A redesign does not begin by choosing new colours. It begins by looking at the existing website from a little distance: what still works, what creates confusion, and what no longer tells the story of the project it is meant to represent.
For me, redesigning first means reading. Reading the structure, the weight of information, the order in which things appear and the kind of impression a page leaves before someone has even started to use it.
Not everything needs to be replaced. Sometimes it is enough to remove noise, change a hierarchy, give an important piece of content more room or make a relationship clearer. In other cases, the tone itself needs to change, building a visual language that brings the website closer to what the project has become over time.
The comparisons below are meant to show that transition. Before and after are the visible result; what interests me most is what happens in between.
Studies referring to external brands are independent visual exercises, not official websites or commissioned work. Trademarks, names, texts, images and related rights belong to their respective owners.
My own website has also been the place where I learned to change my mind.
Over the years I have used it to test very different visual languages. Some were closer to photography, some more graphic, others deliberately illustrated. I do not see those versions as mistakes to hide: they are stages that show how visual taste becomes more precise and, above all, how the understanding of what a page needs to do changes over time.
These are personal experiments, not commissioned work.
For a while I pushed the website towards an almost archival form: large images, a slower rhythm and an approach closer to a personal publication.
It worked well as an authorial narrative, but less well when the website also needed to explain services, skills and ways of working clearly.
Then I tried a more illustrated and openly personal direction. The website almost stopped behaving like a traditional portfolio and became a small visual world of its own.
It was a useful experiment in understanding how much character an interface can carry before the visual language starts to overtake the content.
Earlier still, I had worked in a much harder and more graphic way: strong contrasts, visible structure, frontal composition and very few visual compromises.
It taught me how to build presence and hierarchy without relying on decoration. Many of those lessons stayed with me, even though the final result today is much quieter.
