Journal

Site Research

PC Pro logo Posted: 1st September 1999 | Filed under: Press Articles, Technical
Author: Paul Ockenden
First Appeared in PC Pro 1999

Would you buy a house without first having a full structural survey done? Probably not - unfortunately Paul once did exactly this and he still has the dry rot to prove it. Would a food manufacturing company bring a new product to market without first testing it to see whether people actually like the taste or not? Definitely not. So would a company launch a big expensive new Web site without first testing it on a few users from the target audience? Yes - they would and they do, and unfortunately it happens all of the time.

This reluctance to properly test a Web site prior to launch appears to be traceable to two distinct causes. First, corporate or brand Web sites are usually a co-operation between the marketing and IT departments, and while the marketing people will be totally au fait with user research, it's typically alien to the IT people. The second problem is a more general one, about the way such research is actually conducted - nobody has a good word to say about focus group testing these days. Politics (and in particular 'new' politics) seems to have tarnished this most useful of tools.

This is a shame, as the same tools and techniques that turned most of today's politics into something bland and predictable can also be used to make your Web site exciting and stimulating - if applied correctly. Paul's company, CST, has undertaken user research prior to the construction of a number of large Web sites, and has found most of the feedback to be invaluable. We now recommend this for all big sites, especially those in the £100,000-plus price bracket.

So how should you go about researching your site? We normally like to start at the earliest stages of a project. The site will probably have a theme, a range of possible contents, a few possible names, and some very broad outline design ideas. It'll be tempting to simply go out and ask a few people what they think of these, and indeed many Web agencies and their clients will try to do just that. But there are experts out there who will almost certainly do a better job. Quality of research is critically important. Thinking back to our house analogy, would you buy the Reader's Digest DIY manual and try to survey the house yourself, or would you pay a proper, qualified surveyor to do it for you?

When looking for a suitable expert you'll find that there are the traditional large research bodies, such as NOP (www.nop.co.uk), and a group of smaller and more focused companies like the Internet Research Company (www.internetresearch.co.uk). Both ends of this spectrum have their own merits, so have a chat with a selection of companies to see which best suits your particular project. What we're talking about here is 'qualitative' research. Some of these organisations will also be involved in 'quantitative' research, which is also important but isn't the same thing at all. Qualitative research looks at peoples' likes and dislikes, attitudes, feelings and opinions, whereas quantitative research is numerical - how many people in the UK have access to the Internet, what proportion of Net surfers are under 25 years old, and so on. What the researchers will typically do is sit down with several groups of users and find out exactly what they think of your site content, design and ideas. The group will be led by an experienced 'moderator' whose task will be to keep the discussion focused and stop things spinning off on to some obscure tangent.

These groups should generally be recruited from among the target audience for your site, but it's always a good idea to ensure that you have some representation from right across the spectrum: young and old, male and female, experienced Net users and complete beginners, ethnic diversity and so on. For consumer sites we've found we get good results from family-based groups, where the researchers will visit a family for an evening, show them your site concepts, show them some competitor's sites, and then try to get a meaningful discussion going. It's amazing what can come out of such sessions; not merely comments about the proposed design and navigation of the site, but more fundamental things like whether the consumers felt that a snack brand, say, could legitimately be seen producing a Web site about, say, a particular sport. They might feel this brand has some authority on the subject of cheesy puffs, but none at all when it comes to football or snowboarding.

While your research is being conducted, do make sure that you sit in on a session or two, but just remember to keep quiet. As some geeky ten-year-old trashes your design or navigation system you'll probably feel the need to justify yourself by explaining why you did it that way, but you can't. Remember, you won't be able to explain yourself to every user of your site either - so if the little darling hates your menus, then you had better stop talking and start listening.