Have you ever done Information Architecture(IA) for a University?
A university website IA Case Study
By Erin Toh, Student, General Assembly UXDI -12
Task: Rework NTU’s Electrical and Electronic Engineering Information Architecture (IA).
Disclaimer: The brief for this project was created by the teaching assistants at General Assembly, Singapore to imitate a real-life experience. The content and creator of this article is not affiliated with the company/organisation mentioned.
This was done in a team of two. The timeline given was 10 days.
The first thing we did was create a to-do-list. To rework an IA is no easy task, we needed to map out our time properly given the timeframe as there were many more deliverables expected in this project: Project Log, Research, User Testing, Rework of IA, Prototype in Axure, Usability test, Presentation etc.
#Tip 1: Create a To-Do-List and Project Log. This will help tremendously in getting the project going and knowing if you are on schedule. Refer to the project brief constantly in case you miss anything out.
We were given three personas, and had to choose one to move forward with. We will be moving forward with John’s persona.

Research Steps.
Have you ever been to NTU’s EEE website? If you haven’t, perhaps trying now would be a good idea, I have left you the link, www.eee.ntu.edu.sg.
NTU EEE is ranked among the top 12 schools worldwide in Electrical and Electronic Engineering.
But we found that the website did not reflect the prestigious image of the school. Why?
First we did a Heuristic Evaluation.
What is heuristic evaluation? It is a term for “your own perception”. We relied on the lessons taught and our own sense of design, to figure out what is wrong with the website.
#Step 1: We adopted each persona and went through their journey through the website with each of their needs and pain points in mind. We logged them into an excel sheet.
#Step 2: We mapped their user journeys at the same time. This made it easier later, when we cleaned the images up for a clearer picture.

#Step 3: Going one step further, I wanted to understand whether my views and notes were validated. By recording users, that suited the personas, giving them a task list that “John” would go through. And seeing how they react with the website.

#Tip 2: By being detailed with research, you can speed up the other processes after. By doing our journey and heuristics together, we effectively finished 2 processes in one. Do NOT skim on research.
#Tip 3: You need proper research to back your design decisions up.
Tip #4: If you are using a Mac, ‘Quicktime’ can record quick videos and voice. I also used ‘Discord’ as another screen tool. Where one can chat with your user and watch their screen.
Next, Competitive Analysis.
What is Competitive Analysis? It is a process where we compared NTU EEE’s website with competitors we deemed of the same caliber and prestige.
In this case, we used four different university websites:
- National University of Singapore (NUS)
- Singapore Management University (SMU)
- Singapore University of Social Sciences (SUSS)
- University of Maryland (UMD)
Why these schools?
Three are from local universities, which are often compared. The last is a university website which was voted one of the best online.
We wanted to know if there are similar terms and acronyms used. Also, if there were certain trends in how information is written and categorised across universities.

You can view the detailed version here.
User findings from our research.
- Most gave up after repeatedly going through the website and not finding what they wanted.
- Using a search function, that led them into main NTU website. Confused they wandered around the main website until realizing they weren’t in the EEE NTU website.
- Using the back key on google, and many tabs to navigate.
- Getting lost the moment they tried using the header or footer as it led them away from the NTU EEE website.
- There was not enough content on some pages. Whereas there were other pages, that there were too many things to skim through to read.
- Too many clicks to get to where a user needs to go. (Around 8 clicks)
In summary, people simply got lost and frustrated when trying to finish the tasks given.

Inventory Content of NTU EEE’S Site. No
This is finding out what information is currently residing in the website.
It was one of the most painful processes. We tried using a shortcut by using Screaming Frog SEO. However, that proved rather unsuccessful, as reading it was a pain, and the free version only lets you scan up to 500 items! That was only the tip of the iceberg for us.
We decided the best way was to combed through the website painstakingly logging the content and links into an excel sheet. You can view it here.

Tip #5: It also served as a content audit as well as a current sitemap. Remember to add links to the pages. It helped tremendously when going back to reference pages to improve design.
Synthesising Information.
Findings of the recordings and analysis.
21 Users feel the most important aspect of visiting a University site is the Efficiency of Accessing Relevant Information.

Card Sorting
We used a method called ‘Card Sorting’ from ‘Optimal Sort’ to find what users think categories should go under. We used a hybrid sort, where users are allowed to create their own categories as well as have certain topics that we created for them already.

Our Improved Sitemap.
Using our card sort results, we created an initial improved sitemap.
You can view it here.

Tree Testing
This is another feature from Optimal Sort. Here, we implemented the information from our improved IA and gave users another set of tasks, to see if they could find the information they previously could not.
We did two to three tree tests, to narrow and iterate our IA’s (Sitemaps) till we felt comfortable that it was working the way we wanted it to.

Wire-framing and Prototyping
Moving forward, we sketched our prototypes on paper first, using post-its to add features that we needed.

#Tip 6: We went too shallow by testing this with users, as it didn’t contain enough information at the start. A mid-fidelity wireframe would be much better to test functionality. But this was not wasted. It proved useful to map templates to pages.
Creating a Style-Guide for Our Prototype
Remember the quote earlier from the brochure? This is what I meant by how useful it was. We used school colours as well as brand guidelines and the brochure to create our styleguide to maintain the brand of NTU.

Tip #7: Creating a style-guide will make it easier to add your colours to your website prototype after. It will speed up the process of creating the prototype and make it less troublesome with trying to figure out the body and header text at the same time.
Iterations and Protoyping in Axure RP
You can see the prototype here.

Afterthoughts.
Working in a team is a different experience. I struggled with different things, expectations from and of a team member, deadlines and seeing how detailed we wanted our research to be.
Moving forward.
I would need to be more understanding of how each team member works.
I would like to improve the how the prototype’s functions work, work on the navigation, add breadcrumbs and add features where it would improve the website as a whole not only concentrating on one persona.
Phew that was a long one. Thank you for making it to the end with me!
I hope it would help in your quest in how to go about doing information architecture!
You can also view the full presentation deck here.
