Getting the User

1
Getting the Users AttentionThe Effectiveness
of Two Mediums of Online Advertising

• David Orr
• Steve Pautz
• Chris Coker
• Katherine Hinds
• Mandy Elkins
• dorr, spautz, clcoker, hindsk,
  elkinsa_at_clemson.edu

2
Abstract

• Study to determine which type of internet ad is
  more effective at capturing user attention
• Static, animated
• Participants engaged in a visual search of a web
  page
• Eye tracker data used to determine fixations
  within the banner ad
• Recognition task was given after the web page to
  determine how well the adds promoted recognition
  memory

3
Previous Research Strayer, Drews, and Johnston

• Participants drove simulated highway talking on
  cell phone or not
• Cell phone conversation impaired memory for
  billboards
• Eye tracker showed 2/3 of billboards fixated on
  during experiment
• Indicates failure of divided attention

4
Previous Research Benway and Lane

• Study concerning banner objects
• Questions could either be answered using text
  links or banners
• Text link questions answered 94 of the time,
  banner questions answered 58 of the time
• Altering grouping of banner did not change
  results
• Banner Blindness

5
Previous Research Yantis and the New Object
Hypothesis

• Salient events like motion often thought to
  capture attention
• User presented with large letter comprised of
  smaller letters. Asked to name identity of the
  larger letter
• One of the small letters sometimes exhibited
  motion, response slower during these trials
• New objects in scene may automatically receive
  high attentional priority

6
Four Hypotheses

• Members of the alternating banner group will
  perform more fixations on ads than the static
  banner group
• Members of the alternating group will exhibit
  longer search times
• Members of the alternating banner group will have
  increased recognition memory for the ads
• Members of the alternating banner group will
  exhibit a more negative attitude toward the
  advertisements

7
Methods

• Participants searched a simulated website for the
  answers to six questions.
• Alternating banners or static banners present on
  the screen during the search task.
• Participants eyes were tracked as they searched
  the experimental website.

8
Independent Variable

• Type of banner advertisement present
• - Static banner advertisement
• - Alternating banner advertisement
• Variable manipulated between subjects
• Participants randomly assigned to either group

9
IV Banner Advertisements

• Six ads, each one advertising a different
  product, each ad shown only once
• Each ad shown for the length of one question, ad
  shown changed each time the participant clicks
  the Done button and returned to the homepage
• Order in which the ads were presented randomized
  using a Latin Square design

10
Dependent Variables

• Attention capture, measured by the number of
  fixations performed in the ROI (banner area)
  during entire experiment
• Search times for the experimental questions
• Implicit memory for the advertisement content,
  measured by a recognition task
• Attitude toward the ads, measured by a
  questionnaire

11
Participants

• 10 participants
• 2 male, 8 female
• Age range 18-21

12
Materials

• Simulated website implemented in C and OpenGL
• Replicated an Internet news site, The Herald
  Online
• Contained buttons and text links for the user to
  click on
• Banner always present at the top center of the
  screen

13
Materials

• Recognition task implemented in Flash and taken
  online
• Shown 12 images of ads, only 6 of which had
  actually been present during the experiment
• Attitude questionnaire

14
Materials Eye Tracker Lab
15
Materials Capture Program
16
Materials Capture Program
17
Materials Recognition Task
18
Procedure

• Six questions asked, one at a time, each one read
  aloud to participants
• Participants clicked a Done button present on
  the screen before stating the answer they found
  aloud
• Participants who made incorrect answers were
  allowed to move on
• All search times measured with a stopwatch, began
  when experimenter finished reading question
  aloud, ended when participant clicked Done
  button

19
Procedure

• During experiment, participants were presented
  with 6 different banner advertisements
• Used eye tracker to measure total number of
  fixations
• Each question had a different banner
• Banner changed only when participant clicked
  Done

20
Procedure

• After completing the search task, participants
  performed the recognition task using a web
  browser
• Participants then completed the attitude
  questionnaire (Answered on a 5 point Likert-type
  scale)

21
Results Average Number of Fixations

• Static banner group made an average of 43
  fixations (SD54.19).
• Alternating banner group made an average of 166
  fixations (SD224.8).
• No significant differences between the
  alternating banner group and the static banner
  group (p0.27).
• Alternating banner group made more overall
  fixations than the static banner group.

22
Results Scanpath Example
23
Results Scanpath Example
24
Results Average Number of Fixations
25
Results Search Times

• No significant difference between the groups for
  each of the experimental questions.
• Alternating banner group had a higher overall
  mean search time than the static banner group
• (p0.04).

26
Results Search Times
Static Alternating t-test significance
Q1 44.60 (SD53.3) 26.00 (SD8.9) .062
Q2 37.40 (SD11.5) 35.40 (SD14.5) .942
Q3 34.00 (SD11.1 35.80 (SD20.7) .414
Q4 31.40 (SD24.4) 63.80 (SD41.7) .363
Q5 87.80 (SD51.8) 110.4 (SD46.1) .478
Q6 56.40 (SD40.5) 71.60 (SD50.5) .412
27
Results Recognition Task

• Scored as a ratio of hits and a ratio of false
  alarms
• Alternating banner group exhibited ratios closer
  to chance (.50)
• Static banner group performed better on
  recognition task

28
Results Attitude Survey

• Answered using a 5 point Likert-type scale
• Both groups exhibited a neutral attitude toward
  the ads that were presented
• No significant differences between the groups for
  each question mean

29
Results Attitude Survey
30
Discussion

• Neither banner ad was significantly more
  effective at capturing user attention
• The number of fixations did not promote the
  recognition memory for the ads
• Contrary to the hypothesis, both groups exhibited
  a neutral attitude toward the banners ads
• Overall users pay little attention to the banner
  ads which supports the idea of Banner Blindness
• Further research on the topic should investigate
  other forms of Internet advertising