Why directories make finding the best information easier

Perversely, the more developed the commercial web becomes, the more difficult it is for site owners to ensure that their visitors will find what they want them to find there. Improved data classification provides one realisable solution.

Featured sites

Ten years after the commercial worldwide web was born, it is in chaos. This is not a criticism – as with the largest cities in the world, chaos and dynamism have gone hand in hand. But as with those cities too, there comes a point where the lack of order becomes a problem. The original web pioneer, Sir Tim Berners-Lee, is trying to bring a degree of order with initiatives such as the Semantic Web (see www.w3.org if you need to know more), but those of us who have a more modest role should be thinking about tidying up our neighbourhoods.
Let’s drop the metaphor. What’s at issue is the difficulty of finding what you want on the web and – more important if you are a provider of information – making sure that visitors to your site will find what you want them to find.
There are two broad solutions. One is to organise the data more efficiently – Yahoo! has been the big guy here, using editors to classify every site and fit it into a neat directory. The other is to work on making search engines more efficient. The search approach is winning, simply because the web is outpacing the classifiers. The Open Directory, a project that tried to tackle this by getting us all to classify one another’s sites, has failed so far to live up to early hopes.

Directory approach shows the way

So, search engines have won? In most places yes; or more specifically, Google is winning. But do you always find what you want there? You do not.
Directories can, if well done, pretty much guarantee that you will get to the best information, and there is a place where the search for better classification has not been abandoned. It is called the public sector. The British Library has a website that contains the images of almost 100,000 items, including newspapers articles, maps, pictures and sheet music. Go to the advanced search, select ‘subject search’, and you will be given a choice, then another, then another – for example: People & Society, Food and Drink, Tea. Despite being called a search, this is actually a directory. It looks simple, but it is the result of much labour. The items have been painstakingly classified by the library’s curators to ensure that academics and the general public will find what they want, using the terms that come first into their heads (remembering that academics don’t use the same words as normal people).
David Macken, managing director of System Associates, which built the British Library’s content management system, points out that libraries are so used to classifying that it would be surprising if they did not offer such a service. Museums are similar and, he says, the public sector as a whole is hanging tenaciously to the idea that structured data is easy-to-use data. In the UK a lot of work has been done to ensure that local and national sites work in the same way, enforcing the use of standardised metadata, and even using the same URL structure across sites. Intriguingly, we may eventually be able to go to Direct Gov, and click through to a particular subject area before selecting the district.

Commercial potential in shared classification

It takes a struggle to see why this would be particularly useful, because most people only need to know about their own area. But it could be invaluable in the commercial world. Large groups use their websites to communicate with different target audiences. They should compete fiercely in the quality of information they provide, but do they really need to compete on how it is arranged?
Think about the needs of your visitors – be they customers, investors, would-be employees, journalists, NGOs or whoever. Do they want to learn a different set of classifications in every site? In some areas, such as investor relations, analysts need the same set of information from each company so that they can compare it as they are used to doing in financial reports. Why not use a standard and agreed set of classifications, with the same sections and sub-sections used on every site? There is nothing some people like so much as a good standard-setting committee; here surely, is a place where one is needed. While there is work going on creating XML standards for financial reporting, this is about something much simpler: what links you use, and what the labels would say.
This leads on to another possibility. Would it be possible to have an investor (or media or careers) portal, so that users could pull up and compare equivalent information from any number of companies? It would need a lot of work but yes, it could done. Does this mean the individuality of the site would be sacrificed? Well, there is no doubt that incomprehensible navigation is damaging some companies commercially, but if the basics are standardised, hard-pressed web people could spend more time thinking up genuinely different ways to differentiate their organisations.

What you can do about it

Nice idea, but you may feel it is not something you have too much control over in your own organisation (and anyway, like most nice ideas, we will probably see many blue moons before it comes to fruition). So what can you do? What about taking a page out of the British Library’s book by trying to create a set of classifications, into which every piece of content would fit. If the library can keep the directory to three levels with its volume of data, so can you. Use your curators (i.e. managers) to both create and populate it, ensuring that they think not in terms of their own departments, but in those of their visitors. Keep the words simple, keep them descriptive, and you can’t go too far wrong.
How can you do this when you are hopelessly under-resourced anyway? Well, how much business, investor interest or high quality jobseekers are you losing simply because people do not find what you want them to find? Put like that, how can your company not make the resources available? Worth trying, anyway.

First published 12 January, 2005
< Back to Commentaries