Getting the architecture right...

If you ever want to start a fight, put a bunch of creatives in the same room as a bunch of information architects and get them to agree on how a website should work. I've been involved in enough projects to know that sparks fly very easily.

What do we mean by information architecture? Essentially, it means the way people can find information on our website.

In the beginning, the web didn't really have any proper architecture. Then most sites started developing crumb-trails; those things at the top of the page that tell you where you are on the site (eg: Home > Home, Law and Money > Money > Student Money > Student Bank Accounts). The idea being that if someone was looking for an article about students bank accounts, they'd start on the home page, click on the Home, Law and Money link, then the Money, then the Student Money and finally the Student Loans page.

That works fine except for two major flaws. It assumes that the person arrives on the first page and it assumes they think the same way as you. If you've ever done any card-sorting exercises you'll know how difficult it is to put certain topics in to particular boxes. Using the above example, finding out information on student bank accounts could, quite easily, live in the Student Life section of our Work & Study part of the site.

There are plenty of examples - an article on Drunken rows with mates could be as sensibly placed in Sex & Relationships, Health & Wellbeing, Drink & Drugs or Travel & Free Time. As an extreme example, information on what should be on your payslip is in our Workers Rights section but what that information means is in Benefits & Tax in a completely different part of our site.

There are ways to solve this. You could duplicate the content so that it appears in different places. We've done this a few times (Bouncers and the Law appears both here and here). But it's not very easy to maintain - you have each version of the article.

Because we use a content management system, we could ask it simply to reproduce the same article in different places; but this has the affect of diluting our search engine rating because there are two places you can link to.

The method that most experts are now using is removing crumb trails altogether and using metadata (also known as metatagging) and dynamic indexes. So, in the first example if we tagged the student bank accounts with metadata tags "student money" and "student life" then it would automatically be added to both those indexes.

But that's just the start of using metadata effectively. We have related links on our page (both internal links to other pages, and external links to relevant and useful websites). Currently they are manually added to each page; but using metadata we could automate that process. So, in the example about payslips, if we tagged it "income tax" we could show other related articles and external links tagged with "income tax" (so the article about income tax and a link to the HMRC).

And that's where the information architects come in. Because in order for metadata to work effectively you need to have an agreed vocabulary. If one person tags their content "student life" and the other "university life" or "life on campus" then the computer isn't (yet!) clever enough to realise they are the same thing. Equally there are times when the same expression can have two totally different meanings and you need ways of dealing with that.

Metadata can greater help reduce the workload and allow us to spend more time creating great content rather than putting it into boxes, but we need to get it right.