Thursday, November 16, 2017

search - Is Google Instant good or bad UX?




For power users, it clearly saves time in some circumstances. It also jumps around alot and does things that are very different than before. One person said it is like a guy who finishes your sentences, "More annoying than helpful".


What's the verdict?



Answer




I'm the guy who did all the prototyping on Google Instant, and I attended >90% of the user studies we conducted at Google. (and I'm going to try not to be biased!)


We went through several iterations of prototypes based on our internal testing. What we found worked really well with the current implementation was that study participants either (1) saw the results while they were typing, and proceeded to click what they wanted, or (2) never noticed anything changing because they were hunting-and-pecking on the keyboard, or else too busy looking at the suggestions. That validated our goal that it wouldn't be too distracting. Of course, when a large part of the page changes like it does in Instant, users can't help noticing the change; the important thing is it shouldn't interfere with their task.


Yes, it might feel for certain queries that Google is thinking for you, and auto-completing things for you. Though, this happens much less frequently than you'd think, but you end up remembering more vividly. All the other times when you saved time by not having to type an entire query, or because Google suggested a better query than what you were intending to type, or because you saw a quick preview of the results and decided to refine your query further — all those instances get forgotten more easily because you expect Google to work that way. Try keeping track of all the instances where it worked well for you, then compare against the number of times it felt odd to you.


One of our participants said: “It found a way of asking the question that I hadn't even considered.” (verbatim, from study transcript.) Google Instant helps a lot of users formulate their query better. Many of us here cannot imagine query formulation being a huge problem, but it is. Instant tries to make that just a little better.


We also minimize the amount of flickering on screen when results change. Given the realities of network latency in shipping a new set of results from the server to the client, we have a lot of black magic going on in the interface that attempts to minimize the visual noise when results transition. For example, when new results are available, the old ones fade out quickly through a transition to grey; this was noted as a strong indicator by our users that "Google is thinking"; without it, users didn't realize that results were different, unless there was a huge map or a huge picture in the middle of the new results page.


Here's a longer description of the user studies we conducted.


No comments:

Post a Comment

technique - How credible is wikipedia?

I understand that this question relates more to wikipedia than it does writing but... If I was going to use wikipedia for a source for a res...