Great post from a google designer who left for twitter because he wanted to be in a less data driven env.

It seems like he felt that the Google environement was unable to make daring/basic design decisions due to its focus on data.

Here’s one quote:

“I had a recent debate over whether a border should be 3, 4 or 5 pixels wide, and was asked to prove my case. I can’t operate in an environment like that.”

While I sort-of see where he’s coming from, for me it’s pretty much the opposite.

I’ve seen so many design changes asked for / made by developers, designers, and biz people that have had, when subjected to actual analysis, either a negative or absolutely no measurable effect, if the effect was even measured at all. This includes the pixel changes he’s talking about, but also includes site-wide redesigns done by professional designers in which the redesign had no measurable effect on bounce-rates/time-on-site.

I have found that subjecting things to actual testing / requiring that all changes have “proven their case” is the only way to avoid this.

The only alternative, as I see it, is to make design changes and develop features with no 100% post-release idea of whether the changes are actually working.

Over time, this results in a huge dumpster of design and code that Nobody knows whether or not is actually making any different to customers. Without natural selection of what actually works for customers, I don’t see how an app can evolve to make money.

The point is definitely not that design is irrelevant, but rather that, like everything else (including code), it is unlikely that even the best trained designers have a high enough success rate to warrant not basing decisions on data.

The point is also not that daring decisions shouldn’t be tried. If a design idea seems plausible, and there’s time/money to test it, I don’t see why not. Isn’t that the point of google labs?