Let’s say that you are commissioned to redesign an online office supplies website. You have the client’s product list with which you have to design the site’s information architecture.

How would you come up with a navigation scheme that is effective in helping site visitors find products or topics on the website?

Designing delightful web experiences requires a combination of visually appealing interfaces and an information architecture (IA) that is easy to navigate. Testing navigation structures early in the interface design process ensures a sound IA.

If the menu is not designed yet, then designers can obtain help from their target audience to inform design decisions that determine how information can be organized, structured and labeled in an effective manner.

There are many online research tools that help designers arrive at and evaluate their information architecture.

Different Information Architecture Study Types:

Card sorting, tree testing, screenshot click testing, and unmoderated remote usability testing are research methods that are used in the different phases of an application’s design and development cycle.

Card sorting and tree testing are often used in the formative stages of the design, particularly when a site’s information architecture (IA) is still being designed.

Naturally, there are always exceptions but the type of results from these tests are most often made in the design phase.

Screenshot click testing, prototype testing and unmoderated remote usability testing methods are often used during the middle (Design, Development and Testing) stages of design or during the maintenance stage, once the site is live.

Using more than one research method to arrive at a navigation structure provides a credible argument for why and how a menu design scheme was selected. Card sorting and tree testing are two research methods that are often used in conjunction to create and validate menu design structures.

Let’s follow a sample scenario to make sense of how information architecture studies such as card sorting and tree testing compliment each other.

Sample Scenario:

As a researcher you are commissioned to re-design the navigation structure of an online office supplies website. You have a list of product web pages that need to fit into a meaningful navigation scheme.

Figure 1: Product list that needs to fit into meaningful categories in order to optimize information finding
Figure 1: Product list that needs to fit into meaningful categories in order to optimize information finding

Goal:

The goal of the study is to understand how participants think about the list of topics, and how they categorize them into meaningful groups.

Ideal study type:

Open card sorting is the best study type since it is designed to aggregate different categorization schemes across multiple responses. It visualizes the data and comes up with clusters of categories most commonly arrived at by participants.

In an open card sort participants create and label their own groups of items/cards. It enables the understanding of how participants think about information.

Participant Profile:

  • Age between 25 – 65
  • Gender, approximately 50/50 male/female split.
  • 35 respondents that are comfortable making purchases online

Results:

All online card-sorting software present the results in the form of a dendrogram, distance matrix, frequency by category and frequency by item.

These results from a card sort help you establish relationships between the items, strength of these relationships and in the case of open card sorts, how users think about naming the category of items.

Figure 2: An open card sorting exercise using UserZoom
Figure 2: An open card sorting exercise using UserZoom

Outcomes of Card Sorting:

Card sorting results show:

  • The frequency with which items were categorized together.
  • Which items were difficult to categorize.
  • Which items need redundant categorization.
  • Popular category titles that participants create.

Watch our webinar on card sorting analysis, which explains all aspects of conducting a card sorting study.

Testing the strength of a card sort result

Let’s say that the results from the card sort scenario have demonstrated that participants categorized the list above into 4 main categories: Ink & Toner, Paper & Pads, Office Supplies and Technology Products.

How can we ascertain that the categories Ink & Toner, Paper & Pads, Office Supplies and Technology Products and the items in these categories are robust and that participants would indeed find it easy to find the items? That’s where Tree Testing comes in handy.

Tree Testing 

Tree testing is also known as reverse card sorting. Tree tests help measure the findability of items in a navigation menu. Tree test design entails presenting users with items that are categorized and sub-categorized into a simple navigation menu without any visual cues.

In our example, let’s say the results from the card sorting exercise created the following categories and product grouping: 

Figure 3: A navigation scheme that was arrived at after an open card sort exercise
Figure 3: A navigation scheme that was arrived at after an open card sort exercise

Goal:

The goal of the study is to test the efficacy of the card sorting result and ascertain the findability of items.

Ideal study type:

Tree test is the best study type since it is designed to aggregate different categorization schemes across multiple responses. It enables the evaluation of the findability, labeling and organization of topics in a navigation structure.

An example task in a tree test study would be: “Where would you go to find a mouse pad?

tt5
Figure 4: A tree testing exercise using UserZoom

Participant Profile:

  • Age between 25 – 65
  • Gender, approximately 50/50 male/female split.
  • 20 respondents

Results:

The results from a tree test show the percentage of success and how many participants found the item in the first attempt vs. all attempts.

To have a more in-depth understanding of tree testing analysis, watch our recorded webinar on how to improve content findability & optimize information architecture with a tree test.

Benefits of conducting Online Card Sorting & Tree Testing

  • Easy to execute: Online card sorting and tree testing are a great way to test navigation structure before putting a lot of effort in designing and coding it into any formal layout.
  • Test multiple versions of navigation structures: You can test several different combinations of categorization and see which structure fairs better.
  • Complement each other: Tree tests complement card-sorting studies really well in that, category structures established in the card sorting studies can be re-evaluated in a scenario based tree testing study.
  • Old vs. new navigation design test: Tree tests are good also in a re-design scenario. By presenting a stripped-down tree test of the old vs. new navigation structure, you can measure just how much better (or worse) the proposed navigation menu is to the old one.

Conclusion:

Card sorting and tree testing studies are especially suited in rapid prototyping, iterative and agile development models, where the time between sprints is short and so design decisions must be made rapidly as well.

Using online prototyping and testing tools helps you inform design decisions without compromising on testing with users or time. You no longer have to rely on lab tests to make quick-and-dirty design decisions.


About the Author

Sneha Kanneganti, User Experience Researcher at UserZoom

Screen Shot 2014-12-29 at 11.37.58 AM

Sneha specializes in unmoderated remote usability testing. She also has extensive in-lab usability testing experience with various research methodologies, from in-depth interviews and task based studies to focus groups. Her background is in Information Science and Psychology and loves analyzing results and finding patterns in collective behavioral outcomes.