I was stumbling around the net and came across a couple of videos from Google’s IOs. The IO-2010 was just a couple of days ago, and a lot of interesting things were said on it, but I also found many interesting videos from their previous IOs.
This particular one (http://www.youtube.com/watch?v=6x0cAzQ7PV) is back from 2008, and even thou the internet does evolve with the speed of an atomic bomb, most of what Marissa Mayer says in the video is still true in our days.
And most importantly, what she reveals in the video is not simply some specific tips about how to build great web applications, it is also about how Google operates to really come up with those ideas.
I guess you could say that the secret is trying to touch every knob of your website, change every variable, be it functional or plain aesthetic, and trying to see what effect it does on the user experience.
For example, Marissa said that back in the days when they were starting with the paid search results that were displayed on the Google search result page before the ordinary results, they were making experiments on such seemingly small things as the color of the bar with those paid advertisements. They were trying to see if blue or yellow would make the users feel better. I thought for myself that it would be approximately the same, because, come on, it is just a color, but according to her there was a significant difference and now all the commercial blocks are yellow.
The devil’s in the details. I assume the difference between the two cases was not that high, a couple of percent tops, but if today they make this change that got them 2 percent, tomorrow another one, and if they have been going on like that for a long time, constantly becoming better, changing those small things, those small insignificant variables, then in the end that is what made the magic happen and make Google the best search engine in the world.
Another important variable was the loading speed. They tested whether the users liked the search result pages with 10 or 20 results. (And I am still not sure about their measuring methods, but I guess they have their ways). The experiment showed that the users liked the result pages with only 10 results, but it wasn’t really the number of results that mattered in the end. It turned out, the time difference between google’s indexing server producing 10 results and the same server producing 20 results (20 takes more time to compute of course, even if it’s a a matter of milli-, if not micro-seconds) had the effect on the user experience. And the fact that you right now only get 10 results, which also means that you must click around the pages a lot more, is apparently worth getting the results fast enough.
And this is nothing the users think about consciously. I have never noticed those milliseconds, but apparently I (as an example of a Google user) like to get 10 results very fast instead of 20 results a little slower, even if I cannot consciously perceive that time difference. Basically, this was one of her big points in the keynote video: in order to make good web apps, the developer must know his users better than they know themselves.
And I guess that is the right approach. I mean, I consciously hate some things about google’s apps, mostly their overly simplistic “design”, but I still keep using those apps, so on some level even that design works for me.
There are some more great points Marissa makes in the video, so check it out. If nothing else, it is makes you start thinking about how to switch focus from merely writing some code to making great applications.