Card sorting is a powerful, hands-on tool that we at Stanford Web Services use for helping content creators iron out either the information architecture of their site (meaning the big buckets of their navigation) or to develop categories for their content.
Recently, we used card sorting to develop a secondary sidebar navigation of "Related Content" that crossed the main navigation of a website, and these are my takeaways.
Card sorting is really great for helping people wrap their brains around content using design thinking. By creating cards, each with a single term, category, or page name written on them, we can sort the content of a site, figuring out what things belong together and how best to group the content in logical way.
For a standard card sort to find information architecture (IA), we usually make cards for each known page or chunk of content and then group by subject. For this kind of card sort first instincts work well. But what do we do if already have the primary IA figured out and we're trying to find a secondary way to sort our content? We have to change up our thinking a bit.
Prepping cards for the sort: Include full navigational path
The first thing to do before a card sort is prep the cards. In a standard IA page sort, we include the page title and maybe a little additional information for context. But in our case, where we are trying to determine a secondary way to sort content that is different than the navigation scheme, we needed to know more than just the content itself.
I wanted to be able to see quickly if we were grouping pages that were already grouped by our primary navigation, so I printed the full navigational path with each level on a new line. I did not include cards for the top level navigation pages, since that level of page would not feature our sidebar block of "Related Content."
Where "Curriculum" is the page name and the terms above are the navigation path to get there.
Trick of the trade: cards not stickies
Nothing will hamper a card sort more than the cards sticking to the table! Be sure to use actual cards or pieces of paper not post-its for making your cards.
Getting away from subject sorts
As we started on our card sort exercise, I noticed that we instinctively were grouping content into their navigational structure. Everything in the Connect section got put in a pile with the label Connect.
I thought, "Hmmm, we seem to just be remaking our navigation. How do we break out of that tendency?"
Indeed, card sorting is often used to develop primary navigation: grouping content into the most logical buckets and then building menu structures based on those buckets. We had a natural prediliction for grouping by subject. So we needed a different way of thinking.
User-centered thinking: Grouping by process, circumstance, and role
The deeper we dived into the card sort the more we started to think out of the box. We started grouping based on facets other than subject: things like process, circumstance, and role. The key we found was to consider user roles and the situations that could bring people to our website.
I found myself saying, "If I were on this page, what might have brought me here? And what else could I want to know about elsewhere in the site?"
Below are a few hypothetical examples from an academic department website.
Example #1: Considering a Physics Major
I'm a new Undergrad and I'm considering majoring in Physics. I'm looking at the curriculum for the major. I don't really know what I want to study yet and I could probably use some advice. I might want to just Minor in Physics or see what kind of research I could be involved in.
Through the circumstance above, we just identified some pages from throughout the site that are relevant to a category "Considering Major":
- Academics > Undergrad > Physics Major > Curriculum
- Academics > Undergrad > Physics Minor > Curriculum
- Connect > Advising > Undergrad Advisors
- Research > Centers and Programs > Undergrad Research Centers
We can then anticipate that there might be a number of other similar categories like "Current Major", "Graduating Major", etc.
Example #2: Graduate Financial Assistance
I'm a graduate student and I'm living with limited means. I'm looking at the financial aid options for my program. I probably also want to know about other financial support that's available, even if it's not directly tied to my degree program. I might also want to reach out to my peers or an advisor.
Through the circumstance above, we just identified some pages from throughout the site that are relevant to a category "Graduate Financial Assistance":
- Academics > Graduate > PhD > Tuition > Financial Aid
- Research > Grants > Department Project Grants
- Research > Grants > Grant Recipients
- Connect > Advising > Graduate Advisors
- Connect > Student Groups > Graduate Student Weekly Meetings
Dividing big categories: Aim for 5-6 items in each
The key to developing a related content grouping like this (or any grouping in general) is to limit the number of items in each category to 5-6 for scannability. Readers have a hard time processing more items than that, so you'll serve them better by keeping the number of items in each collection small. Below are some techniques we used to keep our lists shorter.
Tag the top page for a mini-section, not the details
One way we kept our groupings smaller was to only tag top level items of a section. So rather than tagging a page detailing how to apply to a specific grant, we just focused on the grant's top level page.
Break up large groups into more specific sub-groupings: add facets
If we had large broad groups like "Graduating", we took the opportunity of breaking them out by other facets. Here's an example of "Graduating" which is a process category , divided up by adding roles and circumstances.
- "Prepping for Graduation (Undergrad)"
- "Prepping for Graduation (Grad)"
- "Post-Graduation (Undergrad Alums)"
- "Post-Graduation (Grad Alums)"
In conclusion: Know your users
So, there you have it. We found that the key to developing a useful Related Content taxonomy was to know our users and their common motivations. I think that's the key to a lot of things!!
Have you ever made a Related Content taxonomy? How did you go about it?