking for items of relevance to their purpose.

Search engines, alas, play a vital part in all of this, too. Internal search engines must be capable of returning useful information on a range of searches. Useful information includes the titles of the documents, the relevance of each document in relation to the search string, the context of the search string in the surrounding text, the URL, and the date and size of the documents found. External search engines (such as Google and Yahoo) should be taken into account by building pages with search engine optimisation (SEO), right from the beginning. This is often hard, partly because SEO is still rather under-developed, but also because site design and SEO do not always make good bedfellows. So why the 'alas'? Search engines are lazy librarians. The information is out there, and waiting to be found. However, few people can articulate what they are looking for in a way that makes sense to almost all current search engines. Serendipity is par for the course here (perhaps this is not so bad after all), and until search engines become fully-automated librarians, then searchers are unlikely to get any wiser when looking for answers.

Overall, efficiency wins

Sites need to be built with more than a cursory consideration of development and maintenance. Often overlooked entirely, this aspect of site construction can make or break the site as it grows larger. The more that structure can be divorced from presentation (either through stylesheets, database and content management systems, and the like), the more the site can be manipulated to suit changing needs or styles.

The less coding that is required to put content in place, the better for all concerned. ALIAnet is a good example of this aspect of site management. Prior to version 5, the simple act of moving a collection of pages from one directory to another required massive recoding. Now, entire directories can be relocated and re-purposed in seconds, with obvious benefits for those responsible for coding, and to those who submit material for mark-up. Over time, the deployment of stylesheets and good coding reduces the amount of resources required to maintain and build the site. Modifications can be made more quickly, and ongoing maintenance costs can either be reduced or diverted to other projects.

So when should all this take place?

All sites benefit from being reviewed frequently. All aspects of a site's operation should be considered, with the ultimate aim of making it work better and more efficiently for all constituents. It can be a difficult operation. The latest revision of ALIAnet could best be described as that of changing the wings of a Jumbo Jet flying at 35 000 feet with a full load of passengers. The process of removing the wings, building new ones, and bolting them back into place should be treated with careful and well-planned consideration. A Jumbo Jet can only sustain a certain percentage altitude loss before passengers become very restless, and if the wings are re-affixed in the wrong location, the plane might not recover from the dive sufficiently to carry on to its destination. Planning, research, and a full engagement with the passengers makes the journey as rewarding as any true adventure can be.