I mentioned a while back that I would be writing on the process of improving this site. Up until this point, I have been making small, incremental changes throughout the site to see if they would impact key traffic metrics. My goal was to set up some A/B testing scenarios on my blog, but performing tests on anything but the most superficial of elements proved to be disruptive. I then decided to move to Plan B and implement custom event tracking with Google Analytics. In short, the results have not only been revealing, but unexpected. I always knew in a theoretical level how important deriving empirical data was in order to justify a design, but soon after I began to observe detailed click patters, that opinion became concrete. If I was writing my Golden Rules of Site Redesign article today, it may be a little different.
Basic web analytics just does not cut it
I check my web analytics often. Too often. However, your basic statistics such as page views, bounce rate, etc. only address the macro view of how your site is performing. I had a sense that certain areas were either working or not working, but I did not have empirical data to back it up. After a little less than a month of slowly tracking more and more custom events, I have a better idea of how people are using my site than I ever have — I truly regret that it took me this long to implement such a simple element. Moving forward, I plan to implement custom tracking events for all my future projects. It just does not make sense not to do it.
Your design choices, no matter how informed, are guesses
The more I look at the patterns that are coming in from my custom events, the more I am convinced that I would never have predicted these outcomes. The results that came in not only surprised me, they surprised my peers. Without being able to peer into the future, our choices are guesses. They may be educated guesses, but nonetheless guesses. The good thing is, we are not building a car or a house. We do not have to get the optimal design on the first shot. Our designs are be easily changed after the initial design — and they likely should. There are simply too many variables to keep track of and address correctly. Better to acknowledge our decisions are simply guesses until we can get some hard data to inform future choices.
Think, do, react
Designing the initial idea gets you halfway through the process. However, it’s really that set of decisions that occur based off the objective and subjective feedback from that initial idea which finishes the job. Sweating the details on the initial design may be important on a philosophical level, but it may end up just being wasted cycles. I played baseball all the way up until the beginning of my freshman year of college. Playing sports to that point taught me the importance of trusting a decision and committing to it. In most sports, there simply is not the time to dwell on how to handle a scenario — you simply have to make the choice you think is right at the moment. There is a strong chance it was not the best choice, but that is all a part of learning how to play the game. The important thing is to analyze your play after the game and learn from the poor choices in order to avoid them in the future. I really to think this train of thought can be a strong guide for how to consider design. I am not saying to only make snap decisions, but it can be far too easy to overthink problems that may not even be answerable with your current understanding of it. Better to be aware that you are making the best decision you could come to with the clear understanding that it is likely not the perfect one. We should all think before we do, but the magic happens after we react.
I will be going into the gross details on the data I have received on how people have been interacting with this site. I think you will be just as surprised I was.