Main Content

Determining the Navigation for Your Website

When we reconfigured the Information Architecture (IA) of the Unitarian Universalist Association website ( in 2011, we relied heavily on user input (from surveys, constituent feedback, and data from Google Analytics) and testing (surveys, card sorting, and tree testing) to maximize our chance of success. Here are some of the resources we used and the lessons we learned.

Reading, Tools, and Data

  • A Practical Guide to Information Architecture is a fantastic overview of basic IA principles—Donna Spencer lays it all out in an easy-to-read format, and points to several useful tools and processes.
  • Getting The Website Information Architecture Right: How to Structure Your Site for Optimal User Experiences is another great guide.
  • Optimal Workshop has online card-sorting and tree-testing tools so you can get input from people all over, and their results-analyzing tools are really impressive. If your tests are small, the site is free; larger tests can be affordable if you schedule them well (the site offers monthly and yearly subscriptions—we used a year-long subscription during the major redesign, and now save up several tests at a time to take best advantage of an occasional month-long subscription to test refinements).
  • Google Analytics, of course. This is where we gathered the bulk of our data about what was “most popular” on our site. We looked at page visits and site searches—what sections were getting the most traffic, and what content were people searching for most often (and what words were they using to do so). If you're building a new site, perhaps you can borrow some of this data from a similar site, like a neighboring congregation.

Reorganization Goals

We had a few goals when we started the project:

  • Lessen the depth of our site.
    We had three main navigational options, and that meant that the majority of our content was several clicks away and several layers deep. We wanted to make sure our most popular content required fewer clicks to get to.
  • Shorten our URLs.
    We wanted to make sharing our site via speech, print, email, and social media easier, so shorter (but meaningful) URLs were important. Being able to say/write, “Go to” was so much nicer than “Go to and search for ____” or “Go to and click on ____ then on ____ and then ____” or “Go to”
  • Make it easy to find stuff via navigation.
    We wanted to use language and groupings that would be obvious and unambiguous to as many people as possible. When we weren’t sure about how to label a resource or a category of resources, we looked at what people were typing into our site search.

    For example, we learned that people searched for “email lists” more than “listservs” and “religious education” more than “faith development.” Where the language people were accustomed to using wasn’t incorrect or misleading, we used it, even in preference to our own branding. (Sometimes we used both, if the numbers were about even, i.e. “Our Whole Lives Sexuality Education.”)

    We ran some card-sorting exercises to help determine how resources ought to be grouped together, and then we thoroughly tree-tested those organizations to make sure they made sense to as many people as possible. We did our best to eliminate vague labels, also known as “evil attractors” or “dirty magnets”—labels that attract clicks when they shouldn’t. “Membership” turned out to have too many meanings to people to be useful as a category label, even though lots of people suggested it during our card sorts, so we used “Growing Your Membership” to more clearly convey what would be in that section. We added cross-referencing links when there was more than one “best” place to put something.

Lessons Learned

Don't Argue; Test

When egos or ideas conflict during the process, put it to the users. Everyone has the tendency to believe that their way of thinking is somewhat universal, but letting a whole bunch of people show you how they think about your content will clarify where the best home is for each of your resources.

Use Topic- or Task-Based Organization

In most cases, a topic-based or a task-based organization will serve people best. 

Why audience navigation usually doesn’t work gives some fantastic examples illustrating why people find audience-based navigation confusing. An audience-based organization can work if the groups in question are totally distinct and separate from one another, and if the tasks related to each group are totally different. The UUA Health Plan page is a pretty good example of an audience-based organization (the headers) with a subsequent task-based organization (the bullet points). Most of the time, our audiences aren't so distinct, so a task- or topic-based structure will probably work better.

Communications and Social Media for UU Congregations is a good example of a topic-based structure. How to Run the Welcoming Congregation Program is a good task-based list.

You can have more than one kind of organization—you might use your sidebar to organize things topically, and your landing page to organize by task (or vice versa). (Either way, good usability and search engine optimization dictate that everything in the sidebar ought to also be linked in the page body, and everything in a section ought to be included in the sidebar.)

Sometimes an alphabetical organization might work, if the items you're listing have universally agreed-upon names and spellings (like U.S. states).

Listing by importance can also be handy: people will pay more attention to the top of your text area than anyplace else on the page, so putting your most popular or most important resources there is rarely a losing proposition.

Often data from your site's analytics can offer support to one approach or another, and content experiements (A/B tests) can help you determine what works and what doesn't for the people trying to use your site.

Share, Print, or Explore

For more information contact