From: Norman J. Jacknis (norm@jacknis.com)
Date: Tue Jul 11 2000 - 22:20:21 EDT
Another interesting story about the use of the Internet. I would recommend going to the site referenced in this summary -- http://www.poynter.org/eyetrack2000/index.htm
Regards,
Norm
----------
It is one thing to look at the statistics that a website normally
gathers, such as where traffic comes from and what pages are more
popular, but in terms of tracking audience behavior across
websites, especially as links are followed, the numbers are
hard to come by.
This has some bearing as to whether people are looking for
specific information that informs a specific personal
need, or are looking for ways to access a broad range of
sources and perspectives; and whether people look for
information outside of established, or familiar, news entities.
The May release of the initial findings from a recent study by
Stanford University and the Poynter Center (http://www.poynter.org),
a Florida education and research group which focuses on
journalism, attempt to fill a portion of this research gap.
What did they find? Try this:
-- People look at text before they will look at photos or graphics
on a page. If people do look at graphics, it is online after
they have returned from reading the full text of an article,
even if it is on the same page;
-- People look for headlines and news summaries more than they
do articles, with some three times as many briefs viewed than
articles. People do, however, read at least 75% of an online
article's content compared to the 30% of a print article's
content;
-- Online readers are looking to news sources for both general
and personal concerns, and to examine other perspectives on
news, as evidenced by the fact that 10% of the subjects made
a concerted effort to read foreign news sources for additional
insights on news items, and that 58% of news summaries and 97%
of articles opened were opinion pieces (including columnists,
letters to the editors, and editorials);
-- Internet advertising actually catches the eyeballs of users,
with some 45% of banner ads catching a reader's attention for
a full second 22% of non-advertising graphics are actually
looked at, for at least one second 64% of photographs are looked
at for about 1.25 seconds
We know, we know. You're thinking, great, one second does not speak
well for the attention span of online users. But consider what one
second means as far as online content is concerned. To do this,
we'll jump back to the origins of this survey.
Though the Poynter Center has conducted an "Eyewatch" survey for
about ten years that studies the reading habits and patterns of
print media audiences, this recent study started four years ago
under Stanford University communications profess Marion Lewenstein,
as an effort to videotape people who read news online at home and
work. At the time, researchers found that:
-- users read news in 30 minute blocks of time
-- people gravitated towards the news sources they were familiar
with and/or utilized in the real world
-- customized news services that delivered clippings were not popular
-- online users still read, watched, or listened to the traditional
print and broadcast versions of news sources
Two years ago, the researchers wanted to learn what kind of
information users were paying attention from online news sources,
and where their attention was focused most on certain pages as
they followed the links in news stories. So they decided to look
at they eye movements of users.
For this survey, tiny cameras affixed to headsets monitored the
eye movements. The software used allowed the study to actually track
not only general surfing habits, but also recorded the eye movements
of each user into a database.
A little background is in order. You might recall that
the smallest amount of information on your computer is a bit (short
for binary digit). One single bit has only one of two values, a "0"
or "1". When you combine a sequence of bits into larger units, you
can yield more complex types of information, such as "bytes" which
is a string of 8 consecutive bits.
When you hear computer folks mentioning things like "32-bit
computers," they are referring either to the number of bits that
can be processed at a given time (more bits processed in a certain
amount of time makes a machine faster) *or* the number of bits it
takes to define a portion of a machine's memory (the larger the size
of memory spaces on a machine means that it can process larger and
more complex programs). You may also see the term bit used when
referring to graphics. In this context, it means how many bits
are used to define one dot in an image. So, for example, 32-bit
graphics can represent more accurate colors, while 1-bit graphics
give you one-color images.
The human brain, by comparison, can process roughly one bit of
visual information in 1/10th of a second, for the purposes of the
study, this was counted as an "eye fixation." Think about it: you
are processing 10 bits of information a second as you scan a
news headline or a full web page. The researchers hoped that by
looking at where eyeballs stopped on a page, they would have a
better idea of what specifically people were reading and processing.
For the survey, researchers selected 67 subjects from two cities,
Chicago, Illinois and St. Petersburg, Florida. The subjects spent
a combined total of 40 hours viewing some 6,000 pages from 211
unique news sources, yielding some 608,063 eye fixations and
24,530 mouse clicks. Subjects were asked to look only at news
websites, although they could look at whatever sites they wanted,
for as long as they wanted, as many times as they wanted. News
sources examined were broken into seven categories:
-- National (e.g. New York Times and Washingtonpost.com)
-- Local (with respect to the research locations)
-- Customized (e.g. MyYahoo or MyExcite)
-- Hometown (those covering a specified local geographic area of interest)
-- Specialty (e.g. ESPN or ZDNet)
-- Foreign (like the BBC, Candian, or media from other countries),
-- Non-Traditional (e.g. TheSmokingGun.com or National Enquirer).
Each test subject, on average, viewed up to 6 news provider
websites during individual web viewing sessions of around 34 minutes.
If you count each subject's visit to a news site once, that's about
426 different visits to 211 news providers. If you include multiple
visits to a site, the subjects viewed the 211 sites some 610 times
total.
The longest amount of time any user spent on a specialty/non-standard
news provider website was around 20 minutes, versus a maximum of
46 minutes on a general news provider site (this includes the
National, Local, Hometown, and Customized categories above,
encompassing breaking news and content on national/international,
local, sports, entertainment, business, etc.)
One-third of the news sites viewed were those of mainstream
news providers (national and local broadcasters and or sites
providing customizable news feeds). Some 59% of all visits
(including revisits) were to the mainstream general news sites,
which were also the sites most likely to be accessed first
(79% of all users), revisited (by 64% of users) and for the
longest period of time. About 40% of participants started with
the website of their local newspaper or broadcast station.
Interestingly, the most revisited sites tended to be "hub" sites
that redirected people to other sites, and not "portals" that
sought to draw in users by providing comprehensive information.
Users liked the ability to leave and revist a web entity's site in
order to feel as if they were literally experiencing another
perspective. In post survey interviews, the reasons users had for
leaving and revisiting a particular news site were:
-- following a content-related link to another article but wanting
to return to the source
-- hunting for related information not available as a hypertext link
from the website
-- taking advantage of features like on-site polls or
forwarding information to others after reading an article
As these are only the preliminary findings, the researchers are
still combing through a large stack of information. But the
findings reinforce both conventional wisdom and common sense. There
is, for example, a difference between looking at something online
and actually reading and processing content.
More importantly, because text appears on a page before graphics,
you basically have a better chance of grabbing someone's attention
with well written text than multimedia content. Also, if people look
at text on a page before they look at photos or graphics,
consideration needs to be paid to quality of an Internet connection
or good web page design when looking at how to get your message
across.
People also look for branding online as much as content. Reputation
as a reliable, if not trusted news source, matters just as much online
as it does offline. If you, therefore, are looking to guide a
user's online experience through links, banner ads, etc. you must
ensure that your name or organizational identity is firmly, clearly,
and consistently established, whether it be the name of the
organization placed on each animated frame of a banner ad, or anchored
on a web page you create.
More information on the study can be found on the Stanford Poynter
Project website (http://www.poynter.org/eyetrack2000/index.htm).
Ryan Turner
NPT Project
OMB Watch
This archive was generated by hypermail 2.1.2 : Fri May 31 2002 - 23:55:01 EDT