Statistically spoken, statistics are just a method to define results by categorising them. The whole design process is a cultural thing and antropomorh thing.

I think that this shows that a designer must have a mixture of creative and scientific thinking else the final product will either be all eye candy or a goal-orientated solution that looks like a bucket of smashed crabs.

In a team, the designer often loses control over the more complicated functionality of the site. He/she should be liased with at all times to ensure consistency and usability.

Johan, no doubt there is a strong amount of subjectivity (colors, texture, typography to some extent), but once you go one layer deeper such as informational structure, organization, hierarchy to name a few - there are an abundance of studies that can give us a glimpse into what people understand and what people do not. We should not limit our responsibility to just making something look good - our design should improve the user’s overall experience, accessibility to information and cognition of the site we’re designing for.

Additionally, our improvements, if we truly are interested in the user experience, should ultimately lead to increased use of the site in one way or another - hopefully in the areas we were hired to improve. As I said in the article, most actions people take on a website are relatively concrete and easy to record. If our design truly aided the site, certain numerical metrics should show just that. If they show the opposite, we need to take that sort of evidence very seriously.

Also I found that flash websites have been used to experiment a whole lot with different forms of information architecture. From my own experience, you really benefit from playing around with playing aroundkeyframes and time-based animations, it are great ways to find out ways to present content.