SOME RANDOM DUDE

Anyone that has read Andy Rutledge in the past knows that his style is intended to immediately get your attention. The articles begin with a strong opinion that is usually backed up with good thought. Nonetheless, I was a little taken aback when I read a recent article of his, Objectivity Be Damned, where he seems to pit designers against analytical statisticians - “Objective data is helpful in this arena, but it’s the designer’s intuitive ‘feel’ that must prevail else design is the realm of statisticians.” My argument is that designers for the web need to take a lesson from their fellow statistician. Most of the work on the internet that designers do is directed towards leading users to definable and definitive actions. Therefore, I suggest that designers ultimately cannot rely solely on “feel” or subjective solutions to solve these objective problems.

Past Mistakes by the Design Community

Rutledge suggests that designers have inside knowledge on how to create experiences that lead to the productive use of a site. If designers had inside knowledge about user experience, then scientific studies analyzing user interaction with interfaces would not have any substantive findings to the contrary. The studies would have found that all those well designed sites performed swimmingly and everyone would have gone home. However, we, the web design community, have found an innumerable amount of internet users have been left perplexed and frustrated.

Web statistics are not perfect by any means. Nonetheless, over long periods of time with a large enough segment to draw from, those numbers can begin to give amazing insight into patterns, trends and characteristics of your site.

Like many other human beings, we address issues from an egocentric point of view, i.e., if I understand it, then everyone must as well. Through the years, designers have learned that all those great ideas we had about how to improve the user experience including splash pages, flash intros and pop-ups sites, to name a few, were not so great after all. We know now that the public detests those design elements due to, you guessed it, statistical data. I can see many designers saying, “Those design elements were done by bad designers”. If our job is so subjective, who gets to decipher the good from the bad design? I would argue that we already use very objective arguments to judge a design/designer’s success. To be fair, these poor design choices were not really our fault - no one had any idea what they were doing. The internet was budding and there was not a user’s manual to follow. Web designers did the best we could, we took chances, but ultimately we ended up, doing as much good as we did bad. Still, no harm, no foul, this experimentation helped move design on the web forward - we just need to learn from our mistakes. How better to learn from our mistakes than utilizing research and objective data? Invaluable data can be extrapolated to give us a glimpse into how people use our site, potential problems with different facets and the before/after results after changes were made. We cannot read the minds of our audience. Fortunately, we have a record that may give us a clue as to what they were thinking. If our job is to communicate and to create a visual construct that facilitates various goals, we would be foolish to not have that data be an integral part of our design process.

Measuring Design Success (and Failure) Quantitatively

Designers need to consider the full picture as part of design, including how users interact with the site, navigate through information, different sections, etc. Numbers and data become important when taking these elements into account. There is no doubt that numbers can be a very humbling power on a designer. However, if the goal for a particular project is to increase traffic/exposure to a particular area of a site and the results show the traffic has gone down, the design has failed. Period. It does not matter how much better the site looks, if certain tangible goals were expected to be met from a new design and the numbers show a negative response, the design did not succeed.

Numbers can give us an idea what we are doing right or wrong. What the numbers cannot always tell us is how to improve or fix the design. Any statistic can be twisted, this is true. However, using that argument to lessen the significance of statistics seems problematic when compared to the uncertainty of decisions made purely under subjective direction. Web statistics are not perfect by any means. Nonetheless, over long periods of time with a large enough segment to draw from, those numbers can begin to give amazing insight into patterns, trends and characteristics of your site. Those numbers are not going to give you a solution, however they can give you a metric to conclude that, a) you have a problem, b) you have fixed that problem or c) you have not fixed the problem.

Leave Your Ego At the Door

A large portion of web designers obviously did not understand what general users wanted. Ironically, it was the engineers that taught us. Websites like Google created unobtrusive, simple, function-centric sites that flourished while many of the behemoth, overdesigned sites (does Boo.com ring a bell?) slowly (or quickly) lost traction. We have learned from our lessons. Yet, the hubris of saying we designers have some innate ability to know what users want that cannot be backed up with some level of objectivity creates the risk of repeating old mistakes. Rather, the analytical, statistical method of studying traffic trends, patterns and responses to site changes can give us a glimpse of how users actually interact with the site. The data provides designers with an invaluable tool to move forward in improving upon different areas of a site’s design.

With AJAX taking a greater role in web interfaces everyday, we will, no doubt, allow our superior creative ability to create chaos once again.

I think we all know that getting users to interact with a site the way you want can be a lot like herding cats. You can spend months trying to put yourself in the mind of the user only to find that they are confused once they are brought into a usability session. Through years of trial and error, designers have a better idea of what works and what does not on the web. Yet, each site has a unique set of problems to solve. There is not going to be a whitepaper or usability study for every problem that needs solving for each project. We are definitely going to have to fly blind in some cases and use our “feel” to best address each issue. Nonetheless, the second quantitative data becomes available on those specific areas, we would be foolish to not only take it seriously, but to use it as a metric as to whether the proposed solution was, in fact, the right one.

We must be very clear - the visual style of a site is only the surface of its design - scratching beneath its surface should reveal much more thought and creative execution. If we consider the design process to stop at the visual level, then yes, our job immediately becomes much less about a synthesis of logical process and much more of a subjective feel. If that is the case, we are not doing much more than applying virtual wallpaper.

Future Mistakes to Make

Designers are still learning lessons about how to design traditional web interfaces more effectively. However, the web is still maturing at a frightening speed. When Flash first came out, we made sure to let our “feel” ruin people’s perception of the format for years. With AJAX taking a greater role in web interfaces everyday, we will, no doubt, allow our superior creative ability to create chaos once again. Honestly, that is OK - without experimentation the web would become quite a boring place. Nonetheless, we need to judge our experiments’ success and relevancy by objective figures on how users react. There are plenty of times when users can actually like the look and feel of the site and before they know it, unbeknownst to them, they find another website to fulfill their needs.

The Experience Site Loophole

Frankly, a large majority of websites on the internet are driven by goals that lend themselves nicely to numbers - increased traffic, increased registration, greater sales, more downloads. Experience sites are quite different. Since the site is less action-driven, getting “lost” for short periods of time may not always be a bad thing and possibly encourage deeper interaction due to its entertainment value. Hi Res! made a living on those sorts of sites with wild success. To be honest, experience sites are a designer’s dream come true as their desire for experimentation are not only asked for, but are necessary. Subjectivity here is king as these sites are almost always about feel and less about tangible actions for the user to commit. At the end of the day, we need to know what we are designing and why. Ultimately, we need to accept that experience sites are the exception to the rule and keep them in their proper context. We definitely can learn valuable design lessons from them, but we need to be aware that comparing them to most sites is like comparing apples to oranges.

Conclusion

Much of design on the internet is directed towards simple, tangible and recordable goals. Ultimately, those goals need to be gauged a success or failure based on objective data. Designers should tackle subjective problems with creative subjective solutions and tackle objective problems with creative objective solutions. Better yet, tackle both types of problems with both creative subjective and objective solutions. Nonetheless, you better have some objective results at the end of the day. Good design should produce positive tangible results well beyond a pretty exterior. If we have in fact achieved such goals, it is in our interest to tout such gains. Either way, it is not in our interest to ignore solid statistics - it ultimately hurts the client with a potentially sub par product and it denies the designer invaluable feedback and the possibility of growth. beautiful, design, empirical, logic, numbers, objectivity, statistics, subjectivity web design