It was bound to happen sooner or later (and I’m not surprised that it happened sooner).
Bryan Eisenberg posted “gaze plots” generated by Feng-Gui of Facebook and Google+ this morning, and asked the G+ community for feedback on what they thought. So I replied…
First though, thanks to Bryan for sharing this, because it is important to think of design from a behaviour standpoint, and having these discussions will bring more awareness of that.
So, here’s Bryan’s original post and my reply:
“I just ran a Feng-Gui visual attention analysis on FaceBook stream versus Google Plus stream. I am going to write up my full analysis but I am curious about your impressions of the gaze plots. What does this show about what each one’s primary goals are?”
I think it’s important to look at these “gaze plots” with the understanding that they are generated by a computer.
Yes, Feng-Gui and other systems use algorithms based on vision science, but they’re generally quite poor at taking intent into consideration – that is, our gaze behaviour can change quite significantly depending on what we want to do, and computers typically can’t account for that.
These tools use properties of visual salience (e.g., contrast, colour, edges, etc.) along with things that we know are visually important to humans (e.g., faces), and typically profess to represent the first 3-5 seconds of behaviour (I believe Feng-Gui does 5 seconds).
So you can look at these “gaze plots” and assume that they’re reasonably accurate given the constraint that the “person” looking at the screen has no specific task or goal they’re trying to accomplish and has never seen the screen before.
There are a couple things that immediately jump out as questionable to me:
1) The first “fixation” in both cases is over on the right side of the page. For G+, it’s on the “Start a Hangout” button, probably because it’s big and green and stands out from the rest of the page. But I would wager that very few people actually look there first, especially if they’ve been to the G+ homepage before.
For FB, it’s the ad for Crazy, Stupid, Love. Humans are extremely adept at ignoring ads, especially when we’re familiar with an interface and know where they are (not that everyone would ignore it – maybe you like to check out the ads… but it’s again unlikely that you would do it first).
Most eye tracking studies reveal that Western cultures typically look at a web page from left to right and top to bottom (the same way we read), in an F-shaped pattern. Yes, there are deviations from this, but the “gaze plots” here don’t resemble this at all.
2) There is no “attention” paid to the stories appearing the in the stream/news feed. Again, this is extremely unlikely, because most people will be coming to these pages for one of two reasons: they want to read what their friends are up to, or they’re going to navigate to a different page (the navigation isn’t getting much “attention” either).
Anyway, I will stop my diatribe there. Thanks for sharing, Bryan, as thinking about design from a behaviour perspective is always important. I’m quite confident in saying that Google has probably done a bunch of real eye tracking and user experience research on the G+ interface and has taken those results into consideration.
What Do You Think?
So you’ve heard my position on it. What do you think?