A UX Picnic with Donna Spencer: The FAQ of information architecture
As an information architect, Donna Spencer often gets asks a million and one questions about what she does and how she does it. Through the years, some of the same questions kept popping up over and over again, and they’re what Donna calls The FAQ of IA. Interestingly, most of the questions begin with a “Why can’t we just…?”
While some of them sound easy, they’re actually a lot more complicated than what meets the eye (or ear in this case).
Why can’t we just arrange this by audience?
This one seems completely obvious, but there’s a lot of thought and reasoning as to why we don’t arrange things by audience (typically). We have customers, we think of them in particular groups, we create our products around those groups, so why don’t we just use that on the website? It’s not completely out of question to think about it this way — until you work through it.
To make this work you need two things: People need to figure out who the audience is, and your content needs to split neatly across those audiences. Even these sound straightforward until you start digging into them. For example, a lot of universities use a split of an audience scheme with prospective students, current students, staff, and alumni.
It might sound OK to split up your audiences like this, because this is what universities are made up of. However, it’s possible for you to belong to more than one of those audiences at one time. You could be a prospective student for psychology, while you’re currently studying art history, while you’re also tutoring on the side as a staff member.
Sometimes you know which audience you belong to when you perform a particular task, but other times it can be extremely difficult. Audience schemes only really work when the content separates nice and neatly around the audiences. This is usually quite rare — there’s always content that would suit all audiences or none of them.
Why can’t we just arrange it by format?
Let’s say a website has different types of content: books, calculators, brochures, planners, workshops and videos. Some may think it makes sense to just arrange all of this content by the format it’s in. For example, grouping together all of the books and not separating these by topic. However, this means that users will have to look through all the types of content to find exactly what they’re looking for.
Some people also ask “Why can’t we just order alphabetically?” Alphabetical sort order is acceptable in a few situations, such as when you have a known item information seeking task. It’s a lot easier to find stuff in a long list if you know what you want and what to call it, like in an A to Z index.
Picture a TV network’s list of programmes, and say you were trying to find information about Doctor Who. In this case, you’d just scroll down to the D section and find Doctor Who there — this situation is perfectly fine for alphabetical sort order. However, in any other situation, A to Z is a terrible way to sequence your information. In reality, there’s actually no relation to any of the letters in the alphabet, they’re completely arbitrary; it’s just a learned scheme that we know.
Make sure you keep the order in mind, too. When we look at a list of things, we subconsciously think that there’s some order to them, and that there’s some kind of grouping or priority. Our basic assumption is that things at the front of the list are more important than things further down the list. For example, an e-commerce site may have its product categories displayed with the most popular or prioritized product category first.
In usability testing you might sometimes hear people say “I wish it was in alphabetical order”, but the reason they’re saying that isn’t because they wish it was in alphabetical order; the reason is because you’ve asked them to look for something. They have a word in their head and they’re pattern matching around the screen trying to figure it out. They wouldn’t say that to you if they were looking for a product to buy for their mother-in-law for Christmas. They’re not going to say “I wish those were in alphabetical order” then, so watch for that — just because they say it doesn’t mean you should do it.
Why can’t we just do the design now and pour in the IA later?
Depending who you’re working with, this question can come up a lot. However, remember that it doesn’t make sense to add the IA to anything — you start with it, because it matters.
As an example, picture a content model for a conference. A conference has events, and events have presentations. Presentations are put on by speakers, who present at presentations. Presentations also have a time, date, length, and a topic. From this we can figure out how we might be able to construct navigation and how someone might move around the site. We can look at this and figure out that we can link from the event to the presentations, and the event to the speakers, and between the speakers and the presentations, but we can’t create a list of speakers by topic because that doesn’t fit into the content model. We can’t create a list of events by topic because that doesn’t fit into the content model. We start with this information, we don’t pour the IA onto the design later.