Eyetracking study

Google Eye Tracking Report
July 2005

Released by Enquiro, Eyetools and Did-It

Prepared by
Gord Hotchkiss, Enquiro
Steve Alston, Enquiro
Greg Edwards, Eyetools Inc.

PDF $149.00 Print & PDF $179

How Searchers See and Click on Google Search Results


Contact Information:

Enquiro
Search marketing is an incredibly effective channel for connecting with potential
customers. And the more you know about how your customers interact with search,
the more successful that channel will be. At Enquiro, we never stop exploring how to
make search more successful.

www.enquiro.com
1-800-277-9997

Eyetools
Eyetools provides tools and services to measure eye-movement as people look at
webpages to quantify what people read, what they don’t read, what they glance at,
what they skip and what they never see... and then we correlate this to their clicks,
comments and actions.

www.eyetools.com
1-916.792.4538

Did-It
Jupiter Research named Did-it.com #1 in market suitability among all agencies
evaluated in their December 2004 “SEM Agency Constellation Vendor Evaluation”.

www.did-it.com
1-800-932-7761

2 Copyright 2005 Enquiro Search Solutions Inc. Reproduction in whole or in part of information obtained from
this study is prohibited without prior, written permission.


Contents
Foreword 5
Introduction 6
Importance of the Golden Triangle 7
What Is Google’s Golden Triangle? 7
The Golden Triangle 7
The Golden Triangle and the Google Effect 8
Interaction within the Golden Triangle 10
General 10
Entry Point 10
Reorientation 11
First Fixation of the Eye 12
Conclusions Regarding First Fixation Point 13
First Signiﬁcant Scan Activity 14
Insight: What You See Is What You Click 16
Where the Clicks Happened 21
Interaction with Top Organic 22
F Scan Patterns 23
Interaction with Lower Organic 26
Impact of OneBox 29
Impact of OneBox on Scan Activity 32
Impact of Top Sponsored Ads 34
Is Top Organic Worth the Premium? 35
Interaction with Side Sponsored Ads 36
Insight: Why We Ignore Advertising 40
How We Scan a Listing 41
Regression Analysis of the Importance of Factors in a Click-Through 42
Semantic Mapping 44
The Impact of Default Attractors in Semantic Mapping 47
The Role of Intention 47
Attractors in Consumer Research 47
Product Details 48
Brands 49
Prices 50
Trusted URLs 51
Reading of Titles vs. Descriptions. 52
Insight: The Role of Brand in Search 53
Insight: Mapping Search Behavior Against the Buying Funnel 55
Insight: Right Brain vs. Left Brain in Search 56
Impact of Bolded Search Queries and Icons 58
Insight: Blink, Thin Slicing and the Art of Search 60
Insight: General Search vs. Vertical Search Intention 62
Insight: Mapping Searches vs. Navigation Searches 66
Conﬁdence with Search Results 68
First Visit vs. Repeat Visits 71
Insight: Doing a Search Results Landscape and SWOT Analysis 74
Insight: Solving the Riddle of the Sponsored/Organic Multiplier 75

Copyright 2005 Enquiro Search Solutions Inc. Reproduction in whole or in part of information obtained from
3


Searcher Behavior 76
The Quick Click 77
The Linear Scan 78
The Golden Triangle Scan 79
The Deliberate Scan 80
The Pick Up Search 81
The Importance of the Deliberate Searcher 82
Demographic Analysis 83
Note on Demographic Analysis 83
Men vs. Women and Their Search Patterns 83
Insight: Online Patience and the Sexes 85
Education 87
Age 90
Summary of Demographic Analysis 92
The Waiting-for-Load Scan 93
Success of Links 94
Methodology 96
Overview of Methodology 96
Observed vs. Self-Reported Research Methodologies 96
Image Management 96
Memory Deﬁcit 97
Behaviors Done at a Subconscious Level 97
Comparisons of Free vs. Prompted vs. Scripted 98
The Treasure Hunt Syndrome 99
Interpreting Aggregate Heatmaps 100
Colored Regions 100
Purple X’s (G) 100
Red Lines (H) 100
Dotted Line (I) 100
Background Image 101
Interpreting Individual Searcher Session Images 102
Green Dot (A) 102
Circles (B) 102
White Halos (C) 102
Orange Lines (D) 102
Black Lines (E) 102
Numbers in Black Boxes (F) 103
Blue Lines (G) 103
Red X’s (H) 103
Red Dot (I) 103
Scrolling Pages and Following the Mouse 104
A Word About Sample Size 105
The Study Sample 106
General Demographic Breakdown 106



Foreword

At a recent networking reception, I ran into Geoff Ramsay from eMarketer. Geoff started by saying
that he liked the research we did and that he had included some of it in his collected reports. Then
Geoff asked what he warned would be a loaded question. “Is the research valid? Do you stand
behind the numbers?” As so often happens in these situations, I started to respond, someone else
joined our conversation and I don’t think I ever finished answering Geoff’s question. So, in a much
delayed and roundabout way, here you go, Geoff.

This will be the fourth research white paper that Enquiro has produced. In this process, we have In this study we found
created an approach that has refined itself in the past year. a number of things that
we think are indicative
Our goal is not just to throw a lot of numbers and study results at you. When we undertake a of a bigger, and all too
project, we have an idea of the types of things we’ll find. Sometimes we’re right on track, and vague, picture.
sometimes we’re completely surprised. But regardless of the findings, we try to tell you a story,
supported by the research findings.

In this study in particular we found a number of things that we think are indicative of a bigger,
and all too vague, picture. They start to provide more insight on how it is we search. As we start
to find these indicators, we often step beyond the “safe” area defined by the study findings and
speculate on what it is we might be seeing. We think that, as unscientific as it is, the true value of
our studies lies in this speculation.

So, to answer Geoff, in designing our studies we try to keep the methodology as sound as possible.
We stick to accepted methods of sampling and analyzing data. We work with recognized experts
in the field to ensure that there are no unintentional biases or inaccuracies embedded in the results.
So, yes, we stand behind our numbers.

But, we also try to look beyond the findings and sometimes stretch a little to “connect the dots”
using our growing base of insight into search behavior and our domain expertise to put forward
some scientific guesses. We think it’s what sets our research apart. In this study, we’ve tried to
indicate these areas with the prefix “Insight”. Anywhere you see this preceding a title, we’ve
moved a little beyond the data from the study and tried to interpret what that data means.

Hope that answers your question Geoff.

One last note. On behalf of the partners who helped create this research initiative, Kevin Lee
at Did-It, Colin Johnson and Greg Edwards at Eyetools, and Steve Alston and myself, Gord
Hotchkiss, at Enquiro: thank you for your interest.

5


Introduction
Since we first mentioned the Golden Triangle at Search Engine Strategies in New York at the end
of February, the image has been posted to dozens, or perhaps hundreds, of forums, blogs and
web sites. I have spoken of the Golden Triangle in San Francisco, Toronto, Orlando and London,
England since then, and most members of the audience had heard of it. For many, the image
confirmed what they had already intuitively known. For others, the intensity and concentration
of the scan patterns was surprising.

Although we have been The research, despite the fact that it was only conducted on Google (or perhaps because of it), has
doing research on how caught the attention of all the major search engines. Although we have been doing research on
people use search how people use search engines for a year and a half, this was the first time we had findings with
engines for a year and such visual impact.
a half, this was the first
time we had findings We have fallen in love with eye tracking technology and plan to continue down the path we
with such visual started.
impact.
In this study, we think any search marketer, or even developers of search technology, will gain
insight that will make a significant impact on their search strategies. Not since our original white
paper, Inside the Mind of the Searcher, have we been afforded such a glimpse into the mental
processes that govern search sessions. In this paper, you’ll learn about the complex interplay
between page elements such as Top Sponsored links, OneBox results (Google’s News, Product or
Local listings that appear just above the organic results), organic results and Side Sponsored ads.

You’ll learn how our eyes move around the page in excruciating detail, how we scan listings,
what catches our eye and what motivates us to click on one listing over another. Is position
everything? If we don’t find what we’re looking for right off, do we lose confidence in the
listings? Does the eye move differently the first time it sees a results page versus its return to
the same page? How many listings do we look at before we decide? Do different demographic
groups scan differently? These are just a few of the questions we try to answer.

Like any research project, it will answer some questions, but likely will raise more. This is okay,
as long as it helps push the boundaries of our knowledge of how people search. We feel confident
that this white paper accomplishes this.

One last note, and this is very important to get the full value of this paper. We make extensive
use of images from the study, both aggregate heat map images that show the cumulative activity
from several sessions and individual page session images. There’s a bit of a trick to interpreting
these images. Before you go too much further, we highly recommend reading Interpreting
Aggregate Heat Maps and Interpreting Individual Search Sessions.



Importance of the Golden Triangle

Since we first talked about the appearance of the Golden Triangle, there has been much discussion
in forums and blogs about how this confirms the importance of being number one. In reality,
as we looked more at the individual behavior of searchers and how they interacted within this
area, we saw that the reasons that led to the formation of the triangle were more complex and
interdependent than the simple conclusion that position is all important.

What Is Google’s Golden Triangle?

The Golden Triangle is a distinct area of intense eye scan activity that is shown in the diagram
below. It’s important to understand that the Golden Triangle pattern is seen in first time visits to a
results page. We’ll explain this more in the Search Confidence section.

The Golden Triangle

If your listing is not in
the Golden Triangle,
your odds of being
seen by a searcher are
dramatically reduced.

Source: Eyetools, Inc
Generally speaking, if your listing is not in the Golden Triangle, your odds of being seen by a
searcher are dramatically reduced.

7


Let’s carry this offer to a bricks-and-mortar analogy. Think of the search results page as a
shopping mall. Think of the eye activity as foot traffic. In a mall, you have anchor tenants who
attract the majority of traffic. Usually, malls try to have two or three anchor tenants distributed
evenly around the perimeter of the mall so foot traffic is generated moving from anchor tenant
to anchor tenant. All the other tenants take advantage of this by catching the attention of the foot
traffic as they walk by.

The (search page) The Google search results page has one anchor, and it’s usually in the upper left corner. The
anchor is the number anchor is the number one organic result. That’s what we’re all looking for. Everything else is a
one organic result. detour on the way. Yes, Top Sponsored results get high visibility and a reasonable number of
Everything else is a click throughs, but more often than not, it’s because they happens to catch our attention while we
detour on the way. are looking for the top organic listing. With rare exceptions, we’re not looking for a sponsored
result.

Let’s go back to our analogy. What would happen if a mall had only one anchor in a corner of the
mall that every single shopper went to see? You would see a flurry of foot traffic in this corner
of the mall and little everywhere else. The further you got from the anchor, the less traffic you
would see. That’s the Golden Triangle. And in this case, the anchor is very easy to see. It’s firmly
established in the upper left corner of the search results page. That’s where our eyes first go to see
the top organic listing. We may have to reorient ourselves from that point, but that’s where we’ll
start.

The Golden Triangle and the Google Effect

In looking at the Golden Triangle, there’s an important point to keep in mind. The Golden
Triangle exists more because of the effectiveness of Google than of the searcher’s preoccupation
with position. Let us introduce you to the Google Effect.

One of the advantages we had that led to us to being able to separate out the importance of
position from other factors was a study by the University of Tulsa that was being conducted at
the same time as the eye tracking study. In this study, we did a structured conjoint analysis of all
the factors that led to searchers clicking through on a particular listing in a given page of search
results. We looked at position, presence of trusted brands, presence of trusted URLs, relevant
words from the search queries in the title, relevant words from the search query in descriptions
and the presence of words like “review” and “consumer information”. In the end, the number
one factor in precipitating a click through on a particular listing was relevant words in the title,
with relevant words in the description being second most important. Position came in after these
in significance.

So why the Triangle? Why the overwhelming number of click throughs on the number one
listings in the two different studies we conducted (this Eye tracking Study and an October 2004
study, The Role of Search in B2B Buying Decisions)? Obviously, position is important.

It appears that the predominance of click throughs on the number one listing comes more from
a conditioned behavioral response due to a natural pattern in how we assimilate information



presented in a textual format and the actual background functionality of the engine than from the
fact that we always click on the number one listing, no matter what. Dr. Akhilesh Bajaj, one of the
authors of the University of Tulsa study, calls this the Google Effect.

Here, in layman’s terms, is a quick explanation of the Google Effect. Google is relatively good at If Google is doing its
getting the best listings to the top. Over the time we have used Google, we have come to realize job properly, there’s a
this. So in the beginning of our search session, with our confidence level high (see Confidence high likelihood that the
with Search Results) we restrict our scan activity to the area of the page most likely to have the most relevant result
best listings. This is the Golden Triangle. We start reading information in this area with a familiar will be the first one we
upper left orientation. We scan listings and the majority of us (about 72 percent according to the see.
Business to Business study done previously) click as soon as we see something of interest. So, if
Google is doing its job properly, there’s a high likelihood that the most relevant result will be the
first one we see. We’re not clicking it because it’s number one. We’re clicking it because Google
has done its job just as we expected it to and put the best result in the first place we’ve been
conditioned to look. We still take the time to scan the title and sometimes the description to ensure
it is relevant, and more often than not, it is.

Proof that this is taking place can be seen by looking at what happens on the second visit to the
same results page (see First Visit vs. Repeat Visits). If position were the only important factor, we
would keep seeing the Golden Triangle pattern repeated, just shifted lower than it was before. But
the fact is, the Golden Triangle pattern breaks down completely on the second and subsequent
visits. The Google Effect has been broken and our confidence that Google will put the best result at
the top is gone. Suddenly, position becomes a negligible factor.

In summary, 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. This distinction is important to
understand.

9


Interaction within the Golden Triangle

General
The Golden Triangle represents the aggregation of activity we see from the first visit to a search
results page. It shows the entry point of the eye, the subsequent reorientation and then the typical
scan activity that results.

We’ll look more closely at the typical eye activity we saw in this section of the results page
throughout the study.

Entry Point

We saw a remarkable In this study, we saw a remarkable consistency in the entry point of the eye. This at first puzzled
consistency in the entry us, until we overlaid the Google home page and saw that the entry point lined up with the button
point of the eye. you click to launch your search. In our study, all searches were launched from Google’s home
page, rather than from a tool bar.

For monitors with 1024 X 768 resolution (the resolution used in the study) this puts the entry
point just left of center and slightly down from the first result, when there are no top paid results
or OneBox.

 Faded image shows
location of home
page search buttom,
superimposed over
search results

 Hot spot shows
entry point of eye
which was looking at
this button


If searches are launched from Google’s home page, this entry point becomes a matter of some
significance. If there are no Top Sponsored ads shown, this spot is near the bottom of the first
organic listing. If one sponsored ad is shown, it’s just under the first ad. If two are shown, it
usually ends up between them.



In any case, this means that when Top Sponsored ads are shown, there is a great likelihood that
the eye will start its journey within this section if the search is launched from Google’s home page.
We didn’t test entry points from searches launched from a search tool bar, so we don’t know if this
persistence of position would hold true, or if the eye would reorient itself during page loading.

Reorientation
Upon page loading, the ﬁrst thing the eye does is reorient itself. The common pattern is a quick
upswing to the top left of the page. Google has a light blue bar that separates the search results We use this bar as a
from the above navigation. It seems we use this bar as a visual reference point and move to just visual reference point
below this on the far left side. This is the position in which we expect to see the ﬁrst result. It seems and move to just below
that most eyes take a quick orientation swing using peripheral vision to locate a visual reference this on the far left side.
point like the blue bar to anchor themselves for further scanning.

 Shows using the lighta
blue bar at the top as
visual orientation point,
from which the scanning
activity starts.


11


First Fixation of the Eye

In general, the first fixation of the eye was as follows on the first visit to a page of results (we
haven’t included initial fixations from return visits, as the patterns are significantly different and
are discussed in First Visit vs. Repeat Visits).

All First Visits to a Page – 199 Sessions

First Fixation in Top Organic 38.7%
First Fixation in Top Sponsored 43.2%
First Fixation in Side Sponsored 1.0%
First Fixation in Google OneBox 14.6%
Other 2.5%

However, the breakdowns aren’t as simple as they appear in the above table. Top Sponsored and
OneBox results didn’t appear for every search, so we’ve broken them down to account for these
variables.

No Top Sponsored Present, No Google OneBox – 41 Sessions

In this case, fixation points were consistent through all participants. Every eye anchored in the
upper left on the first organic listing.

First Fixation in Top Organic 100%

Top Sponsored Present, No Google OneBox – 63 Sessions

When Top Sponsored ads When Top Sponsored ads were present about 57% of the visitors fixated on the Top Sponsored
were present about 60% ads. The remainder, 38%, fixated instead on the top Organic result, without really looking at the
of the visitors had their Top Sponsored. Only one person went straight to the Side Sponsored ads.
first fixation in this area.
Other 4.6%

No Top Sponsored but Google OneBox Present – 24 Sessions

Again, 54% of the visitors fixated on the Google OneBox results, with the remainder, 33%
scanning down and fixating on the top Organic listing instead. Again, one person went directly
over to the Side Sponsored results.

First Fixation in Google OneBox 54.2%
Other 12.5%

Top Sponsored and Google OneBox Present – 72 Sessions

The results here were interesting to compare. The chances of somebody immediately finding and
fixating on the top organic listings drops dramatically when more Golden Triangle real estate is
taken up by Top Sponsored and OneBox results. Only 10% of the visitors managed to do it. The



majority, almost 70%, fixated in the Top Sponsored ads. About 21% fixated in the OneBox results.
The implication seems to be that even if you intend to go straight to the organic results, the more
that appears above them, the more chance there is that you’ll be distracted.

First Fixation in Google OneBox 20.8.%
Other 2.7%

Conclusions Regarding First Fixation Point
This split second decision that determines where the first fixation point will occur is essential in
laying the path for what will follow. The first fixation point of the eye has a strong correlation to Once the eye anchors
the link eventually chosen. Once the eye anchors in a section of the results page, the chances that a in a section of the
link will be chosen in that section go up dramatically. results page, the
chances that a link
It’s important to note that when no other top results were shown, 100% of the participants kept will be chosen in
their fixation point in the top organic, indicating that this is the section they intended to find. that section go up
But when either Top Sponsored or Google OneBox results appeared, we saw that a significant dramatically.
percentage of participants quickly scanned past these to the top organic results. This lends further
credence to the view that the intended target for many searches is the organic results and any time
spent in other sections is the result of a diversion from the intended activity.

Comparison of First Fixations and First Clicks

Areas of First Click Top Top Side Bottom Other3
Sponsored Organic1 Sponsored Organic2
Areas of First Fixation

Top Sponsored 22.1% 0.0% 37.2% 7.0% 9.3% 22.1%
OneBox 17.8% 14.3% 50.0% 3.6% 3.6% 10.7%
Top Organic 6.8% 1.3% 44.9% 0.0% 10.8% 29.7%
Side Sponsored 0% 0% 0% 66.7% 0% 33.3%

1 Includes top 4 organic listings
2 Includes all organic listings lower than 4th place
3 Includes going to next page, new searches and other choices, e.g. clicking on image search or Froogle.

13


First Significant Scan Activity
After recording the first fixation point, we then looked for significant scan activity. The
distinction between the two is important. A fixation point is simply a split second stop on the eye
path. It indicates a quick orientation check, often on a visual reference such as a bolded search
query or even a bit of white space between listings. It doesn’t indicate a cognitive assimilation of
information.

A fixation point is
simply a split second
stop on the eye path.

 First Fixation Point, showing a momentary stop in the eye’s path

 First significant scanning, as shown
by the orange lines. This indicates
reading activity

Scan activity, on the other hand, indicates the familiar left to right reading of text. In our
individual session images, this reading is indicated by orange lines.

We broke down the first significant scan activity in two different ways. First of all, an overall
view of first scan activity across all sessions, and then a comparison between where the first
fixation point occurred and where the first significant scan activity occurred. This allowed us
to explore the importance of first fixation points as an indicator for where searchers start their
activity.

Overall Significant Scan Activity – 199 Sessions

Location First Fixation First Scan Activity
OneBox 14.6% 14.6%
Side Sponsored Ads 1.0% 2.0%
Top Organic 38.7% 46.7%
Top Sponsored Ads 43.2% 31.2%
Other 2.5% 5.5%



 Whilefirst fixation point
their
more people had

 in the top sponored
ads, a number of
participants moved
down to the top organic
to start scanning
results


As can be seen, in actual significant scan activity, top organic reverses the trend from first fixation
point and edges out the Top Sponsored, indicating that although some visitors first looked at these
Top Sponsored ads, they quickly moved down to top organic before actively reading listings.

Finally, let’s look at the preceding table and extend it to see where the actual clicks ended up
occurring. To see the direct correlation between areas first scanned and links clicked, check the
table below:
Location First Fixation First Scan Activity First Click

OneBox 14.6% 14.6% 2.5%
Side Sponsored Ads 1.0% 2.0% 4%
Top Organic 38.7% 46.7% 42.2%
Top Sponsored Ads 43.2% 31.2% 13.5%
Other 2.5% 5.5% 37.5%4

4
In this case, Other includes bottom organic (lower than top 4), hitting next, launching a search and
following any other navigation links.

As can be seen from the chart below, there is a strong correlation between where searchers start
Only 7.5% of the people
reading and where they eventually click. While there are always people moving from other areas
who started scanning
to the top organic listings (46.7% of the people who started scanning in the Top Sponsored ads
in the organic listings
ended up choosing an organic link), there isn’t much movement the other way. The majority of
actually ended up
people who start scanning in the organic listings either choose an organic link (71%), click through
choosing a sponsored
to the next page, revise their search or take some other action (21.5%). Only 7.5% of the people who
link.
started scanning in the organic listings actually ended up choosing a sponsored link.

Areas of First Click Top Top Side Bottom Other7
Sponsored Organic5 Sponsored Organic6
Areas of First Scanning

Top Sponsored 25.8% 0.0% 37.1% 6.5% 9.6% 21.0%
OneBox 11.5% 15.4% 46.2% 3.9% 0.0% 23.1%
Top Organic 7.5% 1.1% 52.7% 0.0% 18.3% 21.5%
Side Sponsored 0% 0% 25% 75% 0% 0%
5
Includes top 4 organic listings
6
Includes all organic listings lower than 4th place
7
Includes going to next page, new searches and other choices, e.g. clicking on image
search or Froogle. Cont’d on Page 17

15


Insight: What You See Is What You Click

It’s not really surprising that there’s a strong correlation between where eyes tend to go and where
clicks are made. There is a certain amount of common sense here. But there is a factor at play that
should be expanded on, as it provides further insight into the nature of our interaction with a
search engine results page.

The reason there is such a strong correlation is that we make very fast decisions on a search results
page (See Blink, Thin Slicing and the Art of Search). Especially on the first click, we don’t give
it a lot of thought before we choose a listing. Decisions are made quickly and without a lot of
deliberation for the majority of us. In the study, The Role of Search in Business-to-Business Buying
Decisions, a full 72% of respondents said they clicked on the first listing that seemed to offer what
they were looking for. Because of self-reporting issues in these types of surveys, we suspect the
actual percentage is even higher.

Searcher Behavior While Clicking on Links Percent
click on first link of interest 72.0
read all listings first, then decide 25.5
Valid
none of the above 2.5
Total 100.0

The Role of Search in B2B Buying Decisions, Enquiro, October 2004

We spend less than In our study, we found that an average of just under seven seconds elapsed before a selection of a
two seconds per listing listing was made in the first visit to the search page. In that time, an average of almost four listings
actually reading what were scanned. So, we spend less than two seconds per listing actually reading what it is that listing
it is that listing has to has to offer us. Not a lot of time here for deliberate thought or rational decisions.
offer us.
So, with decisions made this quickly, and with the overwhelming majority of us clicking the
first time we see a listing that catches our attention, without considering other listings first, the
importance of where eyes first fixate on a search results page becomes quite clear. If we know
where eyes go first, being in this section of the page and providing the factors that play a part in
motivating a click-through becomes essential to our search marketing strategy.



Again, as with fixation points, to get the true picture, we have to break down scan activity further
depending on whether Top Sponsored ads or OneBox results appeared. To give some comparison,
we’ll also include the percentages from first fixations.

It’s interesting to note
the tight, compact
shape of the Golden
Triangle when there are
just organic results in
this area.


No Top Sponsored Present, No Google OneBox - 41 Sessions

Top Organic 100% 100%

As with fixation points, everyone went to the top organic results and started reading there.
It’s interesting to note the tight, compact shape of the Golden Triangle when there are just organic
results in this area. As we’ll mention again, the triangle shape is definitely stretched vertically by
the appearance of OneBox and Top Sponsored results. Also note the dominance both in visibility
and click-throughs of the number-one organic listing.

17


Top Sponsored Present, No Google OneBox – 63 Sessions

When Top Sponsored
ads were present,
about 60% of the
participants first
fixated here.


When Top Sponsored ads were present, about 57% of the participants fixated on the Top
Sponsored ads. The remainder, 38%, fixated instead on the top Organic result, without really
looking at the Top Sponsored. Only one person went straight to the Side Sponsored ads.

Top Organic 38.5% 54.0 %
Top Sponsored 56.9% 34.9%
Other 4.6% 11.1%



No Top Sponsored but Google OneBox Present – 24 Sessions

About 54% fixated in
the OneBox results
(when Top Sponsored
results weren’t
present).


About, 54% of the visitors fixated on the Google OneBox results, with the remainder, 33% scanning
down and fixating on the top Organic listing instead. Again, one person went directly over to the
Side Sponsored results.

Top Organic 33.3% 25%
OneBox 54.2% 58.3%
Other 12.5% 16.6%

19


Top Sponsored and Google OneBox Present – 72 Sessions

When more of the
Golden Triangle is
taken up with Top
Sponsored and OneBox
results, the chances of
the first scan activity
happening in the
top organic lessen
dramatically.


As we mentioned in talking about fixation points, when more of the Golden Triangle is taken up
with Top Sponsored and OneBox results, the chance of the first scan activity happening in the
top organic area lessen dramatically. In this case, about 12.5% of the visitors did move from their
first fixation point to the organic listings before starting to read. This came mainly from the Top
Sponsored ads.

Top Organic 9.7% 22.2%
Top Sponsored 69.4% 55.6%
Google OneBox 20.8.% 19.4%
Other 2.7% 2.8%



Where the Clicks Happened

Second & Later Where the Clicks
Listing Clicked Overall First Visit
Visits Happened:
OneBox 1.96% 2.6% 0% 53.6% Organic
27.6% Other
Organic 1 26.0% 32.3% 6.5% 11.7% Top Sponsored
Organic 2 4.7% 3.6% 8.1% 5.1% Side Sponsored
Organic 3 5.9% 5.7% 6.5% 1.96% OneBox
Organic 4 5.5% 6.2% 3.2%
Organic 5 3.9% 2.1% 9.7%
Organic 6 1.6% 1.0% 3.2%
Organic 7 2.8% 2.1% 4.8%
Organic 8 1.2% 1.6% 0%
Organic 9 0.39% 0.52% 0%
Organic 10 1.6% 1.6% 1.6%

Total Organic 53.6% 56.7% 43.6%
Total Top Organic (1-4) 42.1% 47.8% 24.3%
Total Lower Organic (5 – 10) 11.5% 8.9% 19.3%

Side 1 1.96% 1.6% 3.2%
Side 2 0.39% 0% 1.6%
Side 3 0.79% .52% 1.6%
Side 4 0.39% 0% 1.6%
Side 5 0 0% 0%
Side 6 1.2% .52% 3.2%
Side 7 0.39% .52% 0%

Total Side Sponsored 5.1 % 3.16% 11.2%

Top Sponsored 7.8% 9.4% 3.2%
Top Sponsored 2 3.9% 4.7% 1.6%

Total Top Sponsored 11.7% 14.1% 4.8%

Other (next, new search, etc.) 27.6% 23.4% 40.4%

21


Interaction with Top Organic

There are many variations of how participants reach the top organic results, but once they get
there, there tends to be a similarity in how they interact with them. They start at the top, looking
for visual anchors (see How We Scan a Listing) and when one catches their attention, they
quickly read the title. Then, if interested, they will pick up additional detail from the description,
either through active reading or by picking up detail through their peripheral vision.

They start at the top,
looking for visual Entry point (from Google Search Button), then
anchors, and when one visual orientation using blue sponsored box
at top. They use this to navigate to the top
catches their attention, organic listing, where they start scanning

they quickly read the 
title.




F Scan Patterns

In our original release about the Golden Triangle, we said that these patterns generally resembled
an F-shaped scan pattern. Starting in the upper left corner, we scan down vertically and when
titles catch our attention, we start reading the title horizontally. If we choose not to click and
continue to look at other listings, we begin to scan down vertically until we see the next listing that
catches our attention, again switching back to a horizontal scan. This repeated pattern tends to
create a series of F’s down the page.

This repeated pattern
tends to create a series
of F’s down the page.

Most of the vertical Movement tends to be along the left
side of the page, slightly in from the far left side.


 When a visual cue catches the eye, lateral scanning occurs

 This continues down the page in a
repeating F shaped pattern

 Finally, when a listing appears
to offer what’s needed, a
clickthrough occurs


The vertical scanning down tends to be aligned to the left side of the page, and is near the
beginning of the title, although this path often seems to be slightly right from the actual start of the
title.

23


Some searchers appear to be quite methodical in moving down the page.

Some searchers
appear to be quite
methodical in moving
down the page.

In this case, the person moved
methodically down the page to
the bottom, then looked at the
top Side Sponsored ad, scrolled
down and resumed a methodical
but less sequential scan of the
bottom listings.




Others show this pattern to the bottom of the fold, also the bottom of the Golden Triangle, and
then bounce back to reread listings more carefully before making their selection.

Others... bounce back
to reread listings more
carefully before making
their selection.
Here we see a methodical scan
of the page to the bottom of
the visibile results, than a move
back to reread the OneBox and
top Sponsored Listings.


25


Interaction with Lower Organic

Interaction with the lower organic listings is not signiﬁcantly different from the F scan pattern
described before. Generally, this process occurs on a return visit to the page, when the searcher
continues down the page until a listing is found that attracts enough attention that the searcher
decides to click on it. Again, there seem to be a few ways this is done.

First, there’s a quick scan down the page, quickly clicking on the ﬁrst link of interest.

This process occurs on
a return visit to a page,
when the searcher
continues down the  On returning to the to the the searcher
page,
page until a listing is quickly scans down
page they want to scan
part of the

found...

 Then active scanning starts on listings
of interest




Secondly, there’s a two-step scan, where listings are scanned in groups of three to ﬁve (generally
whatever appears on one page without scrolling down), titles are quickly scanned and then the
searcher returns to the listings that catch their attention to read further.

Listings are scanned
in groups of three to
ﬁve..and the searcher
returns to the listings
that catch their
attention..

 A very quick visualgoing back to one3in
or 4 listings, then
scan down over
 this group to actively read


27


Finally, there’s a more deliberate and linear scan, working their way down the page and reading
each title in turn. Some will read all listings visible before making their decision, and some will
read deliberately but will click on listings of interest as they encounter them. The difference in
this case is that this interaction with the lower organic listings often happens on the ﬁrst visit to
the page.

There’s a more
deliberate and
linearscan, working
their way down the
page and reading each
title in turn.




Impact of OneBox

Google’s OneBox results introduced an interesting variation in general scan activity. They seemed
to impact the overall patterns in a couple of signiﬁcant ways.

Google’s OneBox
results introduced an
interesting variation in
general scan activity.

Aggregate Heat Map of Activity from all Sessions with OneBox results Present

First of all, there seems to be a genuine curiosity about OneBox results. We did see a fare amount
of scan activity in this section when they were present. If there was scan activity at all in this
area, which there was in about 22.5% of the searches, the participants spent an average of 2.12
seconds actively reading the results. Although this might not sound like much, compare this with

29


an average of 2.04 seconds actively scanning the Top Sponsored ads or an average 3.68 seconds
actively scanning the top three organic listings. Remember, the OneBox results are usually two
or three single lines, so both the Top Sponsored and the top organic results would usually offer
more text to scan.

This was particularly true when the results shown were product results from Froogle, Google’s
shopping engine. It seems that searchers are beginning to recognize the fact that results can offer
interesting information.

As we look more at what attracts attention on the search results page, it appears that OneBox
results have even more than this going for them. There’s an icon presented that acts as a natural
visual anchor for the eye (for more on the impact of graphic icons, see Impact of Bolded Search
Queries and Icons ). Then, in the case of the product search, there’s another proven visual anchor,
the bolded search query. Finally, there are often recognized brands presented (see The Role of
Brand in Search). All these are proven magnets for attracting the eye.

In the case of a product search, there’s yet another eye attractor. Prices. We saw a couple of
instances where in a speciﬁc product search, the searcher would take a quick scan over the
results, keying in on both the product description and the prices.

It seems that searchers
are beginning to
recognize the fact
that these results
can offer interesting
information.
The orange line indicates a scan of
 prices in OneBox results


In this study, the scenarios we presented put the searchers more in a product research mode
than in a purchase mode. It would be fascinating to do further research to see if this presentation
of prices and quick product information would prove even more compelling when in purchase
mode.

Despite attracting a reasonable amount of attention, the OneBox results didn’t fare nearly as well
on the click-throughs. As we mentioned before, for all searches launched in the study, about
22.5% of them included some scanning activity in the OneBox. But only ﬁve people, or 2.3% of



the total, actually clicked on a OneBox link. In looking at the success rate of these clicks, not one
of them led to a page where the searcher found what they were looking for. In every case, they
clicked back to the search results page and tried again. Lots of people are looking, but not many
are clicking.

It was not apparent why OneBox click-throughs weren’t higher, but we’ve heard some potential
reasons anecdotally. A Google product manager that ventured the opinion that many searchers
may think OneBox results are sponsored ads and so are a little reluctant to click. We also know
that people tend to be creatures of habit online and are leery about clicking into the unknown.
Perhaps it was not just aversion to potential advertising, but also confusion about OneBox’s
purpose that suppressed click-throughs.Finally, in looking at the success rate of the clicks that did
happen in the OneBox, maybe it was simply that searchers didn’t trust the quality of the results
they’d ﬁnd there.
We suspect that the
OneBox is yet to come
Regardless of the reason, we suspect that the OneBox is yet to come into its own and will
into its own and will
eventually play a bigger part in our online search activity. It’s the key to increased verticalization
eventually play a bigger
for Google down the road and increased search awareness and smoother integration will both
part in our online
lead to increased importance for OneBox results. As a point of interest, click-throughs on OneBox
search activity.
results were signiﬁcantly higher among men and study participants with higher education. Again,
we suspect click throughs would be substantially higher if the scenarios presented were more
geared to purchase activity.

31


Impact of OneBox on Scan Activity

Now let’s look at how the appearance of another piece of search real estate affected general
search activity.

Whether by intention or happenstance, Google has placed their OneBox in the prime real estate
on the page. We suspect Google knows exactly what they’re doing. When present, Google
OneBox is in the prime section of the Golden Triangle. So, what does this do to the shape of the
Triangle?

In looking at the difference between aggregate activity with OneBox present and with it not
present, we see that the Triangle stretches itself to accommodate the OneBox. The reason is clear
when we start looking at the actual scan activity. In the above section, we mentioned that 22.5%
of searchers spent some time looking in the OneBox. Let’s further break out the entire picture of
OneBox activity.

OneBox not presented 49.7%
OneBox presented but not scanned 27.8%
OneBox presented and scanned 22.5%

Whether by intention
or happenstance,
Google has placed
their OneBox in the
prime real estate on
the page.

Notice how the
appearance of
OneBox stretches
the Golden Triangle
vertically.




As can be seen, although 22.5% of participants scanned the OneBox results, 27.8% of the
participants quickly saw them and skipped right past them. Because of the way the aggregate
heat maps are compiled, this would still show eye activity but may be misleading when trying to
identify significant scan activity in this area.

Because the heat map images shown can be misread (see Interpreting Aggregate Heat maps) let us
offer some guidance in looking at these ones in particular.

Because the OneBox appears in the highest traffic section of the page, there was a significant When OneBox results
percentage of visitors who fixated there first. When OneBox results were present, 21.9% of were present, 21.9% of
participants had their first fixation in this area. Further, 21.1% remained in the area and had their participants had their
first scan activity there as well. From there, they tended to go down to the organic results, or, in far first fixation in this
fewer cases, up to the Top Sponsored results. area.

For those who had their first fixation in the Top Sponsored ads, when they scanned the OneBox
results, it was a quick pass-through to the top organic results. Only 4.8% of this group did their
first significant scanning in the OneBox results. While this showed as eye activity in the aggregate
heat map, it was not significant in most cases.

When the first fixation happened in the Organic listings, in 6.7% of the cases we saw a visual swing
up into the OneBox results before significant scanning started.

When looking at the impact of OneBox’s impact on scanning, it broke down into two factors.
When these results were present, about a quarter of our group spent some time actively looking
at them before moving on to another section of the page. The remainder tended to go to either
Top Sponsored or top organic sections of the page. While there may be a few fixation points in the
OneBox results, there’s little active scanning there. So while OneBox results are in the middle of
the Golden Triangle and appear to have high visibility, this is only really true for one out of four
visitors. The rest of the activity comes from these results being sandwiched between the two most
popular parts of the page.

33

Eyetracking study

Recommended

Recommended

More Related Content

Similar to Eyetracking study

Similar to Eyetracking study (20)

More from WSI Egypt

More from WSI Egypt (15)

Recently uploaded

Recently uploaded (20)

Eyetracking study