rahrahfeminista2

Five posts to my livejournal explained, from thoughtcatalog
Though I don't post here much anymore, I felt it appropriate to post. You know what? Tumblr might be less spammy and easier to post to, but I do miss my community here.

Hi friends. I've moved to Tumblr for a bit. My tumblr is called Underthinking. I'm hoping this means more posts because tumblr makes it that much easier to put things up and make things colorful. I would like to see you there!

I just read Cliff Kuang's Google Instant Proves Google's Design Process is Broken.

In it, he argues that good designers would have intuitively known the kinds of things Google belabored by putting things through prototype usability tests. He argues that Google evidences an obsession with killing intuition and inspiration by staying close to the provable and data-driven color tweaks and efficiency improvements. He builds from a post by ex-Google visual designer Doug Bowman in which Doug details the paralysis that comes from designing by the sword of quantitative analysis. Now, I agree with Doug's assessment that data-driven design is a paralyzing. I would even go further and say that it functions to hand power over to engineering modes of thinking and iterating over other ways of understand users, products, and innovation. (Less conspiracy theory and more emergent bias and power structure.)

However, I largely find the article to be one sided and adds little to Doug's argument. Worse, it mischaracterizes potential benefits of instant search in an effort to valorize designers as intuitive mindreaders.
I have a bunch of issues with the article.

First, the tone bothers me because it rides the whole hindsight is 20/20 thing. While New Coke's failure might be clear once you tried it out, I'm sure 10 "classically trained" designers or anthropological empaths would have have had different conjectures about whether New Coke would have worked. This is why real designers, not just Google engineers, prototype -- empathy is a starting point for guessing. Empathy isn't inception or mind-reading.

My second issue with it is that I think Kuang misses why instant search might have fans. I'm pretty sure Google wouldn't have launched such a radical and expensive change to the UI without some strong evidence that it helps some users. I know that query formulation is a big challenge for the majority of web searchers -- the kind of people who search a few times a week rather than a few times an hour. For those people, formulating queries is really hard. They're not super fluent in how Google interprets their queries. Also, when one searches in a domain they're not super familiar with, it can be hard to formulate a sufficiently narrow query without already knowing the kinds of materials and keywords within the document corpus. For example, how can you know that you should type "ruby stone" to eliminate pages about the Ruby programming languages unless you know such a language exists? In such cases, instant search lets them type until they find something they recognize as good enough (satisficing) or offers them the sort of fast feedback they can use to refine their query easily and smoothly, particularly if they have a slow internet connection or aren't fluent in keyboard shortcuts. I wouldn't have guessed it was about efficiency and shaving off seconds, but rather providing an experience of flow, control, and highly interactive feedback -- the kind of thing people get when they interact with and monitor the response of people they interact with but not so much with page-based interfaces. Kuang instead takes the cheap shot of jumping to conclusions about those engineers being at it again with their seconds shaving.

Third, Kuang says that Google shows you a bunch of irrelevant stuff now because they show so many results pages as you type. If you're an expert user who is confident in in your query, by all means ignore the page refreshes happening under the box. People adapt their visual practices with experience. But "irrelevant stuff" for a lot of less search-fluent users is feedback that helps them guide and refine their queries by showing them how their current one is doing by winnowing down the relevant chunk of the web google is returning to them.

Now, I don't think Google does itself any favors starting off with some boilerplate motivation like "the web has more information than libraries so we need to change search!" When information might mean bytes, and bytes can be porn, banner ads, and xanax spam, I'm hardly set up to be impressed about the value of this feature. I'm not sure why Google wouldn't highlight the much more interesting potential benefits of the feature. I hope I'm not wrong in inferring that there are these benefits.

Everything I'm saying is just my guess. I haven't talked to anyone about instant search and I haven't been at Google for three years. My opinion is my own but informed by the sorts of conversations I was immersed in while I was there. I'd love to hear what you all think or know.