Media Coverage

New Media Age: Make it easy

So you're building a Web site. It's the online face of your company. It'll have e-commerce. It'll have downloadable catalogues. It's even going to have a community section so people can see how caring and sharing you are. You want it to be easy to use, slickly professional, bang up to date in look and feel. When the designers come back with a prototype, you think it's all of these.

But how do you know? It looks great to you; what about to your users? "A lot of people just seem to go by their instincts, which is a very subjective and fallible measure," says Jakob Nielsen, Web usability guru and author of several books on the subject. Assuming that every consumer has the same taste, Internet experience and intelligence level as your company's marketing department can be fatal.

It's not known exactly how many users give up in despair at not being able to find something on a site, not being able to navigate, not being able to buy what they want to. But research suggests that between 10% and 40% of Web visits end in frustration. "People find so many sites insufferable, they just leave them and never go back," Nielsen shrugs.

At present, companies wanting to test the usability of their sites can do so themselves by setting up a lab and inviting testers in off the street, or have it done for them by one of the dozens of UK companies now offering site auditing. Usability consultants each have their own methodology, but the most common is taking a sample of the site's target audience and getting them to run through certain tasks while under observation.

These processes raise their own questions, however. How do you objectively assess the work of your usability consultants? Are there internationally recognised standards for Web site usability? The International Standards Organisation (ISO) declares it has no work going on in this area, nor plans to address it. The World Wide Web Consortium (W3C) gets a little bit closer with its Web Content Accessibility Guidelines, but these are solely intended to open the Internet to users with disabilities, and don't govern general usability issues. The Internet Engineering Task Force (IETF) prefers to concern itself with the architecture and smooth running of the Internet, rather than design issues.

In short, "there aren't any Web usability standards," says James MacAonghas, head of Internet research at Aqute Research. The closest the Internet has got are sets of companies that have evolved their own standards for user interfaces for their software or their sites. But these aren't endorsed by any independent or officially recognised body.

Are standards needed? Perhaps surprisingly, though, MacAonghas adds that a lack of standards isn't a bad thing: "No-one is losing any sleep over the lack of Web usability standards. Developers muddle along pretty well without them," he feels.

"Although standards are usually a good thing, they would have limited effect since most Web sites aren't professionally built," he says. "And current levels of usability aren't so bad that much time should be spent on standards - they're a nice-to-have rather than a need. Maybe in a few years' time when everything's settled down, but not now." Indeed, even if there were standards, would people stick to them, MacAonghas asks? "Probably not. [Experience shows that] developers aren't very good at adhering to these standards when they are attempted. Partly because usability standards are difficult to define, partly because most site designers like to think of themselves as artists not rule followers, and partly because the pressures of any Web project make it difficult getting anything done on time, let alone complying to standards." Measurable criteria of usability would at least provide a benchmark by which designers could be judged by their clients. "If you have an untested site, you don't really know what your customers are experiencing. You don't know how many potential customers you may be turning away," points out Peter Chaplin, CEO of Netusability, a behavioural marketing company that performs and promotes usability checks.

In the absence of internationally agreed standards, companies must rely on guidelines gleaned from usability experts and set out by usability consultants themselves. Chaplin gives some general principles that he recommends.

"It's not so much about how the site looks, although that is important, but more about being task-oriented and goal-oriented," he explains. "Try to match your visitor's goals. Define some task scenarios that visitors might want to undertake and try them out. It's crucial to have a framework that identifies tasks people might want to achieve using the site." Give testers a series of tasks to perform and then note and compare how long it takes them - or indeed, whether they manage to complete the tasks at all. Netusability espouses a scorecard concept, marking sites on both design and navigation issues and on task completion. The company's software tools help to map out a site as a series of tasks that visitors are likely to perform and lets the designer quantify financially each one's importance, to calculate the value lost through non-completion by visitors.

As an example of poor usability, Chaplin recalls his own recent use of Expedia.com to book two hotel rooms on the same night, one for adults and one for children. This proved impossible, as the site tried to amalgamate the room requests into one room for five people. In the end he had to book twice under different names, so the hotel wouldn't cancel one of the bookings in the belief it was a duplicate. If only the Expedia designers had explored such a simple scenario they would have included it within the site's capabilities.

"If customers can achieve their tasks, then you can achieve your business goals. It's as simple as that," he says.

Red Sherriff, another usability specialist firm, posits its approach as the possible basis for future usability standards. Its method is to run sites through a list of 200 questions, each with a yes or no answer, scoring 1 or 0 accordingly. The questions can also be weighted to give greater importance to certain issues, forming a complete usability index.

Sample questions include: is there a home button? Is there a list or map of stores, branches or offices? Can the shopping basket be viewed prior to paying? Does the site record previous orders and let you start a new list based on those? Companies seeking an introduction to usability before committing to a full-scale audit could begin by investing in a book by Jakob Nielsen. His own site, nngroup.com, embodies his central tenet of design: keep it simple, as demonstrated by his favourite site, Google. The second rule is to do as other people do.

"Users don't have time to learn how to use an individual site," he warns. "They want to learn a few principles and apply them to all sites. So you'll be successful if you follow standards. If you try to do anything different, you'll fail." A simple search box into which the user can type a few keywords is essential. Using only a few subtle colours rather than a riot will help visibility. (Indeed, BT Exact notes that designers should bear in mind that colour blindness affects one in 12 males, so it's best not to rely on colour alone to cue a link.) "Anything fancy fails," states Nielsen.

While keen to emphasise the importance of usability in the development of the Web, Nielsen is pessimistic about the prospect of the Internet industry coming up with standards to govern Web usability. "It's chaos now," he says.

Given the absence of standards at the international level, are there any other organisations that might step into the breach? There are several groups of professionals drawn together by usability issues, and these may be in a position to evolve standards or endorse comprehensive good practice guidelines in the future. "The one everyone knows is the Usability Professionals Association (UPA) based at Loughborough University (upassoc.org), which is very good. In the US, the government group the National Institute of Standards and Technology (NIST) is also good," says Chaplin.

In addition, the British Computer Society has a human computer interaction group (bcs-hci.org.uk), which deals in part with usability issues but doesn't specialise in Web usability. This September, the BCS HCI group will run its annual conference, alongside The European Usability Professionals Association, at South Bank University in London.

Membership of these associations appears to confer some credibility on those claiming to be experts in the usability field. However, membership alone doesn't guarantee true expertise. These organisations don't yet offer the kind of professional accreditation available in other fields, such as the Certified Network Engineer or Certified Lotus Professional in the corporate computing world.

The BCS HCI group is moving towards such accreditation, with a draft document available at its Web site. Under the proposed scheme, professionals would be tested against published criteria on usability consultancy, planning user-centred design, evaluation and testing, with options on additional testing for more specialised competencies, including design training courses and technology transfer for others. Again, these criteria wouldn't be confined to Web site usability, but encompass all aspects of human-computer interface design.

Catriona Campbell of the Usability Company has been working closely with BCS HCI. Contrary to ISO's assertions that it has no Web usability standards, she believes that two existing ISO standards can be brought into play. ISO 13407 defines a methodology for building human-centred design into software, including the need for user testing. ISO 18529 provides guidance on "descriptions of human-centred processes structured for use in process assessment and process improvement".

Campbell says: "The BCS HCI Usability Practitioners Accreditation Scheme will use these standards, and develop new standards and methodologies. I'm hopeful it will come into practice in 2003." Usability-based backing Technology professionals themselves may try to take a leading role in promoting usability, but they risk being labelled as self-serving by clients who see in usability consultants just another hawker touting for a fee. A more effective push towards taking up usability standards and more serious testing should come from venture capitalists, argues Netusability's Chaplin. Backing for any fledgling company should come with strings attached, in the form of usability audits for that company's Web sites, particularly if it's planning to embark on e-commerce activities.

There has been far too little push from VCs in this direction so far, Chaplin believes, which is a glaring omission on their part. Only by usability testing can the backers ensure that they're spending their money wisely. If they wouldn't buy a car without a test drive, why should VCs pour their cash into sites that might break down at the first bend? Yet they do. Perhaps surprisingly, The Usability Company reckons only 5% of UK Web site owners conduct any sort of independent usability testing.

In the end, the approach to usability companies choose may need to depend on the money they have to spend. A full usability audit can run to thousands of pounds, depending on which of the dozens of usability consultancies you choose.

"It's worth paying for a professional audit," says Martin Filz, VP of Red Sherriff, which offers testing starting at £500. He argues that given the amount of cash it takes to set up a decent Web site in the first place, not testing it is a false economy.

In the absence of recognised standards by which to benchmark sites for free, that's the choice: paying a usability consultant or sitting down with a book about usability and a few mates to run through the site, performing set tasks to a stopwatch. The price of the latter will be a few beers - plus all the customers you turn away because they can't use your site.

Back to top

 
© The Usability Company 2007