There seems to be a great riff these days between design and analysis. Usually the argument from the aestheticians seems to be self-preserving rather than logical. In many ill-conceived articles and posts, these authors even go so far as to question the scientific method and the unequivocally powerful field of statistical inference. To me, these arguments are plain provincial. On the other hand, the analytically inclined camp also fails to appreciate the potential and relevance of informed graphic design.
Most websites are meant to create revenue for the owner, whether directly through sales or indirectly through product promotion and brand building. It is difficult to find a web site that does not derive some kind of financial benefit for the owner. (There are of course some, but small by proportion.)
One of the few times that purely aesthetic design is truly significant is when it adds value to the company’s brand. This is not insignificant, however, since brand for some companies can be worth the majority of their market capitalization. Despite this, it is buffoonery to discount empirical analysis when architecting user experiences. The real-estate that is your canvas is, without a doubt, the most valuable planar surface in the world and design decisions can make or lose the client millions if not billions of dollars. It is not by coincidence that large companies that heavily rely on their web properties for revenue, like Yahoo and Amazon, hire teams of economists and MBAs to analyze every link and pixel on each of their main pages.
Often, those that snub empirical analysis in the web design process do so because they are uncomfortable working with it and leveraging it to inform their design decisions. In order to effectively use the information provided in usability, interaction, and other studies, one must not only understand how to employ their findings but also their limitations.
I often think, that this polarized discourse must have occurred many times before in history. For example, I often imagine that a similar debate must have occurred between holistic healers and modern medical doctors in the 19th century. The holistic healers where skilled tradesmen that believed in their herbal cures and discounted the new ‘scientific-based’ field of medicine. They questioned the validity of ‘medical studies’ because they did not fully understand them or the scientific method behind them. In the end, it turns out that the field of medicine that is based on systematic, empirical studies is clearly effective. It also turns out that the holistic approach also have a great deal of validity; that it why a comprehensive cancer treatment often entails both modern medicine and ages-old holistic healing.
Empirical analysis is invaluable when architecting user experiences; that is clear. The web gives us incredible analytical potential because we have unparalleled insight into users’ actions and habits. Rather than fearing empirical analysis designers must embrace its awesome power. Conversely, analysts must come to terms with the fact that design is also critical and should work with, rather than against, graphic designers to come up with effective solutions that look great and support the brand. After all, albeit effective, not all sites can look like Craigslist.