Redesigning a University Website Through Evidence

In early 2017, I was looking for a research topic for my undergraduate thesis. I was already interested in website design, and through my connections with the Information Systems Directorate at Telkom University, I learned that the university’s main website had become a concern among the rectorate.

The problem became more apparent when I looked at the website myself. Its visual language was heavily inspired by Microsoft’s Windows 8 Metro UI, with large blocks of color and geometric layouts. The reference itself wasn’t necessarily a bad choice. The website that inspired it was well designed. The problem was how that direction had been translated. The color palette felt poorly composed, the visual system was inconsistent, information hierarchy was weak, navigation was confusing, and the entire experience had been designed around desktop use despite many students accessing the web through mobile devices.

The page load time was also slow, although that was outside the scope of my research. Since the website was built with WordPress, I could at least recommend basic measures such as image compression and reducing unnecessary plugins. And the more I looked at it, the more I realized that this wasn’t simply a website that needed to look better. It needed its design decisions to be grounded in evidence.

From Visual Problems to Design Problems

I decided to make the redesign my undergraduate thesis, approaching it as a design-based research project rather than a purely visual redesign. My first step was identifying what information I actually needed. I conducted a literature study and triangulated the findings through visual observations of comparable university websites, interviews with experts, and questionnaires involving 100 intended users. I then organized the findings through a comparison matrix before moving into concept development, prototyping, and final user testing.

One principle became particularly important throughout the process. Telkom University’s Graphic Standard Manual already contained the visual foundation I needed. It covered the identity elements that were relevant to the problems I had identified, yet the existing website didn’t follow it properly.

For a visual communication designer, a Graphic Standard Manual is a standard, not merely a suggestion. If the existing identity system already provides the necessary foundation, ignoring it creates an unnecessary problem. I therefore based the redesign on the existing standard rather than inventing another visual language.

The research also challenged assumptions about how the website should actually be used. Google Analytics showed that mobile was the most common access device, even though the existing website was designed around desktop use and wasn’t responsive at all. This was one of the clearest examples of why I wanted the redesign to be data-backed. An assumption about the audience had already been made for us. The analytics showed that assumption was wrong.

The solution was surprisingly simple. I used Bootstrap CSS to establish a responsive layout rather than building every media query myself. I could have implemented the responsive system manually, and I sometimes do that when experimenting, but this wasn’t the place to introduce unnecessary complexity. The problem was obvious, the requirement was well understood, and an established CSS framework could solve it efficiently. Not every design problem needs sophisticated technology.

Testing the Redesign

After developing the new interface, I tested it again with the intended audience. The comparison was straightforward. The original website received negative responses across the tested indicators, while the redesigned version received positive responses from all three participants.

Users reported that information was easier to obtain, contact information was easier to find, and the main navigation was clearer overall. They also responded positively to the color scheme, text readability, imagery, information grouping, and iconography used throughout the interface.

The result was consistent with the direction established through the earlier research. The redesign wasn’t simply a visual improvement. It addressed the problems identified through observation, user input, and available behavioral data. That distinction became important to how I understood design.

Intuition is Not Evidence

Looking back, I wouldn’t consider a website redesign particularly interesting as a thesis if the argument were simply that the new interface looked better, followed contemporary best practices, or had improved functionality. There was much more to investigate once the design was treated as a response to an actual system and its users.

There were technical considerations, from responsive implementation to performance constraints. There were organizational considerations, such as an existing visual identity that had already been documented but wasn’t being followed consistently. Most importantly, there were human considerations that couldn’t be resolved by looking at the interface alone or considering its visual structure.

One of the concepts this experience made particularly clear to me was false-consensus bias. It is easy to assume that users will behave in a certain way because that is how we expect them to behave. Designers are especially vulnerable to this because experience can make our assumptions feel like intuition. I no longer consider intuition sufficient evidence. Even when a designer has years of experience, assumptions about users remain assumptions until they are tested against reality. This is something I have continued to apply in my professional practice.

The project also reinforced my understanding of the idea that great design is invisible. When a design is bad, users may not know exactly what is wrong, but they can still experience the consequences. They struggle to find information, hesitate over navigation, feel uncomfortable with the visual presentation, or simply leave with a negative impression. But when the design works exceptionally well, the opposite happens. Users don’t necessarily stop to evaluate the interface or recognize the decisions behind it. They simply use it, move through it, and leave with a positive experience.

That is ultimately why I see design as more than visual composition. The interface is only the visible result of a much larger reasoning process. Good design begins by understanding the problem, questioning our assumptions, and using whatever evidence is available to determine what the solution actually needs to be.