Running card sorting online is easier than you think and ensures users can find answers to their questions, thereby improving conversion.
This guide is sponsored by UX Metrics
- 1. Create an Account on UX Metrics and Login
- 2. Create a New Open Card Sort
- 3. Add Cards to Your Online Card Sort
- 4. Publish and Share Your Online Card Sort
- 5. Analyse Your Results
- 6. Run Closed Card Sorting Online
- 7. Create a Tree Test for Testing Your Information Architecture.
- 8. Assign Tasks for Your Tree Test
- 9. Analyse Your Tree Test Results
The following quick reference guide uses a free account with UXMetrics to carry out card sorting online. This process will help improve the information architecture of your site and overall site usability.
1. Create an Account on UX Metrics and Login
If you wish to run card sorting online, you will need a tool to manage the process of allowing participants to organize cards. By far, the best tool available is UXMetrics that offers a generous free account.
It also allows participants to carry out card sorting on mobile devices, unlike many of the competitors.
Once you have registered, login.
2. Create a New Open Card Sort
Once you have logged in select “Card Sorts” and then “New Card Sort”.
Enter a name for your test. This is only for your reference and will not be seen by participants. The instructions for participants and the thank you message can be left as-is unless you wish to customize them.
Click “Next” and select an “open” card sort. This type of card sort allows users to organize cards into any categories they wish.
3. Add Cards to Your Online Card Sort
To add cards to your card sort, sort click the “New Card” button. I recommend not adding more than approximately 30-40 cards. More than that, you are in danger of overwhelming participants, and sessions will become painfully long.
Each card should represent an individual type of content. That content might represent a benefit of your offering, a feature, a question users might have, or an objection that could stop them from acting. It might also include a task that users may wish to complete on the website. Focus on the most critical content if 30-40 cards limit you.
Once you have added all of your cards, click “next.” If you have a pro account, you can upload a logo and change a few other settings. Otherwise, click “Next” again.
After you have previewed your card sort, click “publish.” UX Metrics will give you a URL you can share with participants.
To recruit participants, I recommend using people outside of your organization and, ideally, people who are your target audience. Recruitment through a mailing list or your social media channels can be effective.
Failing that, you can use friends or family. Do not worry if participants are not experts in your field. If you can create a clear information architecture for non-experts, it will also help experts reduce their cognitive load when navigating your site.
According to the Nielsen Norman Group you should aim to test with at least 15 participants.
5. Analyse Your Results
Once all participants have completed the test, log in to UXMetrics, select your card sort, and click “End this Study.” Now you can click “View Report.”
Click on the “Groups” tab to see all of the groupings that users created when organising your cards.
Pay particular attention to the groups that several participants created. Reoccurring groups are a sign of a label that matches people’s mental models.
You will also encounter groups that are similar but not the same. For example, some participants may label a group “about,” while others use the term “about us.” Merge these groups by selecting their associated select boxes and clicking “Merge Selected Groups.”
Using the data available, you should be able to weed out rarely used groupings and merge others until you have the basis for your information architecture.
6. Run Closed Card Sorting Online
If you are working with a lot more content than you could be covered with the 30-40 cards of your initial open card sort, you may wish to run a second test called a closed card sort.
A closed card sort asks participants to organize cards into predetermined groups. These would be the groups you identified in the previous step.
Because participants are working with predefined groups, you can ask them to organize considerably more cards.
By running a closed card sort, you can check that your sections work for all of your content, not just the top 30 cards.
To run a closed card sorting online, repeat steps 1-3 but select “Closed” instead of “Open.” You will also need to define your groups when asked.
7. Create a Tree Test for Testing Your Information Architecture.
The final stage of card sorting is to check that people can find content within your new information architecture.
You can do this by creating a “tree test” in UXMetrics. Click the “Tree Tests” menu option and then “New Tree Test.”
As before, give the test a name and click “Next.” Now enter your information architecture using the tools available.
8. Assign Tasks for Your Tree Test
Once you have entered your site structure, you can assign a series of tasks for users to complete by navigating through the tree.
Limit yourself to 3 or 4 tasks and focus those tasks on finding critical content from your site. For example: “Imagine that you want to update your credit card details. Where would you find that?”
Avoid using words in the site structure as that will bias your results.
For each task click “set correct choice” so the test knows when a participant has succeeded.
You can now complete setting up and running the test as before.
9. Analyse Your Tree Test Results
Once you have finished running the tree test, you will be able to access the results. Pay attention to the percentage of people who completed the task and whether they did so by the most direct route (directness).
Also, note the length of time it took them to complete the task. The faster they did it successfully, the better your information architecture.
If users make mistakes, look at where they were going wrong and amend your site structure accordingly by changing misleading labels.
But, if you see a high success rate and participants achieve this reasonably fast, you can be confident you have created an effective information architecture.