I attended a working session recently to plan out the information architecture of an organization facing profound problems with their existing websites.
The organization's web content is deep and they tried to solve this by having two websites. A recent stakeholder revolt convinced the organization back to the drawing board with the website. This time they were seeking the input of a the various different stakeholder groups (always a great idea).
I decided to participate as I love information architecture (IA). I relished categorizing things since I was a youngster and as a website manager among my favourite duties was planning out (and replanning) the IA of a site.
Here are some tips I gathered from my experience and the session.
The expert introduced us to the pioneering work in this field by Richard Saul Wurman (here's a brief history of IA).
The expert also recommended when preparing to do an IA start by listing all the stakeholders and user groups. A stakeholder and user group may be one in the same, but not always. Stakeholders need to be considered as even if they don't use the website, the final IA still has to be fine by them. It is also important to consider, the expert noted, how a user group changes over time and the resulting changes in information needs. For example, a student changes (hopefully) over time to an alumnus or a prospective customer to an owner.
It is important to know how your users think about your content - how they would categorize, structure, and name in their own words. There are various ways to do this. I have done focus groups and user testing and they are good sources, but I think card sorting is the way to go for preparing for IA. Card sorting is where you ask users to "participants organize topics into categories that make sense to them and they may also help you label these groups" according to Usability.gov.
The IA expert also stressed that it is crucial when labeling the navigation elements "to not worry about being interesting, be clear. The content can be interesting, your navigation should be clear." I recommend using the words your users would use for the labels as much as reasonably possible.
We also discussed at the session on the necessity to make trade-offs and prioritizations. For example, you can't design the IA of a website both for a novice and for an expert user at the same time. They likely won't think about the content and names in the same way the more familiar they get with your topic.
You can consider duplicating content to get around this or other situations, and while this may work it can also make a website more of a quagmire. So thread carefully with duplication as the best solution depends on the unique considerations of each website and its users.
Another challenge we wrestled with is whether to keep users out of content that doesn't apply to them (via separate sites or log-in) or have all your content open to all for transparency. Again, the best solution depends on the individual website.
My biggest tip, however, is invest in a good search engine - the best you can afford. Then customize and maintain it. Great IA and navagation aids can never direct every user all the time to the correct content and some people are search dominant (they go right go the search feature as soon as they arrive). Make sure to configure the search settings with how your users will use it. This means your search engine should support natural language, stem, and fuzzy search (which I believe are standard nowadays). Many search engines will allow for a customized thesaurus. Find out the actual words your uses use to think of your content (via focus groups, surveys, and anecdotal feedback mechanisms). As the essence of search engines is indexing words, it is also a good idea to write your content so that keywords are used.
In going through our organization's various content and functions that the website was required to house, it quickly became overwhelming to have one site that could satisfy (let alone delight) every stakeholder. Doing a website's IA for any organization is something that takes time, careful consideration, and no doubt delicate political wrangling.
It is definitely worth devoting the time, energy, money and machiavellian ploys as truly a good information arcitecture forms the foundation for an effective website.