Keeping an eye on Google – Eye tracking SERPs through the years
It all started with a triangle. And since then, we’ve been keeping an eye (well, several eyes, actually) on the changes that Google makes and how they affect searcher behaviour.
A word of caution…
The heatmaps below show aggregated gaze data, which is assumed to be a proxy for attention. Basically, they show you what parts of the SERP people look at the most. I’ve purposefully left out other data like click through rate, user satisfaction, time to search, query length and variety, and so on, so that we can focus on the way searchers look at the page.
The heatmaps also don’t show the order that searchers look at things, merely the distribution of gaze over the entire duration of a search session.
Basically, the data has been narrowed down to show which results get the most attention.
With those words in mind, take a look at what’s happened to searchers’ gaze behaviour over the last 6 years.
Our landmark research study shows the importance of the “Google Golden Triangle” – the top three or four results on the SERP are the place to be if you want to be seen by a good proportion of searchers.
Research into what makes people click informs us that this is because Google does a good job of surfacing the most relevant results. As pointed out in the report, “the Golden Triangle exists not because we are obsessed with position, but because we expect Google to do its job and put the best result at the top.”
We couldn’t stop there. So we compared Google with Yahoo! and MSN. Despite being given identical tasks on all three engines, searchers covered a lot more real estate on Yahoo! and MSN SERPs than they did on Google.
Along came universal and personalized search. What did this do to behaviour?
Universal search results in chunking and fencing of results – images create mental barriers that cause us to divide the results up into groups around the images and restrict eye movements around them.
Personalized results in the 3rd, 4th, and 5th organic positions (which we mocked up based on previous sites the searcher had been to) drew a lot of attention.
More barriers starting popping up in SERPs – local listings, more blended results. So we did some more eye tracking and wrote a white paper about how all of these things on the SERP cause us to do more chunking and fencing.
Google really shook things up and introduced Google Instant. We put together a quick study that we presented at SMX East of searchers who had never used Google Instant before trying it out.
Brand new Instant users didn’t show much change in behaviour.
So we checked again 7 months later with people who had used Google Instant for a while.
In general, still no big changes in eye tracking behaviour with Google Instant…
… although for particular types of searches (transactional searches shown below) it looked like searchers weren’t scanning as far down the page.
Google is constantly innovating, and the question always comes up about how the changes they make potentially affect searcher behaviour. While we don’t know what the next changes will be, rest assured that we’ll be keeping some eyes on them.